From carl.input at gmail.com Sat Mar 1 06:57:02 2014 From: carl.input at gmail.com (Carl Smith) Date: Sat, 1 Mar 2014 11:57:02 +0000 Subject: [IPython-dev] nush Message-ID: Hi all I don't know if anyone here has taken an interest, but it's the weekend, so it seems like a good time to mention it. I started working on a browser-based programming environment a while back that takes a lot from IPython. Though it's very different internally, it has a lot of similarity at the UX level. There's a few ideas there that were discussed on this list a while back, liked tabbed file editing, as well as some reinvention of IPython concepts, like bookmarking paths and defining commands. If you have Python3 and Chrome or Chromium installed, you should be able to just clone the repo and run the app. There's a fairly complete wiki on GitHub to get you up to speed. If you're an IPython user, you may consider nush a bit redundant, but it does do a lot that IPython, by design, never will. The reason I'm plugging it here is only that it has features that IPython developers may find interesting. Thanks, and all the best. Carl Nush on GitHub https://github.com/carlsmith/nush From doug.blank at gmail.com Sat Mar 1 07:51:54 2014 From: doug.blank at gmail.com (Doug Blank) Date: Sat, 1 Mar 2014 07:51:54 -0500 Subject: [IPython-dev] nush In-Reply-To: References: Message-ID: Nice ideas, Carl... thanks for sharing, both code and ideas! It really did work as easily as you say :) -Doug PS - For those that don't have the time/ability to install and run, you can get an idea of the project in the docs, here: https://github.com/carlsmith/nush/wiki On Sat, Mar 1, 2014 at 6:57 AM, Carl Smith wrote: > Hi all > > I don't know if anyone here has taken an interest, but it's the > weekend, so it seems like a good time to mention it. > > I started working on a browser-based programming environment a while > back that takes a lot from IPython. Though it's very different > internally, it has a lot of similarity at the UX level. There's a few > ideas there that were discussed on this list a while back, liked > tabbed file editing, as well as some reinvention of IPython concepts, > like bookmarking paths and defining commands. > > If you have Python3 and Chrome or Chromium installed, you should be > able to just clone the repo and run the app. There's a fairly complete > wiki on GitHub to get you up to speed. > > If you're an IPython user, you may consider nush a bit redundant, but > it does do a lot that IPython, by design, never will. The reason I'm > plugging it here is only that it has features that IPython developers > may find interesting. > > Thanks, and all the best. > > Carl > > > Nush on GitHub > https://github.com/carlsmith/nush > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From carl.input at gmail.com Sat Mar 1 07:52:53 2014 From: carl.input at gmail.com (Carl Smith) Date: Sat, 1 Mar 2014 12:52:53 +0000 Subject: [IPython-dev] nush In-Reply-To: References: Message-ID: Cheers Doug; glad you liked it. From cyrille.rossant at gmail.com Sun Mar 2 10:03:39 2014 From: cyrille.rossant at gmail.com (Cyrille Rossant) Date: Sun, 2 Mar 2014 16:03:39 +0100 Subject: [IPython-dev] LaTeX2HTML5 in the notebook Message-ID: Hi, I'm trying to load the latex2html5 JS library in the notebook (http://latex2html5.com/installation.html). Whenever I load it, I have an error with Backbone.Layout not being defined. I've tried to load backbone.layoutmanager but the problem remains. Any ideas? Thanks, Cyrille From fperez.net at gmail.com Sun Mar 2 17:57:39 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Sun, 2 Mar 2014 14:57:39 -0800 Subject: [IPython-dev] nush In-Reply-To: References: Message-ID: Hi Carl, Thanks for sharing this! It looks like you've done a great job, and I hadn't heard of the Kano project before (http://www.kano.me). On Sat, Mar 1, 2014 at 3:57 AM, Carl Smith wrote: > If you're an IPython user, you may consider nush a bit redundant, but > it does do a lot that IPython, by design, never will. The reason I'm > plugging it here is only that it has features that IPython developers > may find interesting. > Yup, there's obvious similarities (.commands are reminiscent of %magics, shell/os/python blend, etc). But I'm always happy to see folks exploring these ideas in different directions, and with the freedom that starting a new project provides. IPython is by now an old and stodgy project, saddled with the weight of a large user base. That's on many fronts a great thing, as it means we've been successful in providing something of value to a lot of people. But it also means that we're much more constrained in what we can reasonably do without breaking things for existing users. Best of luck with the project! f -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From carl.input at gmail.com Sun Mar 2 19:49:00 2014 From: carl.input at gmail.com (Carl Smith) Date: Mon, 3 Mar 2014 00:49:00 +0000 Subject: [IPython-dev] nush In-Reply-To: References: Message-ID: Cheers Fernando. Thanks for taking the time. You're right; it is a lot of fun being able to play with these ideas so flippantly. IPython's awesome, but it also a huge deal that can't just try out wacky stuff these days. To much important work depends on it. If IPython wasn't around, I'd have maybe taken a more serious attitude with this project, but you guys do the power tools for interactive Python so well now that it makes sense for other people to do something more off key. No sane person would ever use nush for serious work, or expect dependable APIs, commercial support, security etc. It's for personal computing on hobby boxes, and a little bit punk. Anyway, glad you had fun. All the best Carl From wstein at gmail.com Sun Mar 2 21:56:26 2014 From: wstein at gmail.com (William Stein) Date: Sun, 2 Mar 2014 18:56:26 -0800 Subject: [IPython-dev] meeting in Berkeley Message-ID: Hi, Would any IPython notebook devs be interested in meeting in **Berkeley** on Friday, March 21? Is there any time that day that would work for some of you? -- William -- William Stein Professor of Mathematics University of Washington http://wstein.org From benjaminrk at gmail.com Sun Mar 2 23:15:06 2014 From: benjaminrk at gmail.com (MinRK) Date: Sun, 2 Mar 2014 20:15:06 -0800 Subject: [IPython-dev] meeting in Berkeley In-Reply-To: References: Message-ID: Absolutely! Just about any time should work for me. -MinRK On Sun, Mar 2, 2014 at 6:56 PM, William Stein wrote: > Hi, > > Would any IPython notebook devs be interested in meeting in > **Berkeley** on Friday, March 21? Is there any time that day that > would work for some of you? > > -- William > > -- > William Stein > Professor of Mathematics > University of Washington > http://wstein.org > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From kikocorreoso at gmail.com Mon Mar 3 02:21:12 2014 From: kikocorreoso at gmail.com (Kiko) Date: Mon, 3 Mar 2014 08:21:12 +0100 Subject: [IPython-dev] brythonmagic Message-ID: Hi all, I would like to show you an extension I made: https://github.com/kikocorreoso/brythonmagic It adds a cell magic to the IPython notebook and it allows you to run brython code in the notebook. What is Brython? Brython is a Python 3 implementation made in javascript that allows you to run Python code on the client side as it was js code. In the repo you have a notebook with info to install the extension and examples of use as well as examples of how to use other js libraries from the notebook using a more pythonic syntax. Also, you can have a look to a video I made showing the notebook in action in case you don't want to install the extension: http://youtu.be/adQzjuUX0kw I hope it could be interesting for some one. Best regards. -------------- next part -------------- An HTML attachment was scrubbed... URL: From pi at berkeley.edu Mon Mar 3 12:13:50 2014 From: pi at berkeley.edu (Paul Ivanov) Date: Mon, 3 Mar 2014 09:13:50 -0800 Subject: [IPython-dev] brythonmagic In-Reply-To: References: Message-ID: <20140303171350.GY21162@HbI-OTOH.berkeley.edu> Kiko, on 2014-03-03 08:21, wrote: > I would like to show you an extension I made: > > https://github.com/kikocorreoso/brythonmagic ... > Also, you can have a look to a video I made showing the notebook in action > in case you don't want to install the extension: http://youtu.be/adQzjuUX0kw This looks great, thanks for sharing. The OpenLayers example in particular looks quite fun. Might you consider including Brython as a git submodule? The Brython Py_VFS.js file contains the entire Python standard library, from what I understand. -- _ / \ A* \^ - ,./ _.`\\ / \ / ,--.S \/ \ / `"~,_ \ \ __o ? _ \<,_ /:\ --(_)/-(_)----.../ | \ --------------.......J Paul Ivanov http://pirsquared.org From takowl at gmail.com Mon Mar 3 13:05:20 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Mon, 3 Mar 2014 10:05:20 -0800 Subject: [IPython-dev] meeting in Berkeley In-Reply-To: References: Message-ID: Yep, I'll be here all day as well. On 2 March 2014 20:15, MinRK wrote: > Absolutely! Just about any time should work for me. > > -MinRK > > > On Sun, Mar 2, 2014 at 6:56 PM, William Stein wrote: > >> Hi, >> >> Would any IPython notebook devs be interested in meeting in >> **Berkeley** on Friday, March 21? Is there any time that day that >> would work for some of you? >> >> -- William >> >> -- >> William Stein >> Professor of Mathematics >> University of Washington >> http://wstein.org >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From kikocorreoso at gmail.com Mon Mar 3 17:05:21 2014 From: kikocorreoso at gmail.com (Kiko) Date: Mon, 3 Mar 2014 23:05:21 +0100 Subject: [IPython-dev] brythonmagic In-Reply-To: <20140303171350.GY21162@HbI-OTOH.berkeley.edu> References: <20140303171350.GY21162@HbI-OTOH.berkeley.edu> Message-ID: 2014-03-03 18:13 GMT+01:00 Paul Ivanov : > Kiko, on 2014-03-03 08:21, wrote: > > I would like to show you an extension I made: > > > > https://github.com/kikocorreoso/brythonmagic > ... > > Also, you can have a look to a video I made showing the notebook in > action > > in case you don't want to install the extension: > http://youtu.be/adQzjuUX0kw > > This looks great, thanks for sharing. The OpenLayers example in > particular looks quite fun. Might you consider including Brython > as a git submodule? Right now my brython.js is a slight modification of the official brython release. Right now it is just a proof of concept. If it is interesting (I think it is) for the IPython community I can PR my modifications to the official brython repo and then you can use brython.js from a cdn or from the official brython site (brython.info) so it wouldn't be necessary to load the js libs locally. > The Brython Py_VFS.js file contains the > entire Python standard library, from what I understand. > The py_VFS.js doesn't contains the full python stdlib. It contains some js libs and some python libs. You can take a look here: libs (js): https://bitbucket.org/olemis/brython/src/f13e1f0d94637767c4a5f1b399a9e36c5f8ab5ac/src/libs/?at=default Lib (python): https://bitbucket.org/olemis/brython/src/f13e1f0d94637767c4a5f1b399a9e36c5f8ab5ac/src/Lib/?at=default For instance, most of the Python C libs are not ported (for example, itertools was programmed from scratch in python). Depending your use case, it wouldn't be necessary to use most of the Brython libs. I think that the 'browser' and 'javascript' modules would be necessary to play with the document and other js libs and the rest of the libraries are already in Python so in the IPython notebook environment maybe it is unnecessary to use them in Brython as you could use them in Python more effectively. So, depending your use case, py_VFS.js can be created only containing the libs you need and it could be a very light add-on to brython.js. In the scripts folder of the official brython repo you can find the make_VFS.py and the make_dist.py to create your own js files including just what you need. If someone is interested just let me know and I will try to help or you can ask in the official brython mail list ( https://groups.google.com/forum/?fromgroups=#!forum/brython). > > -- > _ > / \ > A* \^ - > ,./ _.`\\ / \ > / ,--.S \/ \ > / `"~,_ \ \ > __o ? > _ \<,_ /:\ > --(_)/-(_)----.../ | \ > --------------.......J > Cool!! :-D Best regards. Kiko http://pybonacci.wordpress.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Mon Mar 3 17:48:07 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Mon, 3 Mar 2014 14:48:07 -0800 Subject: [IPython-dev] brythonmagic In-Reply-To: References: <20140303171350.GY21162@HbI-OTOH.berkeley.edu> Message-ID: Hi Kiko, On 3 March 2014 14:05, Kiko wrote: > Right now my brython.js is a slight modification of the official brython > release. Right now it is just a proof of concept. If it is interesting (I > think it is) for the IPython community I can PR my modifications to the > official brython repo and then you can use brython.js from a cdn or from > the official brython site (brython.info) so it wouldn't be necessary to > load the js libs locally. > I think it's interesting, and if you think the changes are likely to be accepted back into Brython, go for it. Best wishes, Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From fperez.net at gmail.com Mon Mar 3 19:12:03 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Mon, 3 Mar 2014 16:12:03 -0800 Subject: [IPython-dev] meeting in Berkeley In-Reply-To: References: Message-ID: Hey William, great! I'll be here that day, so that works perfectly. Time-wise it's pretty open, let us know what works for you. See you then, f On Sun, Mar 2, 2014 at 6:56 PM, William Stein wrote: > Hi, > > Would any IPython notebook devs be interested in meeting in > **Berkeley** on Friday, March 21? Is there any time that day that > would work for some of you? > > -- William > > -- > William Stein > Professor of Mathematics > University of Washington > http://wstein.org > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From pi at berkeley.edu Mon Mar 3 19:18:27 2014 From: pi at berkeley.edu (Paul Ivanov) Date: Mon, 3 Mar 2014 16:18:27 -0800 Subject: [IPython-dev] meeting in Berkeley In-Reply-To: References: Message-ID: <20140304001827.GB21162@HbI-OTOH.berkeley.edu> >> On Sun, Mar 2, 2014 at 6:56 PM, William Stein wrote: >>> Would any IPython notebook devs be interested in meeting in >>> **Berkeley** on Friday, March 21? Is there any time that day that >>> would work for some of you? >On 2 March 2014 20:15, MinRK wrote: >> Absolutely! Just about any time should work for me. >> >> -MinRK >> >> Thomas Kluyver, on 2014-03-03 10:05, wrote: > Yep, I'll be here all day as well. Ditto for me. If you're feeling adventurous, you can attend the Python Workers' Party 4-6pm. http://python.berkeley.edu/events/2014/01/24/python-workers-party-rally.html cheers, -- _ / \ A* \^ - ,./ _.`\\ / \ / ,--.S \/ \ / `"~,_ \ \ __o ? _ \<,_ /:\ --(_)/-(_)----.../ | \ --------------.......J Paul Ivanov http://pirsquared.org From tritemio at gmail.com Mon Mar 3 21:18:13 2014 From: tritemio at gmail.com (Antonino Ingargiola) Date: Mon, 3 Mar 2014 18:18:13 -0800 Subject: [IPython-dev] Multiple figures and parallel ipython Message-ID: Hi to all, I'm performing some simulation using an ipython cluster on a single machine. A notebook manages the simulation and display results. For testing I would like to plot some large array, let say 20M of points for each engine. I would like to do a loop like this: for each engine and for each "array slice" plot the slice and close the figure. I use matplotlib inline and close() is needed to release the RAM before the next plot. Right now I'm doing this manually alternating this cell (`s` is the index of the slice, S.emission is a pytables array on disk): %%px --target=0 s = 0 em = S.emission[:, s*2e6:(s+1)*2e6] plt.plot(em.T, alpha=0.5); del em and this cell: %px plt.close('all') Is it possible to turn this into a loop? Thanks, Antonio -------------- next part -------------- An HTML attachment was scrubbed... URL: From dineshvadhia at outlook.com Tue Mar 4 14:04:18 2014 From: dineshvadhia at outlook.com (Dinesh Vadhia) Date: Tue, 4 Mar 2014 11:04:18 -0800 Subject: [IPython-dev] Confused about ipython.parallel namespaces Message-ID: Very confused about ipython.parallel namespaces. Below are two programs and the traceback error. Thanks! > ipcluster start -n 2 => this.py from IPython.parallel import Client def main(): that.two_numbers() return if __name__ == u'__main__': rc = Client() dview = rc.direct_view() dview.block = True with dview.sync_imports(): import that rc[0].apply(main) rc[1].apply(main) => that.py def two_numbers(): print 1+2 return => error traceback > python this.py importing that on engine(s) Traceback (most recent call last): File "this.py", line 17, in import that File "C:\Anaconda\lib\contextlib.py", line 24, in __exit__ self.gen.next() File "C:\Anaconda\lib\site-packages\IPython\parallel\client\view.py", line 511, in sync_imports r.get() File "C:\Anaconda\lib\site-packages\IPython\parallel\client\asyncresult.py", line 126, in get raise self._exception IPython.parallel.error.CompositeError: one or more exceptions from call to method: remote_import [0:apply]: ImportError: No module named that [1:apply]: ImportError: No module named that -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Tue Mar 4 14:39:18 2014 From: benjaminrk at gmail.com (MinRK) Date: Tue, 4 Mar 2014 11:39:18 -0800 Subject: [IPython-dev] Confused about ipython.parallel namespaces In-Reply-To: References: Message-ID: The remote import is calling 'import that' on each engine. If `that.py` is not in the CWD of the engines, then it will not be found. On Tue, Mar 4, 2014 at 11:04 AM, Dinesh Vadhia wrote: > Very confused about ipython.parallel namespaces. Below are two programs > and the traceback error. Thanks! > > > ipcluster start -n 2 > > => this.py > > from IPython.parallel import Client > > def main(): > that.two_numbers() > return > > if __name__ == u'__main__': > > rc = Client() > > dview = rc.direct_view() > dview.block = True > > with dview.sync_imports(): > import that > > rc[0].apply(main) > rc[1].apply(main) > > > => that.py > > def two_numbers(): > print 1+2 > return > > > => error traceback > > > python this.py > > importing that on engine(s) > Traceback (most recent call last): > File "this.py", line 17, in > import that > File "C:\Anaconda\lib\contextlib.py", line 24, in __exit__ > self.gen.next() > File "C:\Anaconda\lib\site-packages\IPython\parallel\client\view.py", > line 511, in sync_imports > r.get() > File > "C:\Anaconda\lib\site-packages\IPython\parallel\client\asyncresult.py", > line 126, in get > raise self._exception > IPython.parallel.error.CompositeError: one or more exceptions from call to > method: remote_import > [0:apply]: ImportError: No module named that > [1:apply]: ImportError: No module named that > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ice.rikh at gmail.com Tue Mar 4 15:29:57 2014 From: ice.rikh at gmail.com (Erik Hvatum) Date: Tue, 4 Mar 2014 14:29:57 -0600 Subject: [IPython-dev] PyQt5? Message-ID: Is anyone else interested in IPython PyQt5 support? PyQt5 support would be great for us at the Pincus Lab as we have a lot of code using PyQt5 that would potentially cooperate with the IPython event loop and integrate nicely - if only IPython supported PyQt5. I did some preliminary work on making IPython 2.0 code from the git repo use PyQt5. The primary problems I've encountered are that much of the QtGui namespace contents have been broken out into QtWidgets (easily fixed - add QtWidgets imports and replace QtGui with QtWidgets as required), and that __init__ behavior for objects multiply inheriting from Qt has changed (could be tricky, I haven't dug into this yet). Adding PyQt5 support alongside the existing support for PyQt4 and pyside qt4 is perhaps possible, though not something I need. Anyway, I have the impression that PyQt5 support is not on the radar at all, and I want it to be! It would be a real let down if IPython were stuck on Qt4 for the next decade due to IPython users writing new software for PyQt4 in order to be compatible with IPython... Thanks, Erik From dineshvadhia at outlook.com Tue Mar 4 15:37:01 2014 From: dineshvadhia at outlook.com (Dinesh Vadhia) Date: Tue, 4 Mar 2014 12:37:01 -0800 Subject: [IPython-dev] Confused about ipython.parallel namespaces Message-ID: I thought that dview.sync_imports() imported modules into the namespace of each engine. How do you make "that.py" be in the current working directory of the engines for the import to be successful? Thanks. -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Tue Mar 4 15:39:11 2014 From: benjaminrk at gmail.com (MinRK) Date: Tue, 4 Mar 2014 12:39:11 -0800 Subject: [IPython-dev] PyQt5? In-Reply-To: References: Message-ID: You are right that PyQt5 is not really on our radar. The Qt code in IPython is not getting much attention at this point - it basically works, and we do not have any immediate plans to do more interesting things with it. I think it is too late to get PyQt5 support into IPython 2.0, but I think it is quite reasonable for 3.0. We already have a compatibility shim in IPython.external.qt, so that is where I would start to look for dealing with whatever changes PyQt5 makes. -MinRK On Tue, Mar 4, 2014 at 12:29 PM, Erik Hvatum wrote: > Is anyone else interested in IPython PyQt5 support? PyQt5 support > would be great for us at the Pincus Lab as we have a lot of code using > PyQt5 that would potentially cooperate with the IPython event loop and > integrate nicely - if only IPython supported PyQt5. > > I did some preliminary work on making IPython 2.0 code from the git > repo use PyQt5. The primary problems I've encountered are that much > of the QtGui namespace contents have been broken out into QtWidgets > (easily fixed - add QtWidgets imports and replace QtGui with QtWidgets > as required), and that __init__ behavior for objects multiply > inheriting from Qt has changed (could be tricky, I haven't dug into > this yet). > > Adding PyQt5 support alongside the existing support for PyQt4 and > pyside qt4 is perhaps possible, though not something I need. > > Anyway, I have the impression that PyQt5 support is not on the radar > at all, and I want it to be! It would be a real let down if IPython > were stuck on Qt4 for the next decade due to IPython users writing new > software for PyQt4 in order to be compatible with IPython... > > Thanks, > Erik > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Tue Mar 4 15:46:59 2014 From: benjaminrk at gmail.com (MinRK) Date: Tue, 4 Mar 2014 12:46:59 -0800 Subject: [IPython-dev] Confused about ipython.parallel namespaces In-Reply-To: References: Message-ID: On Tue, Mar 4, 2014 at 12:37 PM, Dinesh Vadhia wrote: I thought that dview.sync_imports() imported modules into the namespace of > each engine. How do you make "that.py" be in the current working > directory of the engines for the import to be successful? Thanks. > That is what it does. It performs the same import locally and on all engines. What it does not do is send local modules to the engines before importing them. Are your engines running on your local machine? If so, you can ensure the engines are running in the same cwd as the client: rc[:]['cwd'] = os.getcwd() rc[:]execute("import os; os.chdir(cwd)") > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From dineshvadhia at outlook.com Tue Mar 4 17:34:48 2014 From: dineshvadhia at outlook.com (Dinesh Vadhia) Date: Tue, 4 Mar 2014 14:34:48 -0800 Subject: [IPython-dev] Confused about ipython.parallel namespaces In-Reply-To: References: Message-ID: Got it and it works! I'm reviewing and testing ipython.parallel on a local machine. How would this situation change on a cluster of machines? From dineshvadhia at outlook.com Tue Mar 4 17:52:42 2014 From: dineshvadhia at outlook.com (Dinesh Vadhia) Date: Tue, 4 Mar 2014 14:52:42 -0800 Subject: [IPython-dev] ipython.parallel and localhost Message-ID: On a local machine, with > ipcluster start -n 4 is it possible to find out from ipython.parallel the port number being used with localhost (127.0.0.1) for each engine? -------------- next part -------------- An HTML attachment was scrubbed... URL: From fperez.net at gmail.com Tue Mar 4 18:07:37 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Tue, 4 Mar 2014 15:07:37 -0800 Subject: [IPython-dev] PyQt5? In-Reply-To: References: Message-ID: On Tue, Mar 4, 2014 at 12:39 PM, MinRK wrote: > You are right that PyQt5 is not really on our radar. The Qt code in > IPython is not getting much attention at this point - it basically works, > and we do not have any immediate plans to do more interesting things with > it. I think it is too late to get PyQt5 support into IPython 2.0, but I > think it is quite reasonable for 3.0. +1, thanks Min. -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Tue Mar 4 18:54:29 2014 From: benjaminrk at gmail.com (MinRK) Date: Tue, 4 Mar 2014 15:54:29 -0800 Subject: [IPython-dev] Confused about ipython.parallel namespaces In-Reply-To: References: Message-ID: > How would this situation change on a cluster of machines? You would need to make sure to distribute the file to all of the machines where you engines run. On Tue, Mar 4, 2014 at 2:34 PM, Dinesh Vadhia wrote: > Got it and it works! > > I'm reviewing and testing ipython.parallel on a local machine. > > How would this situation change on a cluster of machines? > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Tue Mar 4 18:55:32 2014 From: benjaminrk at gmail.com (MinRK) Date: Tue, 4 Mar 2014 15:55:32 -0800 Subject: [IPython-dev] ipython.parallel and localhost In-Reply-To: References: Message-ID: On Tue, Mar 4, 2014 at 2:52 PM, Dinesh Vadhia wrote: > On a local machine, with > > > ipcluster start -n 4 > > is it possible to find out from ipython.parallel the port number being > used with localhost (127.0.0.1) for each engine? > The engines do not listen on any ports. Only the controller listens on ports. These are listed in ~/.ipython/profile_default/security/ipcontroller-*.json > > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ice.rikh at gmail.com Tue Mar 4 22:44:27 2014 From: ice.rikh at gmail.com (Erik Hvatum) Date: Tue, 4 Mar 2014 21:44:27 -0600 Subject: [IPython-dev] PyQt5? In-Reply-To: References: Message-ID: Thanks, guys. I'll put some thought into shimming pyqt5 along side pyqt4 and pyside. If the pyqt related code is static, then I can just dump a patch set on here, and it will turn up as a useful google result for the silent majority of pyqt5 users ;) -Erik On Tue, Mar 4, 2014 at 5:07 PM, Fernando Perez wrote: > > On Tue, Mar 4, 2014 at 12:39 PM, MinRK wrote: >> >> You are right that PyQt5 is not really on our radar. The Qt code in >> IPython is not getting much attention at this point - it basically works, >> and we do not have any immediate plans to do more interesting things with >> it. I think it is too late to get PyQt5 support into IPython 2.0, but I >> think it is quite reasonable for 3.0. > > > +1, thanks Min. > > > -- > Fernando Perez (@fperez_org; http://fperez.org) > fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) > fernando.perez-at-berkeley: contact me here for any direct mail > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From jenshnielsen at gmail.com Wed Mar 5 05:04:30 2014 From: jenshnielsen at gmail.com (Jens Nielsen) Date: Wed, 5 Mar 2014 10:04:30 +0000 Subject: [IPython-dev] PyQt5? In-Reply-To: References: Message-ID: There is a pull request with some work towards a QT5 backend for matplotlib https://github.com/matplotlib/matplotlib/pull/2471. It might be useful to coordinate with this. At lease the mapping between backend used for the %pylab and %matplotlib would likely need updating when matplotlib has a qt5 backend. Jens On Wed, Mar 5, 2014 at 3:44 AM, Erik Hvatum wrote: > Thanks, guys. I'll put some thought into shimming pyqt5 along side > pyqt4 and pyside. If the pyqt related code is static, then I can just > dump a patch set on here, and it will turn up as a useful google > result for the silent majority of pyqt5 users ;) > > -Erik > > On Tue, Mar 4, 2014 at 5:07 PM, Fernando Perez > wrote: > > > > On Tue, Mar 4, 2014 at 12:39 PM, MinRK wrote: > >> > >> You are right that PyQt5 is not really on our radar. The Qt code in > >> IPython is not getting much attention at this point - it basically > works, > >> and we do not have any immediate plans to do more interesting things > with > >> it. I think it is too late to get PyQt5 support into IPython 2.0, but I > >> think it is quite reasonable for 3.0. > > > > > > +1, thanks Min. > > > > > > -- > > Fernando Perez (@fperez_org; http://fperez.org) > > fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) > > fernando.perez-at-berkeley: contact me here for any direct mail > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From konrad.hinsen at fastmail.net Wed Mar 5 07:02:46 2014 From: konrad.hinsen at fastmail.net (Konrad Hinsen) Date: Wed, 5 Mar 2014 13:02:46 +0100 Subject: [IPython-dev] Adding custom JavaScript code Message-ID: <21271.4710.607283.650456@Konrad-Hinsens-MacBook-Pro-2.local> Hi everyone, I am trying to add JavaScript code to the file static/custom/custom.js in my profile. As a first test, I simply uncommented the "custom button" example. That makes no difference at all. I tried other additions, and made some explorations in the Browser console, which all led me to conclude that custom.js is not used at all. I have only one profile (the default one), so I am rather sure I am not using a wrong one. Pretty much everything I find about this subject on the Web is for IPython 1.x. Is custom.js still supposed to work with 2.x? Does anyone have a working example? Or ideas for debugging this? Konrad. From vasco+python at tenner.nl Wed Mar 5 07:05:41 2014 From: vasco+python at tenner.nl (Vasco Tenner) Date: Wed, 05 Mar 2014 13:05:41 +0100 Subject: [IPython-dev] --pylab flag deprecated Message-ID: <53171315.5000908@tenner.nl> Hi, I used to start ipython notebook with the --pylab flag, so I have all kinds of usefull functions directly available. By guessing and tab-completion many usefull functions, for data analysis popped up. (like plot, display etc). Now the --pylab flag will deprecate soon, what is the best alternative for an interactive environment. I already found it usefull to import from pylab: from pylab import * However, the --pylab flag did more: it also loaded a bunch of ipython functions, like IPython.display.display. Is there a preffered way to load this? To configure this in some enviroment? I just found the %pylab magic. Should I put this in the top cell of all my notebooks? What should I advice to my starting collegues? Groet, Vasco From doug.blank at gmail.com Wed Mar 5 07:10:42 2014 From: doug.blank at gmail.com (Doug Blank) Date: Wed, 5 Mar 2014 07:10:42 -0500 Subject: [IPython-dev] Adding custom JavaScript code In-Reply-To: <21271.4710.607283.650456@Konrad-Hinsens-MacBook-Pro-2.local> References: <21271.4710.607283.650456@Konrad-Hinsens-MacBook-Pro-2.local> Message-ID: On Wed, Mar 5, 2014 at 7:02 AM, Konrad Hinsen wrote: > Hi everyone, > > I am trying to add JavaScript code to the file static/custom/custom.js > in my profile. As a first test, I simply uncommented the "custom > button" example. That makes no difference at all. I tried other > additions, and made some explorations in the Browser console, which > all led me to conclude that custom.js is not used at all. > > I have only one profile (the default one), so I am rather sure I am > not using a wrong one. > > Pretty much everything I find about this subject on the Web is for > IPython 1.x. Is custom.js still supposed to work with 2.x? > Does anyone have a working example? Or ideas for debugging this? Others will no doubt have detailed help, but here is an example (the drag and drop images on a notebook) that works with IPython 2.0: https://bitbucket.org/ipre/calico/src/master/notebooks/profile/static/custom/custom.js#cl-52 We copy all of those files into new profiles and works fine. As for debugging, I use the Chrome developer tools (right-click on page, inspect element; select Console tab) to see errors in loading. -Doug > Konrad. > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From fperez.net at gmail.com Wed Mar 5 10:07:42 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Wed, 5 Mar 2014 07:07:42 -0800 Subject: [IPython-dev] --pylab flag deprecated In-Reply-To: <53171315.5000908@tenner.nl> References: <53171315.5000908@tenner.nl> Message-ID: Hi Vasco, the recommended approach, that will make in the long run your notebooks more readable, is to use %matplotlib inline import numpy as np import matplotlib.pyplot as plt atop your notebooks. That will make it easier to know what comes from where. Now, if you *really* want to have a lot of top-level plotting and numerical functions, and are willing to pay the price that for example doing a lot of automatic imports clobbers the built-in python sum() function with numpy.sum() (which can cause some surprises), you can still do %pylab inline at the top. We will probably never remove this, partly because it's still useful for *quick and dirty* analysis where you just want to do something fast with minimal overhead, and is mostly throw-away scratch work. This kind of typing economy is also still very useful at the console. But in the long run, our experience is that notebooks tend to linger for a long time: in that scenario, it's worth paying the price of slightly more verbose code with explicit imports, in order to reap the benefits of readability in the long run. Cheers, f On Wed, Mar 5, 2014 at 4:05 AM, Vasco Tenner wrote: > Hi, > I used to start ipython notebook with the --pylab flag, so I have all > kinds of usefull functions directly available. By guessing and > tab-completion many usefull functions, for data analysis popped up. > (like plot, display etc). > > Now the --pylab flag will deprecate soon, what is the best alternative > for an interactive environment. I already found it usefull to import > from pylab: > > from pylab import * > > However, the --pylab flag did more: it also loaded a bunch of ipython > functions, like IPython.display.display. > > Is there a preffered way to load this? To configure this in some > enviroment? > > I just found the > %pylab > magic. Should I put this in the top cell of all my notebooks? What > should I advice to my starting collegues? > > Groet, > Vasco > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From abie at uw.edu Wed Mar 5 12:53:04 2014 From: abie at uw.edu (Abraham D. Flaxman) Date: Wed, 5 Mar 2014 17:53:04 +0000 Subject: [IPython-dev] --pylab flag deprecated In-Reply-To: References: <53171315.5000908@tenner.nl> Message-ID: This reminds me, is there an easy way to make the first cell in my notebook include these lines by default? And that reminds me, is there an easy way to make the name of a notebook more descriptive than UntitledXXX by default? Thanks for all your work on this amazing project! --Abie From: ipython-dev-bounces at scipy.org [mailto:ipython-dev-bounces at scipy.org] On Behalf Of Fernando Perez Sent: Wednesday, March 5, 2014 7:08 AM To: IPython developers list Subject: Re: [IPython-dev] --pylab flag deprecated Hi Vasco, the recommended approach, that will make in the long run your notebooks more readable, is to use %matplotlib inline import numpy as np import matplotlib.pyplot as plt atop your notebooks. That will make it easier to know what comes from where. Now, if you *really* want to have a lot of top-level plotting and numerical functions, and are willing to pay the price that for example doing a lot of automatic imports clobbers the built-in python sum() function with numpy.sum() (which can cause some surprises), you can still do %pylab inline at the top. We will probably never remove this, partly because it's still useful for *quick and dirty* analysis where you just want to do something fast with minimal overhead, and is mostly throw-away scratch work. This kind of typing economy is also still very useful at the console. But in the long run, our experience is that notebooks tend to linger for a long time: in that scenario, it's worth paying the price of slightly more verbose code with explicit imports, in order to reap the benefits of readability in the long run. Cheers, f On Wed, Mar 5, 2014 at 4:05 AM, Vasco Tenner > wrote: Hi, I used to start ipython notebook with the --pylab flag, so I have all kinds of usefull functions directly available. By guessing and tab-completion many usefull functions, for data analysis popped up. (like plot, display etc). Now the --pylab flag will deprecate soon, what is the best alternative for an interactive environment. I already found it usefull to import from pylab: from pylab import * However, the --pylab flag did more: it also loaded a bunch of ipython functions, like IPython.display.display. Is there a preffered way to load this? To configure this in some enviroment? I just found the %pylab magic. Should I put this in the top cell of all my notebooks? What should I advice to my starting collegues? Groet, Vasco _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From konrad.hinsen at fastmail.net Wed Mar 5 13:59:49 2014 From: konrad.hinsen at fastmail.net (Konrad Hinsen) Date: Wed, 5 Mar 2014 19:59:49 +0100 Subject: [IPython-dev] Adding custom JavaScript code In-Reply-To: References: <21271.4710.607283.650456@Konrad-Hinsens-MacBook-Pro-2.local> Message-ID: <21271.29733.703671.437831@Konrad-Hinsens-MacBook-Pro-2.local> Doug Blank writes: > Others will no doubt have detailed help, but here is an example (the > drag and drop images on a notebook) that works with IPython 2.0: > > https://bitbucket.org/ipre/calico/src/master/notebooks/profile/static/custom/custom.js#cl-52 > > We copy all of those files into new profiles and works fine. Thanks, that means at least that I have a problem that not everyone has. I tried your extensions on my installation, but it makes no difference at all either. There's no sign that anything changes, and in the browser console I can't see any of the definitions made in your JS files. > As for debugging, I use the Chrome developer tools (right-click on > page, inspect element; select Console tab) to see errors in loading. What I need to debug is the mechanism that is supposed to load custom.js. I don't even know if that's in the Python or the JS part of IPython. Konrad. From theodore.r.drain at jpl.nasa.gov Wed Mar 5 14:06:00 2014 From: theodore.r.drain at jpl.nasa.gov (Drain, Theodore R (392P)) Date: Wed, 5 Mar 2014 19:06:00 +0000 Subject: [IPython-dev] IPCluster failing when starting more than a few engines. Message-ID: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB98961@ap-embx-sp20.RES.AD.JPL> Using IPython 2.0.0 dev branch sync'ed on 2014-02-24 11:44:52. Running ipcluster start on a set of machines w/o a shared file system using SSHEngineSetLauncher. I have 6 machines that have between 4 and 12 cores on each machine. If I run ipcluster with 2 engines/machine, it works fine. If I increase it to 3 or higher, I start getting engines that fail to connect. Some failures are failures to connect like look like this: 10:43:30.195 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 10:43:35.196 [IPEngineApp] CRITICAL | Registration timed out after 5.0 seconds Other failures are weirder: 10:43:30.184 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 10:43:30.249 [IPEngineApp] Starting to monitor the heartbeat signal from the hub every 3010 ms. 10:43:30.251 [IPEngineApp] Using existing profile dir: u'.ipython/profile_dev' 10:43:30.252 [IPEngineApp] Completed registration with id 6 10:43:36.273 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). 10:43:39.281 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (2 time(s) in a row). 10:43:42.293 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (3 time(s) in a row). 10:44:36.469 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). 10:44:42.489 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). In the second case, if I connect a client to the controller, there is no engine with ID 6 available even though it seems to be getting some heart beats from the hub. I've tried adding lines like these to my config file and it doesn't help: c.IPClusterStart.delay = 0.5 c.SSHEngineSetLauncher.delay = 0.5 The number of failures increases with the number of engines being started on each machine. Trying to start 12 engines on a single machine is almost a complete failure. Any thoughts on what I should be doing differently? Thanks, Ted From fperez.net at gmail.com Wed Mar 5 16:40:26 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Wed, 5 Mar 2014 16:40:26 -0500 Subject: [IPython-dev] Adding custom JavaScript code In-Reply-To: <21271.29733.703671.437831@Konrad-Hinsens-MacBook-Pro-2.local> References: <21271.4710.607283.650456@Konrad-Hinsens-MacBook-Pro-2.local> <21271.29733.703671.437831@Konrad-Hinsens-MacBook-Pro-2.local> Message-ID: Hey Konrad: JS debugging tip: whenever you see something funky, try reloading the page/app in a *new incognito/private mode window*. Those start with a completely empty cache and are very stateless. I've already been fooled a few times by caching issues that went away once I checked with an incognito window. Credit where it's due, Min taught me this :) f On Wed, Mar 5, 2014 at 1:59 PM, Konrad Hinsen wrote: > Doug Blank writes: > > > Others will no doubt have detailed help, but here is an example (the > > drag and drop images on a notebook) that works with IPython 2.0: > > > > > https://bitbucket.org/ipre/calico/src/master/notebooks/profile/static/custom/custom.js#cl-52 > > > > We copy all of those files into new profiles and works fine. > > Thanks, that means at least that I have a problem that not everyone has. > > I tried your extensions on my installation, but it makes no difference > at all either. There's no sign that anything changes, and in the > browser console I can't see any of the definitions made in your JS > files. > > > As for debugging, I use the Chrome developer tools (right-click on > > page, inspect element; select Console tab) to see errors in loading. > > What I need to debug is the mechanism that is supposed to load > custom.js. I don't even know if that's in the Python or the JS part of > IPython. > > Konrad. > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From claresloggett at gmail.com Wed Mar 5 21:47:35 2014 From: claresloggett at gmail.com (Clare Sloggett) Date: Thu, 6 Mar 2014 13:47:35 +1100 Subject: [IPython-dev] Adding custom JavaScript code In-Reply-To: References: <21271.4710.607283.650456@Konrad-Hinsens-MacBook-Pro-2.local> <21271.29733.703671.437831@Konrad-Hinsens-MacBook-Pro-2.local> Message-ID: +1 ! Also, try navigating to your-url/static/custom/custom.js - you should see your javascript code. That way you can test that you've at least put the file in the location where ipython expects to find it. On 6 March 2014 08:40, Fernando Perez wrote: > Hey Konrad: > > JS debugging tip: whenever you see something funky, try reloading the > page/app in a *new incognito/private mode window*. Those start with a > completely empty cache and are very stateless. I've already been fooled a > few times by caching issues that went away once I checked with an incognito > window. > > Credit where it's due, Min taught me this :) > > f > > > On Wed, Mar 5, 2014 at 1:59 PM, Konrad Hinsen wrote: > >> Doug Blank writes: >> >> > Others will no doubt have detailed help, but here is an example (the >> > drag and drop images on a notebook) that works with IPython 2.0: >> > >> > >> https://bitbucket.org/ipre/calico/src/master/notebooks/profile/static/custom/custom.js#cl-52 >> > >> > We copy all of those files into new profiles and works fine. >> >> Thanks, that means at least that I have a problem that not everyone has. >> >> I tried your extensions on my installation, but it makes no difference >> at all either. There's no sign that anything changes, and in the >> browser console I can't see any of the definitions made in your JS >> files. >> >> > As for debugging, I use the Chrome developer tools (right-click on >> > page, inspect element; select Console tab) to see errors in loading. >> >> What I need to debug is the mechanism that is supposed to load >> custom.js. I don't even know if that's in the Python or the JS part of >> IPython. >> >> Konrad. >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > > > > -- > Fernando Perez (@fperez_org; http://fperez.org) > fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) > fernando.perez-at-berkeley: contact me here for any direct mail > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From fperez.net at gmail.com Thu Mar 6 01:39:03 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Thu, 6 Mar 2014 01:39:03 -0500 Subject: [IPython-dev] A long shot: weird Chinese/Win XP/Anaconda/Notebook bug Message-ID: Hey folks, I know this is a super long shot, but I figured I'd post it in case anyone has an idea... At the workshop on teaching CS with the IPython notebook that just finished up, one attendee came up with a weird bug. The platform was a Windows XP laptop with Chinese locale, running anaconda 1.9.1 32 bit (IPython 1.2.1). Given the error was in the awful cmd.exe console and all menus were in Chinese, I didn't do a very good job of capturing the traceback. Below is what I was able to get. If anyone recognizes the error and there's any chance it's something we can fix on master, that would be great. Cheers, f ### Truncated traceback init() File "C:\Anaconda\lib\mimetypes.py", line 358, in init db.read_windows_registry() File "C:\Anaconda\lib\mimetypes.py", line 258, in read_windows_registry for subkeyname in enum_types(hkcr): File "C:\Anaconda\lib\mimetypes.py", line 249, in enum_types ctype = ctype.encode(default_encoding) # omit in 3.x! UnicodeDecodeError: 'ascii' codec can't decode byte 0xa4 in position 9: ordina not in range(128) 2014-03-06 11:02:31.858 [tornado.access] ERROR | 500 GET /static/custom/custom s (127.0.0.1) 109.00ms 2014-03-06 11:02:32.000 [tornado.application] ERROR | Uncaught exception GET / atic/base/images/favicon.ico (127.0.0.1) HTTPRequest(protocol='http', host='127.0.0.1:8889', method='GET', uri='/static ase/images/favicon.ico', version='HTTP/1.1', remote_ip='127.0.0.1', headers={' nnection': 'keep-alive', 'Accept-Language': 'zh-TW,zh;q=0.8,en-US;q=0.6,en;q=0 ', 'Accept-Encoding': 'gzip,deflate,sdch', 'Host': '127.0.0.1:8889', 'Accept': */*', 'User-Agent': 'Mozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 (KHTML, l e Gecko) Chrome/33.0.1750.146 Safari/537.36'}) Traceback (most recent call last): File "C:\Anaconda\lib\site-packages\tornado\web.py", line 1218, in _when_com ete callback() File "C:\Anaconda\lib\site-packages\tornado\web.py", line 1239, in _execute_ thod self._when_complete(method(*self.path_args, **self.path_kwargs), File "C:\Anaconda\lib\site-packages\IPython\html\base\handlers.py", line 320 in get mime_type, encoding = mimetypes.guess_type(abspath) File "C:\Anaconda\lib\mimetypes.py", line 297, in guess_type init() File "C:\Anaconda\lib\mimetypes.py", line 358, in init db.read_windows_registry() File "C:\Anaconda\lib\mimetypes.py", line 258, in read_windows_registry for subkeyname in enum_types(hkcr): File "C:\Anaconda\lib\mimetypes.py", line 249, in enum_types ctype = ctype.encode(default_encoding) # omit in 3.x! UnicodeDecodeError: 'ascii' codec can't decode byte 0xa4 in position 9: ordina not in range(128) 2014-03-06 11:02:32.000 [tornado.access] ERROR | 500 GET /static/base/images/f icon.ico (127.0.0.1) 125.00ms -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From christophe.pradal at inria.fr Thu Mar 6 05:08:16 2014 From: christophe.pradal at inria.fr (Christophe Pradal) Date: Thu, 06 Mar 2014 11:08:16 +0100 Subject: [IPython-dev] A long shot: weird Chinese/Win XP/Anaconda/Notebook bug In-Reply-To: References: Message-ID: <53184910.9000605@inria.fr> Hi everyone, It seems that this error is the same that the one reported in the Python issues: http://bugs.python.org/issue9291 Cheers, Christophe Pradal Le 06/03/2014 07:39, Fernando Perez a ?crit : > Hey folks, > > I know this is a super long shot, but I figured I'd post it in case > anyone has an idea... At the workshop on teaching CS with the IPython > notebook that just finished up, one attendee came up with a weird bug. > The platform was a Windows XP laptop with Chinese locale, running > anaconda 1.9.1 32 bit (IPython 1.2.1). > > Given the error was in the awful cmd.exe console and all menus were in > Chinese, I didn't do a very good job of capturing the traceback. Below > is what I was able to get. If anyone recognizes the error and there's > any chance it's something we can fix on master, that would be great. > > Cheers, > > f > > ### Truncated traceback > > init() > File "C:\Anaconda\lib\mimetypes.py", line 358, in init > db.read_windows_registry() > File "C:\Anaconda\lib\mimetypes.py", line 258, in read_windows_registry > for subkeyname in enum_types(hkcr): > File "C:\Anaconda\lib\mimetypes.py", line 249, in enum_types > ctype = ctype.encode(default_encoding) # omit in 3.x! > UnicodeDecodeError: 'ascii' codec can't decode byte 0xa4 in position > 9: ordina > not in range(128) > 2014-03-06 11:02:31.858 [tornado.access] ERROR | 500 GET > /static/custom/custom > s (127.0.0.1) 109.00ms > 2014-03-06 11:02:32.000 [tornado.application] ERROR | Uncaught > exception GET / > atic/base/images/favicon.ico (127.0.0.1) > HTTPRequest(protocol='http', host='127.0.0.1:8889 > ', method='GET', uri='/static > ase/images/favicon.ico', version='HTTP/1.1', remote_ip='127.0.0.1', > headers={' > nnection': 'keep-alive', 'Accept-Language': > 'zh-TW,zh;q=0.8,en-US;q=0.6,en;q=0 > ', 'Accept-Encoding': 'gzip,deflate,sdch', 'Host': '127.0.0.1:8889 > ', 'Accept': > */*', 'User-Agent': 'Mozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 > (KHTML, l > e Gecko) Chrome/33.0.1750.146 Safari/537.36'}) > Traceback (most recent call last): > File "C:\Anaconda\lib\site-packages\tornado\web.py", line 1218, in > _when_com > ete > callback() > File "C:\Anaconda\lib\site-packages\tornado\web.py", line 1239, in > _execute_ > thod > self._when_complete(method(*self.path_args, **self.path_kwargs), > File "C:\Anaconda\lib\site-packages\IPython\html\base\handlers.py", > line 320 > in get > mime_type, encoding = mimetypes.guess_type(abspath) > File "C:\Anaconda\lib\mimetypes.py", line 297, in guess_type > init() > File "C:\Anaconda\lib\mimetypes.py", line 358, in init > db.read_windows_registry() > File "C:\Anaconda\lib\mimetypes.py", line 258, in read_windows_registry > for subkeyname in enum_types(hkcr): > File "C:\Anaconda\lib\mimetypes.py", line 249, in enum_types > ctype = ctype.encode(default_encoding) # omit in 3.x! > UnicodeDecodeError: 'ascii' codec can't decode byte 0xa4 in position > 9: ordina > not in range(128) > 2014-03-06 11:02:32.000 [tornado.access] ERROR | 500 GET > /static/base/images/f > icon.ico (127.0.0.1) 125.00ms > > -- > Fernando Perez (@fperez_org; http://fperez.org) > fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) > fernando.perez-at-berkeley: contact me here for any direct mail > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: From konrad.hinsen at fastmail.net Thu Mar 6 12:14:20 2014 From: konrad.hinsen at fastmail.net (Konrad Hinsen) Date: Thu, 6 Mar 2014 18:14:20 +0100 Subject: [IPython-dev] Adding custom JavaScript code In-Reply-To: References: <21271.4710.607283.650456@Konrad-Hinsens-MacBook-Pro-2.local> <21271.29733.703671.437831@Konrad-Hinsens-MacBook-Pro-2.local> Message-ID: <21272.44268.659045.106995@Konrad-Hinsens-MacBook-Pro-2.local> Fernando Perez writes: > JS debugging tip: whenever you see something funky, try reloading > the page/app in a *new incognito/private mode window*. Those start > with a completely empty cache and are very stateless. ?I've already > been fooled a few times by caching issues that went away once I > checked with an incognito window. Excellent advice, because that turned out to be the cause of my problem. I still don't claim to understand how Firefox caches my custom.js to produce the behavior I have seen, but I least I know how to work around it. Clare Sloggett writes: > Also, try navigating to your-url/static/custom/custom.js - you should see your > javascript code. That way you can test that you've at least put the file in the > location where ipython expects to find it. Excellent advice as well because that's how I analyzed my problem. Thanks to both of you! Konrad From fperez.net at gmail.com Thu Mar 6 13:18:04 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Thu, 6 Mar 2014 13:18:04 -0500 Subject: [IPython-dev] A long shot: weird Chinese/Win XP/Anaconda/Notebook bug In-Reply-To: <53184910.9000605@inria.fr> References: <53184910.9000605@inria.fr> Message-ID: Thanks a lot for tracking that down! Bad news for the person who had the problem, but at least it's off our plate. On Mar 6, 2014 5:08 AM, "Christophe Pradal" wrote: > Hi everyone, > It seems that this error is the same that the one reported in the Python > issues: > http://bugs.python.org/issue9291 > > Cheers, > Christophe Pradal > > Le 06/03/2014 07:39, Fernando Perez a ?crit : > > Hey folks, > > I know this is a super long shot, but I figured I'd post it in case > anyone has an idea... At the workshop on teaching CS with the IPython > notebook that just finished up, one attendee came up with a weird bug. The > platform was a Windows XP laptop with Chinese locale, running anaconda > 1.9.1 32 bit (IPython 1.2.1). > > Given the error was in the awful cmd.exe console and all menus were in > Chinese, I didn't do a very good job of capturing the traceback. Below is > what I was able to get. If anyone recognizes the error and there's any > chance it's something we can fix on master, that would be great. > > Cheers, > > f > > ### Truncated traceback > > init() > File "C:\Anaconda\lib\mimetypes.py", line 358, in init > db.read_windows_registry() > File "C:\Anaconda\lib\mimetypes.py", line 258, in read_windows_registry > for subkeyname in enum_types(hkcr): > File "C:\Anaconda\lib\mimetypes.py", line 249, in enum_types > ctype = ctype.encode(default_encoding) # omit in 3.x! > UnicodeDecodeError: 'ascii' codec can't decode byte 0xa4 in position 9: > ordina > not in range(128) > 2014-03-06 11:02:31.858 [tornado.access] ERROR | 500 GET > /static/custom/custom > s (127.0.0.1) 109.00ms > 2014-03-06 11:02:32.000 [tornado.application] ERROR | Uncaught exception > GET / > atic/base/images/favicon.ico (127.0.0.1) > HTTPRequest(protocol='http', host='127.0.0.1:8889', method='GET', > uri='/static > ase/images/favicon.ico', version='HTTP/1.1', remote_ip='127.0.0.1', > headers={' > nnection': 'keep-alive', 'Accept-Language': > 'zh-TW,zh;q=0.8,en-US;q=0.6,en;q=0 > ', 'Accept-Encoding': 'gzip,deflate,sdch', 'Host': '127.0.0.1:8889', > 'Accept': > */*', 'User-Agent': 'Mozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 > (KHTML, l > e Gecko) Chrome/33.0.1750.146 Safari/537.36'}) > Traceback (most recent call last): > File "C:\Anaconda\lib\site-packages\tornado\web.py", line 1218, in > _when_com > ete > callback() > File "C:\Anaconda\lib\site-packages\tornado\web.py", line 1239, in > _execute_ > thod > self._when_complete(method(*self.path_args, **self.path_kwargs), > File "C:\Anaconda\lib\site-packages\IPython\html\base\handlers.py", line > 320 > in get > mime_type, encoding = mimetypes.guess_type(abspath) > File "C:\Anaconda\lib\mimetypes.py", line 297, in guess_type > init() > File "C:\Anaconda\lib\mimetypes.py", line 358, in init > db.read_windows_registry() > File "C:\Anaconda\lib\mimetypes.py", line 258, in read_windows_registry > for subkeyname in enum_types(hkcr): > File "C:\Anaconda\lib\mimetypes.py", line 249, in enum_types > ctype = ctype.encode(default_encoding) # omit in 3.x! > UnicodeDecodeError: 'ascii' codec can't decode byte 0xa4 in position 9: > ordina > not in range(128) > 2014-03-06 11:02:32.000 [tornado.access] ERROR | 500 GET > /static/base/images/f > icon.ico (127.0.0.1) 125.00ms > > -- > Fernando Perez (@fperez_org; http://fperez.org) > fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) > fernando.perez-at-berkeley: contact me here for any direct mail > > > _______________________________________________ > IPython-dev mailing listIPython-dev at scipy.orghttp://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From theodore.r.drain at jpl.nasa.gov Thu Mar 6 15:37:40 2014 From: theodore.r.drain at jpl.nasa.gov (Drain, Theodore R (392P)) Date: Thu, 6 Mar 2014 20:37:40 +0000 Subject: [IPython-dev] IPCluster failing when starting more than a few engines. In-Reply-To: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB98961@ap-embx-sp20.RES.AD.JPL> References: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB98961@ap-embx-sp20.RES.AD.JPL> Message-ID: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F4E@ap-embx-sp20.RES.AD.JPL> Here's some more information. Hopefully someone can help with this as this problem basically makes IPython parallel unusable. I had our SA's disable the firewall and then things work fine. All the engines start up and connect. With the firewall on, I have to add the line "--enginessh=host" to the controller_args input to enable ssh port forwarding for the connections. When I do that, if I try to launch a single engine on 30 separate computers (with a shared file system), I can only connect 5 of them even though ipcluster log reports that they all connected fine. I'm wondering if there is some timing issue w/ running that many SSH port forward calls (it looks like 3 ports per engine are set up). Any thoughts on what I could try to fix this? Ted ________________________________________ From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] Sent: Wednesday, March 05, 2014 11:06 AM To: IPython developers list Subject: [IPython-dev] IPCluster failing when starting more than a few engines. Using IPython 2.0.0 dev branch sync'ed on 2014-02-24 11:44:52. Running ipcluster start on a set of machines w/o a shared file system using SSHEngineSetLauncher. I have 6 machines that have between 4 and 12 cores on each machine. If I run ipcluster with 2 engines/machine, it works fine. If I increase it to 3 or higher, I start getting engines that fail to connect. Some failures are failures to connect like look like this: 10:43:30.195 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 10:43:35.196 [IPEngineApp] CRITICAL | Registration timed out after 5.0 seconds Other failures are weirder: 10:43:30.184 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 10:43:30.249 [IPEngineApp] Starting to monitor the heartbeat signal from the hub every 3010 ms. 10:43:30.251 [IPEngineApp] Using existing profile dir: u'.ipython/profile_dev' 10:43:30.252 [IPEngineApp] Completed registration with id 6 10:43:36.273 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). 10:43:39.281 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (2 time(s) in a row). 10:43:42.293 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (3 time(s) in a row). 10:44:36.469 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). 10:44:42.489 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). In the second case, if I connect a client to the controller, there is no engine with ID 6 available even though it seems to be getting some heart beats from the hub. I've tried adding lines like these to my config file and it doesn't help: c.IPClusterStart.delay = 0.5 c.SSHEngineSetLauncher.delay = 0.5 The number of failures increases with the number of engines being started on each machine. Trying to start 12 engines on a single machine is almost a complete failure. Any thoughts on what I should be doing differently? Thanks, Ted _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev From theodore.r.drain at jpl.nasa.gov Thu Mar 6 15:51:37 2014 From: theodore.r.drain at jpl.nasa.gov (Drain, Theodore R (392P)) Date: Thu, 6 Mar 2014 20:51:37 +0000 Subject: [IPython-dev] IPCluster failing when starting more than a few engines. In-Reply-To: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F4E@ap-embx-sp20.RES.AD.JPL> References: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB98961@ap-embx-sp20.RES.AD.JPL>, <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F4E@ap-embx-sp20.RES.AD.JPL> Message-ID: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F73@ap-embx-sp20.RES.AD.JPL> One further bit of information: I'm hitting a hard limit of 5 engines connecting using SSH port forwarding. I can run any number of engines locally and it works fine. Could there be some kind of ZMQ limit or SSH limit? The host machine does spawn a huge number of processes - I count 33 processes created when running ipcluster start with a single remote engine which seems a little excessive. ________________________________________ From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] Sent: Thursday, March 06, 2014 12:37 PM To: IPython developers list Subject: Re: [IPython-dev] IPCluster failing when starting more than a few engines. Here's some more information. Hopefully someone can help with this as this problem basically makes IPython parallel unusable. I had our SA's disable the firewall and then things work fine. All the engines start up and connect. With the firewall on, I have to add the line "--enginessh=host" to the controller_args input to enable ssh port forwarding for the connections. When I do that, if I try to launch a single engine on 30 separate computers (with a shared file system), I can only connect 5 of them even though ipcluster log reports that they all connected fine. I'm wondering if there is some timing issue w/ running that many SSH port forward calls (it looks like 3 ports per engine are set up). Any thoughts on what I could try to fix this? Ted ________________________________________ From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] Sent: Wednesday, March 05, 2014 11:06 AM To: IPython developers list Subject: [IPython-dev] IPCluster failing when starting more than a few engines. Using IPython 2.0.0 dev branch sync'ed on 2014-02-24 11:44:52. Running ipcluster start on a set of machines w/o a shared file system using SSHEngineSetLauncher. I have 6 machines that have between 4 and 12 cores on each machine. If I run ipcluster with 2 engines/machine, it works fine. If I increase it to 3 or higher, I start getting engines that fail to connect. Some failures are failures to connect like look like this: 10:43:30.195 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 10:43:35.196 [IPEngineApp] CRITICAL | Registration timed out after 5.0 seconds Other failures are weirder: 10:43:30.184 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 10:43:30.249 [IPEngineApp] Starting to monitor the heartbeat signal from the hub every 3010 ms. 10:43:30.251 [IPEngineApp] Using existing profile dir: u'.ipython/profile_dev' 10:43:30.252 [IPEngineApp] Completed registration with id 6 10:43:36.273 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). 10:43:39.281 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (2 time(s) in a row). 10:43:42.293 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (3 time(s) in a row). 10:44:36.469 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). 10:44:42.489 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). In the second case, if I connect a client to the controller, there is no engine with ID 6 available even though it seems to be getting some heart beats from the hub. I've tried adding lines like these to my config file and it doesn't help: c.IPClusterStart.delay = 0.5 c.SSHEngineSetLauncher.delay = 0.5 The number of failures increases with the number of engines being started on each machine. Trying to start 12 engines on a single machine is almost a complete failure. Any thoughts on what I should be doing differently? Thanks, Ted _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev From theodore.r.drain at jpl.nasa.gov Thu Mar 6 17:00:13 2014 From: theodore.r.drain at jpl.nasa.gov (Drain, Theodore R (392P)) Date: Thu, 6 Mar 2014 22:00:13 +0000 Subject: [IPython-dev] IPCluster failing when starting more than a few engines. In-Reply-To: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F73@ap-embx-sp20.RES.AD.JPL> References: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB98961@ap-embx-sp20.RES.AD.JPL>, <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F4E@ap-embx-sp20.RES.AD.JPL>, <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F73@ap-embx-sp20.RES.AD.JPL> Message-ID: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99FF2@ap-embx-sp20.RES.AD.JPL> Sorry to keep spamming the list but... It appears the problem I'm having is purely timing (or timeout) based. If I run ipengine by hand after the controller comes up, I can connect more than 5 engines (so it's not a resource problem). I then tried hacking hub.py which has a line like this: self.registration_timeout = max(5000, 2*self.heartmonitor.period) If I change that to 60000 (60 seconds), I can get a few more engines to connect but it's basically guesswork as to how many make it up. And there isn't a config option input for that timeout so that isn't much of a solution even if I could come up w/ a time that worked. At this point I'm thinking I'm going to have to write my own version of "ipcluster" that runs the controller, sets up the port forwards, and spawns the engines. Perhaps if I have more control over how that happens that I can get a cluster that will reliably start up. ________________________________________ From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] Sent: Thursday, March 06, 2014 12:51 PM To: IPython developers list Subject: Re: [IPython-dev] IPCluster failing when starting more than a few engines. One further bit of information: I'm hitting a hard limit of 5 engines connecting using SSH port forwarding. I can run any number of engines locally and it works fine. Could there be some kind of ZMQ limit or SSH limit? The host machine does spawn a huge number of processes - I count 33 processes created when running ipcluster start with a single remote engine which seems a little excessive. ________________________________________ From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] Sent: Thursday, March 06, 2014 12:37 PM To: IPython developers list Subject: Re: [IPython-dev] IPCluster failing when starting more than a few engines. Here's some more information. Hopefully someone can help with this as this problem basically makes IPython parallel unusable. I had our SA's disable the firewall and then things work fine. All the engines start up and connect. With the firewall on, I have to add the line "--enginessh=host" to the controller_args input to enable ssh port forwarding for the connections. When I do that, if I try to launch a single engine on 30 separate computers (with a shared file system), I can only connect 5 of them even though ipcluster log reports that they all connected fine. I'm wondering if there is some timing issue w/ running that many SSH port forward calls (it looks like 3 ports per engine are set up). Any thoughts on what I could try to fix this? Ted ________________________________________ From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] Sent: Wednesday, March 05, 2014 11:06 AM To: IPython developers list Subject: [IPython-dev] IPCluster failing when starting more than a few engines. Using IPython 2.0.0 dev branch sync'ed on 2014-02-24 11:44:52. Running ipcluster start on a set of machines w/o a shared file system using SSHEngineSetLauncher. I have 6 machines that have between 4 and 12 cores on each machine. If I run ipcluster with 2 engines/machine, it works fine. If I increase it to 3 or higher, I start getting engines that fail to connect. Some failures are failures to connect like look like this: 10:43:30.195 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 10:43:35.196 [IPEngineApp] CRITICAL | Registration timed out after 5.0 seconds Other failures are weirder: 10:43:30.184 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 10:43:30.249 [IPEngineApp] Starting to monitor the heartbeat signal from the hub every 3010 ms. 10:43:30.251 [IPEngineApp] Using existing profile dir: u'.ipython/profile_dev' 10:43:30.252 [IPEngineApp] Completed registration with id 6 10:43:36.273 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). 10:43:39.281 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (2 time(s) in a row). 10:43:42.293 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (3 time(s) in a row). 10:44:36.469 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). 10:44:42.489 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). In the second case, if I connect a client to the controller, there is no engine with ID 6 available even though it seems to be getting some heart beats from the hub. I've tried adding lines like these to my config file and it doesn't help: c.IPClusterStart.delay = 0.5 c.SSHEngineSetLauncher.delay = 0.5 The number of failures increases with the number of engines being started on each machine. Trying to start 12 engines on a single machine is almost a complete failure. Any thoughts on what I should be doing differently? Thanks, Ted _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev From aron at ahmadia.net Thu Mar 6 17:08:36 2014 From: aron at ahmadia.net (Aron Ahmadia) Date: Thu, 6 Mar 2014 17:08:36 -0500 Subject: [IPython-dev] IPCluster failing when starting more than a few engines. In-Reply-To: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99FF2@ap-embx-sp20.RES.AD.JPL> References: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB98961@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F4E@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F73@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99FF2@ap-embx-sp20.RES.AD.JPL> Message-ID: On Thu, Mar 6, 2014 at 5:00 PM, Drain, Theodore R (392P) < theodore.r.drain at jpl.nasa.gov> wrote: > At this point I'm thinking I'm going to have to write my own version of > "ipcluster" that runs the controller, sets up the port forwards, and spawns > the engines. Perhaps if I have more control over how that happens that I > can get a cluster that will reliably start up. > Are you dynamically controlling the number of engines? Have you tried switching over to MPI for job startup? -------------- next part -------------- An HTML attachment was scrubbed... URL: From alimanfoo at googlemail.com Fri Mar 7 09:01:36 2014 From: alimanfoo at googlemail.com (Alistair Miles) Date: Fri, 7 Mar 2014 14:01:36 +0000 Subject: [IPython-dev] Fwd: interrupt/abort parallel jobs In-Reply-To: References: Message-ID: Apologies for cross-posting, I originally sent this to ipython-user but a colleague suggested it would be better posted here. ---------- Forwarded message ---------- From: Alistair Miles Date: Fri, Mar 7, 2014 at 11:36 AM Subject: interrupt/abort parallel jobs To: ipython-user at scipy.org Hi all, I know this has been raised before, but I'm using IPython parallel with Sun Grid Engine and I could really do with the ability to interrupt one or more engines without restarting them. As I understand it this is currently not possible? I have tried using AsyncResult.abort() or Client.abort(), but both of those seem ineffective. The former just blocks and hangs. I had a poke around the source code to try and figure out what these are actually doing, but it wasn't immediately clear. Some clarification on what these functions are expected to do would be great, e.g., do these just abort jobs that are queued but not yet executing? Or should it also somehow interrupt running jobs? Just to give you the background, typically I'm setting up some parallel computation, and I set it running but then realise I made a mistake or something isn't running as expected, so then want to interrupt all the engines to cancel the currently running jobs. Of course I can just qdel all the ipengines and then qsub some more, but this is a pain, especially if I have to rerun some common setup on all the engines and/or lose my place in the SGE queue, or if I only need to interrupt some but not all the engines. Thanks, Alistair -- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health The Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Web: http://purl.org/net/aliman Email: alimanfoo at gmail.com Tel: +44 (0)1865 287721 ***new number*** -- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health The Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Web: http://purl.org/net/aliman Email: alimanfoo at gmail.com Tel: +44 (0)1865 287721 ***new number*** -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Fri Mar 7 12:50:41 2014 From: benjaminrk at gmail.com (MinRK) Date: Fri, 7 Mar 2014 09:50:41 -0800 Subject: [IPython-dev] Fwd: interrupt/abort parallel jobs In-Reply-To: References: Message-ID: On Fri, Mar 7, 2014 at 6:01 AM, Alistair Miles wrote: > Apologies for cross-posting, I originally sent this to ipython-user but a > colleague suggested it would be better posted here. > > > ---------- Forwarded message ---------- > From: Alistair Miles > Date: Fri, Mar 7, 2014 at 11:36 AM > Subject: interrupt/abort parallel jobs > To: ipython-user at scipy.org > > > Hi all, > > I know this has been raised before, but I'm using IPython parallel with > Sun Grid Engine and I could really do with the ability to interrupt one or > more engines without restarting them. As I understand it this is currently > not possible? > Remote interrupts require on an adjacent process to the kernel in order to deliver the signal. I have a plan for how to do this (IPEP 12), and hope to have time to implement it in the next release or two. > > I have tried using AsyncResult.abort() or Client.abort(), but both of > those seem ineffective. The former just blocks and hangs. I had a poke > around the source code to try and figure out what these are actually doing, > but it wasn't immediately clear. Some clarification on what these functions > are expected to do would be great, e.g., do these just abort jobs that are > queued but not yet executing? Or should it also somehow interrupt running > jobs? > Abort sends a message on the control channel to the engine(s), indicating that it should not execute the task when it arrives. It cannot abort a *running* task, because the engine won't process the abort message until after it has finished the task to be aborted. > > Just to give you the background, typically I'm setting up some parallel > computation, and I set it running but then realise I made a mistake or > something isn't running as expected, so then want to interrupt all the > engines to cancel the currently running jobs. Of course I can just qdel all > the ipengines and then qsub some more, but this is a pain, especially if I > have to rerun some common setup on all the engines and/or lose my place in > the SGE queue, or if I only need to interrupt some but not all the engines. > The band-aid in the meantime is to keep track of PIDs of engines, and send signals locally or via SSH: pid_map = rc[:].apply_async(os.getpid).get_dict() host_map = c[:].apply_async(socket.gethostname).get_dict() def interrupt_engine(eid): host = host_map[eid] pid = pid_map[eid] if host == socket.gethostname(): # local os.kill(pid, signal.SIGINT) else: !ssh $host kill -INT $pid -MinRK > > Thanks, > Alistair > -- > Alistair Miles > Head of Epidemiological Informatics > Centre for Genomics and Global Health > The Wellcome Trust Centre for Human Genetics > Roosevelt Drive > Oxford > OX3 7BN > United Kingdom > Web: http://purl.org/net/aliman > Email: alimanfoo at gmail.com > Tel: +44 (0)1865 287721 ***new number*** > > > > -- > Alistair Miles > Head of Epidemiological Informatics > Centre for Genomics and Global Health > The Wellcome Trust Centre for Human Genetics > Roosevelt Drive > Oxford > OX3 7BN > United Kingdom > Web: http://purl.org/net/aliman > Email: alimanfoo at gmail.com > Tel: +44 (0)1865 287721 ***new number*** > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Fri Mar 7 18:50:06 2014 From: benjaminrk at gmail.com (MinRK) Date: Fri, 7 Mar 2014 15:50:06 -0800 Subject: [IPython-dev] IPython 2.0.0-b1 Message-ID: Hi all, We are excited to announce the first beta release of IPython 2.0.0. We would very much appreciate your help ironing it out before the final release in a few weeks. We still have a few bugs and lots of documentation we plan to get to before release, but it should be feature complete. Give it a try: http://archive.ipython.org/testing/2.0.0/ A quick summary of changes Major new features: - New modal UI in the notebook (check out the UI Tour in the Help Menu) - New Interactive Widgets (you can check out the tutorialon nbviewer, but they really have to be tested interactively to see how cool they are. - Filesystem navigation from the dashboard (no more uuid urls) - single Python 2 / Python 3 codebase (no more 2to3) - Some notebook security. We will post a more detailed writeup of this one soon. This is the first release of IPython to try the wheel binary format, which is a much smaller download and faster to install, so if you have any issues with the wheel, do let us know. -MinRK -------------- next part -------------- An HTML attachment was scrubbed... URL: From burkhard at ualberta.ca Fri Mar 7 21:32:50 2014 From: burkhard at ualberta.ca (Burkhard Ritter) Date: Fri, 7 Mar 2014 19:32:50 -0700 Subject: [IPython-dev] IPCluster failing when starting more than a few engines. In-Reply-To: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99FF2@ap-embx-sp20.RES.AD.JPL> References: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB98961@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F4E@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F73@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99FF2@ap-embx-sp20.RES.AD.JPL> Message-ID: If I remember correctly I also had difficulties bringing all engines up reliably and it seemed to be due to timing issues with ipcluster. In the end just wrote my own scripts to start up my engines. My script does something like this: ``` for ((i=0;i<$N;i++)); do nohup nice -n19 ipengine --profile=my_profile --ssh=controller_node --log-to-file & sleep 15 done ``` Most of the time I only have two nodes so I just run these scripts by hand, but it shouldn't be difficult to extend the script and and start all engines on a number of nodes. Burkhard On Thu, Mar 6, 2014 at 3:00 PM, Drain, Theodore R (392P) wrote: > Sorry to keep spamming the list but... > > It appears the problem I'm having is purely timing (or timeout) based. If I run ipengine by hand after the controller comes up, I can connect more than 5 engines (so it's not a resource problem). I then tried hacking hub.py which has a line like this: > > self.registration_timeout = max(5000, 2*self.heartmonitor.period) > > If I change that to 60000 (60 seconds), I can get a few more engines to connect but it's basically guesswork as to how many make it up. And there isn't a config option input for that timeout so that isn't much of a solution even if I could come up w/ a time that worked. > > At this point I'm thinking I'm going to have to write my own version of "ipcluster" that runs the controller, sets up the port forwards, and spawns the engines. Perhaps if I have more control over how that happens that I can get a cluster that will reliably start up. > > ________________________________________ > From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] > Sent: Thursday, March 06, 2014 12:51 PM > To: IPython developers list > Subject: Re: [IPython-dev] IPCluster failing when starting more than a few engines. > > One further bit of information: I'm hitting a hard limit of 5 engines connecting using SSH port forwarding. I can run any number of engines locally and it works fine. Could there be some kind of ZMQ limit or SSH limit? The host machine does spawn a huge number of processes - I count 33 processes created when running ipcluster start with a single remote engine which seems a little excessive. > > ________________________________________ > From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] > Sent: Thursday, March 06, 2014 12:37 PM > To: IPython developers list > Subject: Re: [IPython-dev] IPCluster failing when starting more than a few engines. > > Here's some more information. Hopefully someone can help with this as this problem basically makes IPython parallel unusable. > > I had our SA's disable the firewall and then things work fine. All the engines start up and connect. With the firewall on, I have to add the line "--enginessh=host" to the controller_args input to enable ssh port forwarding for the connections. When I do that, if I try to launch a single engine on 30 separate computers (with a shared file system), I can only connect 5 of them even though ipcluster log reports that they all connected fine. > > I'm wondering if there is some timing issue w/ running that many SSH port forward calls (it looks like 3 ports per engine are set up). > > Any thoughts on what I could try to fix this? > > Ted > > ________________________________________ > From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] > Sent: Wednesday, March 05, 2014 11:06 AM > To: IPython developers list > Subject: [IPython-dev] IPCluster failing when starting more than a few engines. > > Using IPython 2.0.0 dev branch sync'ed on 2014-02-24 11:44:52. Running ipcluster start on a set of machines w/o a shared file system using SSHEngineSetLauncher. I have 6 machines that have between 4 and 12 cores on each machine. If I run ipcluster with 2 engines/machine, it works fine. If I increase it to 3 or higher, I start getting engines that fail to connect. > > Some failures are failures to connect like look like this: > 10:43:30.195 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 > 10:43:35.196 [IPEngineApp] CRITICAL | Registration timed out after 5.0 seconds > > Other failures are weirder: > 10:43:30.184 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 > 10:43:30.249 [IPEngineApp] Starting to monitor the heartbeat signal from the hub every 3010 ms. > 10:43:30.251 [IPEngineApp] Using existing profile dir: u'.ipython/profile_dev' > 10:43:30.252 [IPEngineApp] Completed registration with id 6 > 10:43:36.273 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). > 10:43:39.281 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (2 time(s) in a row). > 10:43:42.293 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (3 time(s) in a row). > 10:44:36.469 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). > 10:44:42.489 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). > > In the second case, if I connect a client to the controller, there is no engine with ID 6 available even though it seems to be getting some heart beats from the hub. > > I've tried adding lines like these to my config file and it doesn't help: > c.IPClusterStart.delay = 0.5 > c.SSHEngineSetLauncher.delay = 0.5 > > The number of failures increases with the number of engines being started on each machine. Trying to start 12 engines on a single machine is almost a complete failure. > > Any thoughts on what I should be doing differently? > > Thanks, > Ted > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From p.f.moore at gmail.com Sat Mar 8 06:20:08 2014 From: p.f.moore at gmail.com (Paul Moore) Date: Sat, 8 Mar 2014 11:20:08 +0000 Subject: [IPython-dev] IPython 2.0.0-b1 In-Reply-To: References: Message-ID: On 7 March 2014 23:50, MinRK wrote: > Give it a try: http://archive.ipython.org/testing/2.0.0/ > > A quick summary of changes [...] > single Python 2 / Python 3 codebase (no more 2to3) I haven't checked the code for details, but is there a reason you have both py2 and py3 wheels, rather than a unified py2.py3 one? Paul From takowl at gmail.com Sat Mar 8 11:40:16 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Sat, 8 Mar 2014 08:40:16 -0800 Subject: [IPython-dev] IPython 2.0.0-b1 In-Reply-To: References: Message-ID: On 8 Mar 2014 03:20, "Paul Moore" wrote: > I haven't checked the code for details, but is there a reason you have > both py2 and py3 wheels, rather than a unified py2.py3 one? Different script names: on Python 2, we install ipython2 etc., and on Python 3, ipython3 etc. That appears to be impossible with a single wheel at present. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Sat Mar 8 12:06:36 2014 From: benjaminrk at gmail.com (Min RK) Date: Sat, 8 Mar 2014 09:06:36 -0800 Subject: [IPython-dev] IPython 2.0.0-b1 In-Reply-To: References: Message-ID: > On Mar 8, 2014, at 3:20, Paul Moore wrote: > >> On 7 March 2014 23:50, MinRK wrote: >> Give it a try: http://archive.ipython.org/testing/2.0.0/ >> >> A quick summary of changes > [...] >> single Python 2 / Python 3 codebase (no more 2to3) > > I haven't checked the code for details, but is there a reason you have > both py2 and py3 wheels, rather than a unified py2.py3 one? Entry point suffixes. We install ipython3 on Python 3, and ipython2 on Python 2. I believe metadata format 2 is supposed to add support for this. -MinRK > > Paul > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From tra at popgen.net Sat Mar 8 16:08:07 2014 From: tra at popgen.net (Tiago Antao) Date: Sat, 8 Mar 2014 21:08:07 +0000 Subject: [IPython-dev] git(hub) best practices In-Reply-To: <5310B4DA.7020107@vcu.edu> References: <20140228122316.3a44e691@lnx> <5310B4DA.7020107@vcu.edu> Message-ID: <20140308210807.08bcce3c@grandao> Hi, [Apologies for the late answer] On Fri, 28 Feb 2014 11:10:02 -0500 Chris Friedline wrote: > I have a template that I use that might help you. Check out: > https://github.com/cfriedline/ipynb_template This works very well, thank you. It cannot be used for nbviewer, but it is quite easy to do a script to copy things elsewhere and execute ipython notebook in batch (ipython -c "%run notebook.ipynb") to render the notebooks for nbviewer. Regards, Tiago From p.f.moore at gmail.com Sat Mar 8 16:44:47 2014 From: p.f.moore at gmail.com (Paul Moore) Date: Sat, 8 Mar 2014 21:44:47 +0000 Subject: [IPython-dev] IPython 2.0.0-b1 In-Reply-To: References: Message-ID: On 8 March 2014 16:40, Thomas Kluyver wrote: > On 8 Mar 2014 03:20, "Paul Moore" wrote: >> I haven't checked the code for details, but is there a reason you have >> both py2 and py3 wheels, rather than a unified py2.py3 one? > > Different script names: on Python 2, we install ipython2 etc., and on Python > 3, ipython3 etc. That appears to be impossible with a single wheel at > present. Thanks, I understand now. (BTW, do you install the plain "ipython" name? If you don't, could you, at least on Windows where AFAIK unversioned names are much more conventional?) Paul From ribonucleico at gmail.com Sat Mar 8 16:49:19 2014 From: ribonucleico at gmail.com (Josh Wasserstein) Date: Sat, 8 Mar 2014 16:49:19 -0500 Subject: [IPython-dev] Opening a URL (link to ipynb file) directly on my running notebook Message-ID: Say that I have a link to an .ipnynb file (e.g. on GitHub) like this one: https://github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/blob/master/Chapter1_Introduction/Chapter1_Introduction.ipynb I would like to open that URL directly on my* currently running notebook*. Can I do that directly from the browser? Thanks, Josh -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Sat Mar 8 18:01:34 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Sat, 8 Mar 2014 15:01:34 -0800 Subject: [IPython-dev] IPython 2.0.0-b1 In-Reply-To: References: Message-ID: On 8 March 2014 13:44, Paul Moore wrote: > On 8 March 2014 16:40, Thomas Kluyver wrote: > > On 8 Mar 2014 03:20, "Paul Moore" wrote: > >> I haven't checked the code for details, but is there a reason you have > >> both py2 and py3 wheels, rather than a unified py2.py3 one? > > > > Different script names: on Python 2, we install ipython2 etc., and on > Python > > 3, ipython3 etc. That appears to be impossible with a single wheel at > > present. > > Thanks, I understand now. (BTW, do you install the plain "ipython" > name? If you don't, could you, at least on Windows where AFAIK > unversioned names are much more conventional?) Yes, both versions also install 'ipython'. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Sat Mar 8 18:03:13 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Sat, 8 Mar 2014 15:03:13 -0800 Subject: [IPython-dev] Opening a URL (link to ipynb file) directly on my running notebook In-Reply-To: References: Message-ID: On 8 March 2014 13:49, Josh Wasserstein wrote: > Say that I have a link to an .ipnynb file (e.g. on GitHub) like this one: > > https://github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/blob/master/Chapter1_Introduction/Chapter1_Introduction.ipynb > > I would like to open that URL directly on my* currently running notebook*. > Can I do that directly from the browser? > At present, no: you have to download it and (if necessary) move it to a directory accessible to the notebook. I think opening directly from a URL is something we want to enable when we refactor the dashboard, though. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From theodore.r.drain at jpl.nasa.gov Sat Mar 8 22:41:32 2014 From: theodore.r.drain at jpl.nasa.gov (Drain, Theodore R (392P)) Date: Sun, 9 Mar 2014 03:41:32 +0000 Subject: [IPython-dev] IPCluster failing when starting more than a few engines. In-Reply-To: References: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB98961@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F4E@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99F73@ap-embx-sp20.RES.AD.JPL> <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB99FF2@ap-embx-sp20.RES.AD.JPL>, Message-ID: <0DC1CAB7F6C7FC4A8B54EE1FD49046990FB9A313@ap-embx-sp20.RES.AD.JPL> Thanks Burkhard. I'm going to be launching about 50-200 engines each on a remote machine so waiting a really long time between launches isn't going to be very practical. After beating my head against the wall for a lot of hours on this, I've finally gotten what I think is a robust system for starting up the cluster using SSH. There seems to be two issues: 1) if I launch too many engines at once, some of them with fail either with a timeout or a controller purged request error and 2) if an engine launches on a slow or overloaded machine, it may not finish the registration process with the controller in time and be purged. 2) happens even if I launch a single engine at a time and there is no easy way to fix it with the current inputs that I can see. The first part of the fix is a modification to the HubFactory and Hub classes. I changed the registration_timeout field which defaults to max(10 sec, 5*heartbeat) into a configuration file input on the HubFactory. That lets me set it in the profile to a much larger value (90 sec in my case) which then allows slow engines more time to connect without having to make really long heartbeats. The second part of the script is different launcher logic. I wrote a script that launches a few engines (4 seems to work well for me) 1 second apart. Then it creates a Client and waits for all the engines to connect. Then it launches more engines and repeats until all of the engines have launched and connected. I've submitted an issue https://github.com/ipython/ipython/issues/5302 and the link showing my changes to hub.py which allowed me to get this working. If MinRK thinks it would be useful, I might try and modify the SSHEngineSetLauncher to have this kind of logic. It could be a single integer input to indicate current behavior (input=0) or a new behavior which is how many engines to launch before waiting for them to connect (input>0). Ted ________________________________________ From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Burkhard Ritter [burkhard at ualberta.ca] Sent: Friday, March 07, 2014 6:32 PM To: IPython developers list Subject: Re: [IPython-dev] IPCluster failing when starting more than a few engines. If I remember correctly I also had difficulties bringing all engines up reliably and it seemed to be due to timing issues with ipcluster. In the end just wrote my own scripts to start up my engines. My script does something like this: ``` for ((i=0;i<$N;i++)); do nohup nice -n19 ipengine --profile=my_profile --ssh=controller_node --log-to-file & sleep 15 done ``` Most of the time I only have two nodes so I just run these scripts by hand, but it shouldn't be difficult to extend the script and and start all engines on a number of nodes. Burkhard On Thu, Mar 6, 2014 at 3:00 PM, Drain, Theodore R (392P) wrote: > Sorry to keep spamming the list but... > > It appears the problem I'm having is purely timing (or timeout) based. If I run ipengine by hand after the controller comes up, I can connect more than 5 engines (so it's not a resource problem). I then tried hacking hub.py which has a line like this: > > self.registration_timeout = max(5000, 2*self.heartmonitor.period) > > If I change that to 60000 (60 seconds), I can get a few more engines to connect but it's basically guesswork as to how many make it up. And there isn't a config option input for that timeout so that isn't much of a solution even if I could come up w/ a time that worked. > > At this point I'm thinking I'm going to have to write my own version of "ipcluster" that runs the controller, sets up the port forwards, and spawns the engines. Perhaps if I have more control over how that happens that I can get a cluster that will reliably start up. > > ________________________________________ > From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] > Sent: Thursday, March 06, 2014 12:51 PM > To: IPython developers list > Subject: Re: [IPython-dev] IPCluster failing when starting more than a few engines. > > One further bit of information: I'm hitting a hard limit of 5 engines connecting using SSH port forwarding. I can run any number of engines locally and it works fine. Could there be some kind of ZMQ limit or SSH limit? The host machine does spawn a huge number of processes - I count 33 processes created when running ipcluster start with a single remote engine which seems a little excessive. > > ________________________________________ > From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] > Sent: Thursday, March 06, 2014 12:37 PM > To: IPython developers list > Subject: Re: [IPython-dev] IPCluster failing when starting more than a few engines. > > Here's some more information. Hopefully someone can help with this as this problem basically makes IPython parallel unusable. > > I had our SA's disable the firewall and then things work fine. All the engines start up and connect. With the firewall on, I have to add the line "--enginessh=host" to the controller_args input to enable ssh port forwarding for the connections. When I do that, if I try to launch a single engine on 30 separate computers (with a shared file system), I can only connect 5 of them even though ipcluster log reports that they all connected fine. > > I'm wondering if there is some timing issue w/ running that many SSH port forward calls (it looks like 3 ports per engine are set up). > > Any thoughts on what I could try to fix this? > > Ted > > ________________________________________ > From: ipython-dev-bounces at scipy.org [ipython-dev-bounces at scipy.org] on behalf of Drain, Theodore R (392P) [theodore.r.drain at jpl.nasa.gov] > Sent: Wednesday, March 05, 2014 11:06 AM > To: IPython developers list > Subject: [IPython-dev] IPCluster failing when starting more than a few engines. > > Using IPython 2.0.0 dev branch sync'ed on 2014-02-24 11:44:52. Running ipcluster start on a set of machines w/o a shared file system using SSHEngineSetLauncher. I have 6 machines that have between 4 and 12 cores on each machine. If I run ipcluster with 2 engines/machine, it works fine. If I increase it to 3 or higher, I start getting engines that fail to connect. > > Some failures are failures to connect like look like this: > 10:43:30.195 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 > 10:43:35.196 [IPEngineApp] CRITICAL | Registration timed out after 5.0 seconds > > Other failures are weirder: > 10:43:30.184 [IPEngineApp] Registering with controller at tcp://x.x.x.x:59987 > 10:43:30.249 [IPEngineApp] Starting to monitor the heartbeat signal from the hub every 3010 ms. > 10:43:30.251 [IPEngineApp] Using existing profile dir: u'.ipython/profile_dev' > 10:43:30.252 [IPEngineApp] Completed registration with id 6 > 10:43:36.273 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). > 10:43:39.281 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (2 time(s) in a row). > 10:43:42.293 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (3 time(s) in a row). > 10:44:36.469 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). > 10:44:42.489 [IPEngineApp] WARNING | No heartbeat in the last 3010 ms (1 time(s) in a row). > > In the second case, if I connect a client to the controller, there is no engine with ID 6 available even though it seems to be getting some heart beats from the hub. > > I've tried adding lines like these to my config file and it doesn't help: > c.IPClusterStart.delay = 0.5 > c.SSHEngineSetLauncher.delay = 0.5 > > The number of failures increases with the number of engines being started on each machine. Trying to start 12 engines on a single machine is almost a complete failure. > > Any thoughts on what I should be doing differently? > > Thanks, > Ted > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev From jakevdp at cs.washington.edu Sun Mar 9 12:15:55 2014 From: jakevdp at cs.washington.edu (Jacob Vanderplas) Date: Sun, 9 Mar 2014 09:15:55 -0700 Subject: [IPython-dev] nbconvert changes in 2.0 Message-ID: Hi all, Now that a 2.0 candidate has been released, I'm trying to figure out how to update my pelican notebook plugin to work with the refactor of nbconvert. I wonder if some of the devs might give me some insight into what's changed (the relevant code is here: https://github.com/getpelican/pelican-plugins/blob/master/liquid_tags/notebook.py ) It basically uses the HTMLExporter, a custom Transformer, and a custom pygments highlighter, many of which no longer exist in IPython 2.0. Any help would be appreciated - thanks! Jake Jake Vanderplas Director of Research - Physical Sciences eScience Institute, University of Washington http://www.vanderplas.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From jon.freder at gmail.com Sun Mar 9 13:28:47 2014 From: jon.freder at gmail.com (Jonathan Frederic) Date: Sun, 9 Mar 2014 10:28:47 -0700 Subject: [IPython-dev] nbconvert changes in 2.0 In-Reply-To: References: Message-ID: Hi Jake, I took a quick look at the code you linked- s/_pygment_highlight/_pygments_highlight s/IPython.nbconvert.transformers/IPython.nbconvert.preprocessors s/Transformer/Preprocessor There may be a couple more, but that should be a good starting point. -Jon On Sun, Mar 9, 2014 at 9:15 AM, Jacob Vanderplas wrote: > Hi all, > Now that a 2.0 candidate has been released, I'm trying to figure out how > to update my pelican notebook plugin to work with the refactor of nbconvert. > > I wonder if some of the devs might give me some insight into what's > changed (the relevant code is here: > https://github.com/getpelican/pelican-plugins/blob/master/liquid_tags/notebook.py > ) > > It basically uses the HTMLExporter, a custom Transformer, and a custom > pygments highlighter, many of which no longer exist in IPython 2.0. > > Any help would be appreciated - thanks! > Jake > > Jake Vanderplas > Director of Research - Physical Sciences > eScience Institute, University of Washington > http://www.vanderplas.com > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Sun Mar 9 14:02:03 2014 From: benjaminrk at gmail.com (MinRK) Date: Sun, 9 Mar 2014 11:02:03 -0700 Subject: [IPython-dev] nbconvert changes in 2.0 In-Reply-To: References: Message-ID: This should do it (or get close, anyway): https://github.com/getpelican/pelican-plugins/pull/175 -MinRK On Sun, Mar 9, 2014 at 10:28 AM, Jonathan Frederic wrote: > Hi Jake, > > I took a quick look at the code you linked- > > s/_pygment_highlight/_pygments_highlight > s/IPython.nbconvert.transformers/IPython.nbconvert.preprocessors > s/Transformer/Preprocessor > > > There may be a couple more, but that should be a good starting point. > > -Jon > > > On Sun, Mar 9, 2014 at 9:15 AM, Jacob Vanderplas < > jakevdp at cs.washington.edu> wrote: > >> Hi all, >> Now that a 2.0 candidate has been released, I'm trying to figure out how >> to update my pelican notebook plugin to work with the refactor of nbconvert. >> >> I wonder if some of the devs might give me some insight into what's >> changed (the relevant code is here: >> https://github.com/getpelican/pelican-plugins/blob/master/liquid_tags/notebook.py >> ) >> >> It basically uses the HTMLExporter, a custom Transformer, and a custom >> pygments highlighter, many of which no longer exist in IPython 2.0. >> >> Any help would be appreciated - thanks! >> Jake >> >> Jake Vanderplas >> Director of Research ? Physical Sciences >> eScience Institute, University of Washington >> http://www.vanderplas.com >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jakevdp at cs.washington.edu Sun Mar 9 14:04:02 2014 From: jakevdp at cs.washington.edu (Jacob Vanderplas) Date: Sun, 9 Mar 2014 11:04:02 -0700 Subject: [IPython-dev] nbconvert changes in 2.0 In-Reply-To: References: Message-ID: Hi John, Thanks for the response. I started by making those substitutions, but they resulted in a host of difficult to understand errors, including some obscure ones from within the pygments module. It seems to be a bit more complicated than this! Thanks for the help, Jake Jake Vanderplas Director of Research - Physical Sciences eScience Institute, University of Washington http://www.vanderplas.com On Sun, Mar 9, 2014 at 10:28 AM, Jonathan Frederic wrote: > Hi Jake, > > I took a quick look at the code you linked- > > s/_pygment_highlight/_pygments_highlight > s/IPython.nbconvert.transformers/IPython.nbconvert.preprocessors > s/Transformer/Preprocessor > > > There may be a couple more, but that should be a good starting point. > > -Jon > > > On Sun, Mar 9, 2014 at 9:15 AM, Jacob Vanderplas < > jakevdp at cs.washington.edu> wrote: > >> Hi all, >> Now that a 2.0 candidate has been released, I'm trying to figure out how >> to update my pelican notebook plugin to work with the refactor of nbconvert. >> >> I wonder if some of the devs might give me some insight into what's >> changed (the relevant code is here: >> https://github.com/getpelican/pelican-plugins/blob/master/liquid_tags/notebook.py >> ) >> >> It basically uses the HTMLExporter, a custom Transformer, and a custom >> pygments highlighter, many of which no longer exist in IPython 2.0. >> >> Any help would be appreciated - thanks! >> Jake >> >> Jake Vanderplas >> Director of Research - Physical Sciences >> eScience Institute, University of Washington >> http://www.vanderplas.com >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jakevdp at cs.washington.edu Sun Mar 9 14:05:13 2014 From: jakevdp at cs.washington.edu (Jacob Vanderplas) Date: Sun, 9 Mar 2014 11:05:13 -0700 Subject: [IPython-dev] nbconvert changes in 2.0 In-Reply-To: References: Message-ID: Thanks Min - much appreciated! I'll give that a try. Jake Jake Vanderplas Director of Research - Physical Sciences eScience Institute, University of Washington http://www.vanderplas.com On Sun, Mar 9, 2014 at 11:02 AM, MinRK wrote: > This should do it (or get close, anyway): > https://github.com/getpelican/pelican-plugins/pull/175 > > -MinRK > > > On Sun, Mar 9, 2014 at 10:28 AM, Jonathan Frederic wrote: > >> Hi Jake, >> >> I took a quick look at the code you linked- >> >> s/_pygment_highlight/_pygments_highlight >> s/IPython.nbconvert.transformers/IPython.nbconvert.preprocessors >> s/Transformer/Preprocessor >> >> >> There may be a couple more, but that should be a good starting point. >> >> -Jon >> >> >> On Sun, Mar 9, 2014 at 9:15 AM, Jacob Vanderplas < >> jakevdp at cs.washington.edu> wrote: >> >>> Hi all, >>> Now that a 2.0 candidate has been released, I'm trying to figure out how >>> to update my pelican notebook plugin to work with the refactor of nbconvert. >>> >>> I wonder if some of the devs might give me some insight into what's >>> changed (the relevant code is here: >>> https://github.com/getpelican/pelican-plugins/blob/master/liquid_tags/notebook.py >>> ) >>> >>> It basically uses the HTMLExporter, a custom Transformer, and a custom >>> pygments highlighter, many of which no longer exist in IPython 2.0. >>> >>> Any help would be appreciated - thanks! >>> Jake >>> >>> Jake Vanderplas >>> Director of Research - Physical Sciences >>> eScience Institute, University of Washington >>> http://www.vanderplas.com >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >>> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From patrick.surry at gmail.com Mon Mar 10 10:37:45 2014 From: patrick.surry at gmail.com (Patrick Surry) Date: Mon, 10 Mar 2014 10:37:45 -0400 Subject: [IPython-dev] ERROR: global name 'select_figure_formats' is not defined Message-ID: I'm getting this message in master doing: %matplotlib inline %config InlineBackend.figure_format = 'svg' but matplotlib still seems to work OK. Has the syntax changed or something? Thanks, Patrick -------------- next part -------------- An HTML attachment was scrubbed... URL: From zvoros at gmail.com Mon Mar 10 10:52:35 2014 From: zvoros at gmail.com (=?ISO-8859-1?Q?Zolt=E1n_V=F6r=F6s?=) Date: Mon, 10 Mar 2014 15:52:35 +0100 Subject: [IPython-dev] ERROR: global name 'select_figure_formats' is not defined In-Reply-To: References: Message-ID: <531DD1B3.1080703@gmail.com> I see the same error on the latest from master (commit ID: Updating 1cb67f9..5d81e8d), however, I don't have any problems with %pylab inline %config InlineBackend.figure_format = 'svg' Cheers, Zolt?n On 10/03/14 15:37, Patrick Surry wrote: > I'm getting this message in master doing: > > %matplotlib inline > %config InlineBackend.figure_format = 'svg' > > but matplotlib still seems to work OK. Has the syntax changed or > something? > > Thanks, > Patrick > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: From ribonucleico at gmail.com Mon Mar 10 11:28:32 2014 From: ribonucleico at gmail.com (Josh Wasserstein) Date: Mon, 10 Mar 2014 11:28:32 -0400 Subject: [IPython-dev] Change logs? Message-ID: I just upgraded to IPython 1.2.1 through conda. Where can I find the changelog for this release? When I go to http://ipython.org/news.html I can see the release notes for older versions, but not for 1.2.1. Thanks, Josh -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Mon Mar 10 12:54:15 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Mon, 10 Mar 2014 09:54:15 -0700 Subject: [IPython-dev] Change logs? In-Reply-To: References: Message-ID: Hi Josh, You can see the list of PRs merged and issues closed here: http://ipython.org/ipython-doc/stable/whatsnew/github-stats-1.0.html Thomas On 10 March 2014 08:28, Josh Wasserstein wrote: > I just upgraded to IPython 1.2.1 through conda. Where can I find the > changelog for this release? When I go to http://ipython.org/news.html I > can see the release notes for older versions, but not for 1.2.1. > > Thanks, > > Josh > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Mon Mar 10 13:01:39 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Mon, 10 Mar 2014 10:01:39 -0700 Subject: [IPython-dev] ERROR: global name 'select_figure_formats' is not defined In-Reply-To: <531DD1B3.1080703@gmail.com> References: <531DD1B3.1080703@gmail.com> Message-ID: Try this: https://github.com/ipython/ipython/pull/5318 On 10 March 2014 07:52, Zolt?n V?r?s wrote: > I see the same error on the latest from master (commit ID: Updating > 1cb67f9..5d81e8d), however, I don't have any problems with > > %pylab inline > %config InlineBackend.figure_format = 'svg' > > Cheers, > Zolt?n > > > On 10/03/14 15:37, Patrick Surry wrote: > > I'm getting this message in master doing: > > %matplotlib inline > %config InlineBackend.figure_format = 'svg' > > but matplotlib still seems to work OK. Has the syntax changed or > something? > > Thanks, > Patrick > > > _______________________________________________ > IPython-dev mailing listIPython-dev at scipy.orghttp://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From zvoros at gmail.com Mon Mar 10 14:17:46 2014 From: zvoros at gmail.com (=?ISO-8859-1?Q?Zolt=E1n_V=F6r=F6s?=) Date: Mon, 10 Mar 2014 19:17:46 +0100 Subject: [IPython-dev] ERROR: global name 'select_figure_formats' is not defined In-Reply-To: References: <531DD1B3.1080703@gmail.com> Message-ID: <531E01CA.5010701@gmail.com> Hi Thomas, Many thanks for the prompt fix! Cheers, Zolt?n On 10/03/14 18:01, Thomas Kluyver wrote: > Try this: > https://github.com/ipython/ipython/pull/5318 > > > On 10 March 2014 07:52, Zolt?n V?r?s > wrote: > > I see the same error on the latest from master (commit ID: > Updating 1cb67f9..5d81e8d), however, I don't have any problems with > > %pylab inline > %config InlineBackend.figure_format = 'svg' > > Cheers, > Zolt?n > > > On 10/03/14 15:37, Patrick Surry wrote: >> I'm getting this message in master doing: >> >> %matplotlib inline >> %config InlineBackend.figure_format = 'svg' >> >> but matplotlib still seems to work OK. Has the syntax changed or >> something? >> >> Thanks, >> Patrick >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Mon Mar 10 14:37:28 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Mon, 10 Mar 2014 11:37:28 -0700 Subject: [IPython-dev] Results of last year's user survey Message-ID: Hi all, You may remember filling in a survey of IPython users last year. I finally got round to writing up the results from that, and you can now view them on the website: http://ipython.org/usersurvey2013.html Significant findings include: - A rapid changeover to new IPython versions - More people report using the notebook than the IPython terminal interface - 22% have used IPython on Python 3.3. - Nearly 60 people reported using IPython with the Spyder IDE - People install IPython from a variety of sources, but no one installation method was used by more than half of respondents. I read through all of the suggestions and comments written in the survey, and forwarded some groups of them to people with specific expertise. The write up includes some notes on requests that I saw multiple times. Finally, we're grateful to the over 100 people who used the comments and suggestions field to thank or compliment us. :-) Thanks, Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From ellisonbg at gmail.com Mon Mar 10 15:30:21 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Mon, 10 Mar 2014 12:30:21 -0700 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: Thanks Thomas!! this is really helpful. Cheers, Brian On Mon, Mar 10, 2014 at 11:37 AM, Thomas Kluyver wrote: > Hi all, > > You may remember filling in a survey of IPython users last year. I finally > got round to writing up the results from that, and you can now view them on > the website: > > http://ipython.org/usersurvey2013.html > > Significant findings include: > - A rapid changeover to new IPython versions > - More people report using the notebook than the IPython terminal interface > - 22% have used IPython on Python 3.3. > - Nearly 60 people reported using IPython with the Spyder IDE > - People install IPython from a variety of sources, but no one installation > method was used by more than half of respondents. > > I read through all of the suggestions and comments written in the survey, > and forwarded some groups of them to people with specific expertise. The > write up includes some notes on requests that I saw multiple times. > > Finally, we're grateful to the over 100 people who used the comments and > suggestions field to thank or compliment us. :-) > > Thanks, > Thomas > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From asmeurer at gmail.com Mon Mar 10 16:02:40 2014 From: asmeurer at gmail.com (Aaron Meurer) Date: Mon, 10 Mar 2014 15:02:40 -0500 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: I didn't know you guys were dropping 2.6 support so soon. Regarding hipchat, I hate to say it, but nobody uses it because it sucks. I personally hate it every time I try to use it. The interface is just the worst for people who are not registered. You guys should take a look at gitter.im. It integrates with GitHub, it's free, and it has nice things like backlogs on chats for everyone. I'm a little surprised about the academia vs. industry usage. I suspect there were just more respondents from the academic side than from industry. Aaron Meurer On Mon, Mar 10, 2014 at 2:30 PM, Brian Granger wrote: > Thanks Thomas!! this is really helpful. > > Cheers, > > Brian > > On Mon, Mar 10, 2014 at 11:37 AM, Thomas Kluyver wrote: >> Hi all, >> >> You may remember filling in a survey of IPython users last year. I finally >> got round to writing up the results from that, and you can now view them on >> the website: >> >> http://ipython.org/usersurvey2013.html >> >> Significant findings include: >> - A rapid changeover to new IPython versions >> - More people report using the notebook than the IPython terminal interface >> - 22% have used IPython on Python 3.3. >> - Nearly 60 people reported using IPython with the Spyder IDE >> - People install IPython from a variety of sources, but no one installation >> method was used by more than half of respondents. >> >> I read through all of the suggestions and comments written in the survey, >> and forwarded some groups of them to people with specific expertise. The >> write up includes some notes on requests that I saw multiple times. >> >> Finally, we're grateful to the over 100 people who used the comments and >> suggestions field to thank or compliment us. :-) >> >> Thanks, >> Thomas >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > > > > -- > Brian E. Granger > Cal Poly State University, San Luis Obispo > bgranger at calpoly.edu and ellisonbg at gmail.com > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From asmeurer at gmail.com Mon Mar 10 16:03:26 2014 From: asmeurer at gmail.com (Aaron Meurer) Date: Mon, 10 Mar 2014 15:03:26 -0500 Subject: [IPython-dev] IPython 2.0.0-b1 In-Reply-To: References: Message-ID: So now we will have IPython 2 and the upcoming IPython 3, which have nothing to do with ipython2 and ipython3. Aaron Meurer On Sat, Mar 8, 2014 at 5:01 PM, Thomas Kluyver wrote: > On 8 March 2014 13:44, Paul Moore wrote: >> >> On 8 March 2014 16:40, Thomas Kluyver wrote: >> > On 8 Mar 2014 03:20, "Paul Moore" wrote: >> >> I haven't checked the code for details, but is there a reason you have >> >> both py2 and py3 wheels, rather than a unified py2.py3 one? >> > >> > Different script names: on Python 2, we install ipython2 etc., and on >> > Python >> > 3, ipython3 etc. That appears to be impossible with a single wheel at >> > present. >> >> Thanks, I understand now. (BTW, do you install the plain "ipython" >> name? If you don't, could you, at least on Windows where AFAIK >> unversioned names are much more conventional?) > > > Yes, both versions also install 'ipython'. > > Thomas > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From ellisonbg at gmail.com Mon Mar 10 16:11:30 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Mon, 10 Mar 2014 13:11:30 -0700 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: Does gitter support private rooms and private direct messaging. We use those things on a daily basis and I wouldn't consider moving to something w/o it. On Mon, Mar 10, 2014 at 1:02 PM, Aaron Meurer wrote: > I didn't know you guys were dropping 2.6 support so soon. > > Regarding hipchat, I hate to say it, but nobody uses it because it > sucks. I personally hate it every time I try to use it. The interface > is just the worst for people who are not registered. You guys should > take a look at gitter.im. It integrates with GitHub, it's free, and it > has nice things like backlogs on chats for everyone. > > I'm a little surprised about the academia vs. industry usage. I > suspect there were just more respondents from the academic side than > from industry. > > Aaron Meurer > > On Mon, Mar 10, 2014 at 2:30 PM, Brian Granger wrote: >> Thanks Thomas!! this is really helpful. >> >> Cheers, >> >> Brian >> >> On Mon, Mar 10, 2014 at 11:37 AM, Thomas Kluyver wrote: >>> Hi all, >>> >>> You may remember filling in a survey of IPython users last year. I finally >>> got round to writing up the results from that, and you can now view them on >>> the website: >>> >>> http://ipython.org/usersurvey2013.html >>> >>> Significant findings include: >>> - A rapid changeover to new IPython versions >>> - More people report using the notebook than the IPython terminal interface >>> - 22% have used IPython on Python 3.3. >>> - Nearly 60 people reported using IPython with the Spyder IDE >>> - People install IPython from a variety of sources, but no one installation >>> method was used by more than half of respondents. >>> >>> I read through all of the suggestions and comments written in the survey, >>> and forwarded some groups of them to people with specific expertise. The >>> write up includes some notes on requests that I saw multiple times. >>> >>> Finally, we're grateful to the over 100 people who used the comments and >>> suggestions field to thank or compliment us. :-) >>> >>> Thanks, >>> Thomas >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >> >> >> >> -- >> Brian E. Granger >> Cal Poly State University, San Luis Obispo >> bgranger at calpoly.edu and ellisonbg at gmail.com >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From takowl at gmail.com Mon Mar 10 16:49:44 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Mon, 10 Mar 2014 13:49:44 -0700 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: On 10 March 2014 13:02, Aaron Meurer wrote: > I didn't know you guys were dropping 2.6 support so soon. > Yep. We've been fairly aggressive with Python version requirements for a while - we were one of the first projects to drop 2.5 support as well. We anticipate we'll get some bug reports about it, but we think it's worth it to be able to use 2.7 features reliably. It's so easy to forget what 2.6 didn't have, which is why we had to release 1.2.1 soon after 1.2.0. > Regarding hipchat, I hate to say it, but nobody uses it because it > sucks. I personally hate it every time I try to use it. The interface > is just the worst for people who are not registered. You guys should > take a look at gitter.im. It integrates with GitHub, it's free, and it > has nice things like backlogs on chats for everyone. > We do actually see a couple of questions on Hipchat most days - probably more traffic than the mailing list - but I don't think most of the people using the help chat room would have heard about the survey. Gitter does look interesting, though. Brian: I think Gitter has private rooms if you associate it with private repos. People who can see the repo see the chat room. I don't know about direct messaging. > I'm a little surprised about the academia vs. industry usage. I > suspect there were just more respondents from the academic side than > from industry. > That's quite possible. The results are probably also biased towards English speakers, early adopters, and technically proficient users. Getting a representative sample is pretty tricky. But I think the results give us some ballpark figures. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Mon Mar 10 17:20:55 2014 From: benjaminrk at gmail.com (MinRK) Date: Mon, 10 Mar 2014 14:20:55 -0700 Subject: [IPython-dev] IPython 2.0.0-b1 In-Reply-To: References: Message-ID: naturally. On Mon, Mar 10, 2014 at 1:03 PM, Aaron Meurer wrote: > So now we will have IPython 2 and the upcoming IPython 3, which have > nothing to do with ipython2 and ipython3. > > Aaron Meurer > > On Sat, Mar 8, 2014 at 5:01 PM, Thomas Kluyver wrote: > > On 8 March 2014 13:44, Paul Moore wrote: > >> > >> On 8 March 2014 16:40, Thomas Kluyver wrote: > >> > On 8 Mar 2014 03:20, "Paul Moore" wrote: > >> >> I haven't checked the code for details, but is there a reason you > have > >> >> both py2 and py3 wheels, rather than a unified py2.py3 one? > >> > > >> > Different script names: on Python 2, we install ipython2 etc., and on > >> > Python > >> > 3, ipython3 etc. That appears to be impossible with a single wheel at > >> > present. > >> > >> Thanks, I understand now. (BTW, do you install the plain "ipython" > >> name? If you don't, could you, at least on Windows where AFAIK > >> unversioned names are much more conventional?) > > > > > > Yes, both versions also install 'ipython'. > > > > Thomas > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From asmeurer at gmail.com Mon Mar 10 20:45:11 2014 From: asmeurer at gmail.com (Aaron Meurer) Date: Mon, 10 Mar 2014 19:45:11 -0500 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: Gitter does support direct messages. I don't know if it supports private rooms. Aaron Meurer On Mon, Mar 10, 2014 at 3:49 PM, Thomas Kluyver wrote: > On 10 March 2014 13:02, Aaron Meurer wrote: >> >> I didn't know you guys were dropping 2.6 support so soon. > > > Yep. We've been fairly aggressive with Python version requirements for a > while - we were one of the first projects to drop 2.5 support as well. We > anticipate we'll get some bug reports about it, but we think it's worth it > to be able to use 2.7 features reliably. It's so easy to forget what 2.6 > didn't have, which is why we had to release 1.2.1 soon after 1.2.0. > >> >> Regarding hipchat, I hate to say it, but nobody uses it because it >> sucks. I personally hate it every time I try to use it. The interface >> is just the worst for people who are not registered. You guys should >> take a look at gitter.im. It integrates with GitHub, it's free, and it >> has nice things like backlogs on chats for everyone. > > > We do actually see a couple of questions on Hipchat most days - probably > more traffic than the mailing list - but I don't think most of the people > using the help chat room would have heard about the survey. Gitter does look > interesting, though. > > Brian: I think Gitter has private rooms if you associate it with private > repos. People who can see the repo see the chat room. I don't know about > direct messaging. > >> >> I'm a little surprised about the academia vs. industry usage. I >> suspect there were just more respondents from the academic side than >> from industry. > > > That's quite possible. The results are probably also biased towards English > speakers, early adopters, and technically proficient users. Getting a > representative sample is pretty tricky. But I think the results give us some > ballpark figures. > > Thomas > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From benjaminrk at gmail.com Mon Mar 10 21:08:03 2014 From: benjaminrk at gmail.com (MinRK) Date: Mon, 10 Mar 2014 18:08:03 -0700 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: There are chat rooms for organizations and repos - only members of an organization can join the org room - public repo rooms are public - private repo rooms are private I think there are two things that we use on HipChat that gitter doesn't do: - guest access (GitHub login is required even for public rooms) - push/email notifications for mentions when idle (it supports desktop, but not offline notifications) For the most part, I think Gitter seems a lot better than HipChat, especially for guests (at least those with GitHub accounts). I am not especially fond of HipChat, and the guest experience is pretty horrendous (no native app, no history, frequent crashes, etc.), so I wouldn't mind giving something else a try. -MinRK On Mon, Mar 10, 2014 at 5:45 PM, Aaron Meurer wrote: > Gitter does support direct messages. I don't know if it supports private > rooms. > > Aaron Meurer > > On Mon, Mar 10, 2014 at 3:49 PM, Thomas Kluyver wrote: > > On 10 March 2014 13:02, Aaron Meurer wrote: > >> > >> I didn't know you guys were dropping 2.6 support so soon. > > > > > > Yep. We've been fairly aggressive with Python version requirements for a > > while - we were one of the first projects to drop 2.5 support as well. We > > anticipate we'll get some bug reports about it, but we think it's worth > it > > to be able to use 2.7 features reliably. It's so easy to forget what 2.6 > > didn't have, which is why we had to release 1.2.1 soon after 1.2.0. > > > >> > >> Regarding hipchat, I hate to say it, but nobody uses it because it > >> sucks. I personally hate it every time I try to use it. The interface > >> is just the worst for people who are not registered. You guys should > >> take a look at gitter.im. It integrates with GitHub, it's free, and it > >> has nice things like backlogs on chats for everyone. > > > > > > We do actually see a couple of questions on Hipchat most days - probably > > more traffic than the mailing list - but I don't think most of the people > > using the help chat room would have heard about the survey. Gitter does > look > > interesting, though. > > > > Brian: I think Gitter has private rooms if you associate it with private > > repos. People who can see the repo see the chat room. I don't know about > > direct messaging. > > > >> > >> I'm a little surprised about the academia vs. industry usage. I > >> suspect there were just more respondents from the academic side than > >> from industry. > > > > > > That's quite possible. The results are probably also biased towards > English > > speakers, early adopters, and technically proficient users. Getting a > > representative sample is pretty tricky. But I think the results give us > some > > ballpark figures. > > > > Thomas > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From asmeurer at gmail.com Mon Mar 10 21:39:34 2014 From: asmeurer at gmail.com (Aaron Meurer) Date: Mon, 10 Mar 2014 20:39:34 -0500 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: The guest access of HipChat is really what sucks to me. There is no chat history, so accidentally reloading the page causes you to lose all context. Gitter does email you when a starred room has activity. I think they also email you if you are @mentioned (not positive about that last part, but it would be silly if they didn't). According to https://gitter.im/apps phone apps are coming soon. I've been using it since the public beta started in December and I've been really impressed with how fast it has improved. The only concern with gitter is that they are currently in beta, so it's not clear what will remain free when they leave it. But seeing as you guys are already paying for HipChat, it shouldn't be as big of a concern for you. Aaron Meurer On Mon, Mar 10, 2014 at 8:08 PM, MinRK wrote: > There are chat rooms for organizations and repos > > - only members of an organization can join the org room > - public repo rooms are public > - private repo rooms are private > > I think there are two things that we use on HipChat that gitter doesn't do: > > - guest access (GitHub login is required even for public rooms) > - push/email notifications for mentions when idle (it supports desktop, but > not offline notifications) > > For the most part, I think Gitter seems a lot better than HipChat, > especially for guests (at least those with GitHub accounts). I am not > especially fond of HipChat, and the guest experience is pretty horrendous > (no native app, no history, frequent crashes, etc.), so I wouldn't mind > giving something else a try. > > -MinRK > > > > On Mon, Mar 10, 2014 at 5:45 PM, Aaron Meurer wrote: >> >> Gitter does support direct messages. I don't know if it supports private >> rooms. >> >> Aaron Meurer >> >> On Mon, Mar 10, 2014 at 3:49 PM, Thomas Kluyver wrote: >> > On 10 March 2014 13:02, Aaron Meurer wrote: >> >> >> >> I didn't know you guys were dropping 2.6 support so soon. >> > >> > >> > Yep. We've been fairly aggressive with Python version requirements for a >> > while - we were one of the first projects to drop 2.5 support as well. >> > We >> > anticipate we'll get some bug reports about it, but we think it's worth >> > it >> > to be able to use 2.7 features reliably. It's so easy to forget what 2.6 >> > didn't have, which is why we had to release 1.2.1 soon after 1.2.0. >> > >> >> >> >> Regarding hipchat, I hate to say it, but nobody uses it because it >> >> sucks. I personally hate it every time I try to use it. The interface >> >> is just the worst for people who are not registered. You guys should >> >> take a look at gitter.im. It integrates with GitHub, it's free, and it >> >> has nice things like backlogs on chats for everyone. >> > >> > >> > We do actually see a couple of questions on Hipchat most days - probably >> > more traffic than the mailing list - but I don't think most of the >> > people >> > using the help chat room would have heard about the survey. Gitter does >> > look >> > interesting, though. >> > >> > Brian: I think Gitter has private rooms if you associate it with private >> > repos. People who can see the repo see the chat room. I don't know about >> > direct messaging. >> > >> >> >> >> I'm a little surprised about the academia vs. industry usage. I >> >> suspect there were just more respondents from the academic side than >> >> from industry. >> > >> > >> > That's quite possible. The results are probably also biased towards >> > English >> > speakers, early adopters, and technically proficient users. Getting a >> > representative sample is pretty tricky. But I think the results give us >> > some >> > ballpark figures. >> > >> > Thomas >> > >> > _______________________________________________ >> > IPython-dev mailing list >> > IPython-dev at scipy.org >> > http://mail.scipy.org/mailman/listinfo/ipython-dev >> > >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From nelle.varoquaux at gmail.com Tue Mar 11 02:38:35 2014 From: nelle.varoquaux at gmail.com (Nelle Varoquaux) Date: Tue, 11 Mar 2014 07:38:35 +0100 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: I think I already asked Matthias about this, but why not simply use irc? It has all the features you require, and is the default communication channel for software engineers. On 11 March 2014 02:39, Aaron Meurer wrote: > The guest access of HipChat is really what sucks to me. There is no > chat history, so accidentally reloading the page causes you to lose > all context. > > Gitter does email you when a starred room has activity. I think they > also email you if you are @mentioned (not positive about that last > part, but it would be silly if they didn't). According to > https://gitter.im/apps phone apps are coming soon. > > I've been using it since the public beta started in December and I've > been really impressed with how fast it has improved. > > The only concern with gitter is that they are currently in beta, so > it's not clear what will remain free when they leave it. But seeing as > you guys are already paying for HipChat, it shouldn't be as big of a > concern for you. > > Aaron Meurer > > On Mon, Mar 10, 2014 at 8:08 PM, MinRK wrote: >> There are chat rooms for organizations and repos >> >> - only members of an organization can join the org room >> - public repo rooms are public >> - private repo rooms are private >> >> I think there are two things that we use on HipChat that gitter doesn't do: >> >> - guest access (GitHub login is required even for public rooms) >> - push/email notifications for mentions when idle (it supports desktop, but >> not offline notifications) >> >> For the most part, I think Gitter seems a lot better than HipChat, >> especially for guests (at least those with GitHub accounts). I am not >> especially fond of HipChat, and the guest experience is pretty horrendous >> (no native app, no history, frequent crashes, etc.), so I wouldn't mind >> giving something else a try. >> >> -MinRK >> >> >> >> On Mon, Mar 10, 2014 at 5:45 PM, Aaron Meurer wrote: >>> >>> Gitter does support direct messages. I don't know if it supports private >>> rooms. >>> >>> Aaron Meurer >>> >>> On Mon, Mar 10, 2014 at 3:49 PM, Thomas Kluyver wrote: >>> > On 10 March 2014 13:02, Aaron Meurer wrote: >>> >> >>> >> I didn't know you guys were dropping 2.6 support so soon. >>> > >>> > >>> > Yep. We've been fairly aggressive with Python version requirements for a >>> > while - we were one of the first projects to drop 2.5 support as well. >>> > We >>> > anticipate we'll get some bug reports about it, but we think it's worth >>> > it >>> > to be able to use 2.7 features reliably. It's so easy to forget what 2.6 >>> > didn't have, which is why we had to release 1.2.1 soon after 1.2.0. >>> > >>> >> >>> >> Regarding hipchat, I hate to say it, but nobody uses it because it >>> >> sucks. I personally hate it every time I try to use it. The interface >>> >> is just the worst for people who are not registered. You guys should >>> >> take a look at gitter.im. It integrates with GitHub, it's free, and it >>> >> has nice things like backlogs on chats for everyone. >>> > >>> > >>> > We do actually see a couple of questions on Hipchat most days - probably >>> > more traffic than the mailing list - but I don't think most of the >>> > people >>> > using the help chat room would have heard about the survey. Gitter does >>> > look >>> > interesting, though. >>> > >>> > Brian: I think Gitter has private rooms if you associate it with private >>> > repos. People who can see the repo see the chat room. I don't know about >>> > direct messaging. >>> > >>> >> >>> >> I'm a little surprised about the academia vs. industry usage. I >>> >> suspect there were just more respondents from the academic side than >>> >> from industry. >>> > >>> > >>> > That's quite possible. The results are probably also biased towards >>> > English >>> > speakers, early adopters, and technically proficient users. Getting a >>> > representative sample is pretty tricky. But I think the results give us >>> > some >>> > ballpark figures. >>> > >>> > Thomas >>> > >>> > _______________________________________________ >>> > IPython-dev mailing list >>> > IPython-dev at scipy.org >>> > http://mail.scipy.org/mailman/listinfo/ipython-dev >>> > >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From bussonniermatthias at gmail.com Tue Mar 11 04:16:59 2014 From: bussonniermatthias at gmail.com (Matthias BUSSONNIER) Date: Tue, 11 Mar 2014 09:16:59 +0100 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: Hi all, Le 11 mars 2014 ? 07:38, Nelle Varoquaux a ?crit : > I think I already asked Matthias about this, but why not simply use irc? > It has all the features you require, and is the default communication channel > for software engineers. The Irc channel still exist, one of the problem is the energy barrier to enter it. And for what it is worth, hip chat is really great for beginner user that have no github account you just have to follow the link, enter your name and chat. I personally don't like hipchat, but there is one big advantage for me, being out of schedule with most of the team, @mention me in one of hip chat room send me a text (don't' always have phone data everywhere, or on, so app-things is next to useless to me). Also the migration to hipchat was done more or less at the same time as the "deprecation" of IPython-user list, and I think one of the main goal was to simplify things for newcomers. > There should be one-- and preferably only one --obvious way to do it. This thing is hipchat. But it does not prevent you from asking on irc, stack overflow, twitter? We should just not tell someone who is in need for help that he need to choose, and that he need to install something and learn a new software. As much as I prefer gitter to hipchat, you can have 100ds of room on gitter, which is not friendly for someone not used to the project, and UI is confusing. Finally I've started to see if we could have a hubot to do hipchat/irc/whatever bridge. it seem possible I just miss time to write it. From matthewturk at gmail.com Tue Mar 11 08:13:29 2014 From: matthewturk at gmail.com (Matthew Turk) Date: Tue, 11 Mar 2014 08:13:29 -0400 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: Hi everyone, On Mar 11, 2014 4:15 AM, "Matthias BUSSONNIER" wrote: > > Hi all, > > Le 11 mars 2014 ? 07:38, Nelle Varoquaux a ?crit : > > > I think I already asked Matthias about this, but why not simply use irc? > > It has all the features you require, and is the default communication channel > > for software engineers. > > The Irc channel still exist, one of the problem is the energy barrier to enter it. > And for what it is worth, hip chat is really great for beginner user that have no github account > you just have to follow the link, enter your name and chat. We're a smaller project (by a good amount) than IPython, but we've had good luck with the Freenode web gateway for yt: http://webchat.freenode.net/ You can have it pre-populate the channels list by adding an argument to the iframe like ?channels=ipython . Guest accounts work just fine here. -Matt > > I personally don't like hipchat, but there is one big advantage for me, being out of schedule with most of the team, > @mention me in one of hip chat room send me a text (don't' always have phone data everywhere, or on, so app-things > is next to useless to me). > > Also the migration to hipchat was done more or less at the same time as the "deprecation" of IPython-user list, > and I think one of the main goal was to simplify things for newcomers. > > > There should be one-- and preferably only one --obvious way to do it. > > This thing is hipchat. > > But it does not prevent you from asking on irc, stack overflow, twitter... > We should just not tell someone who is in need for help that he need to choose, > and that he need to install something and learn a new software. > > As much as I prefer gitter to hipchat, you can have 100ds of room on gitter, > which is not friendly for someone not used to the project, and UI is confusing. > > Finally I've started to see if we could have a hubot to do hipchat/irc/whatever bridge. > it seem possible I just miss time to write it. > > > > > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From mcburton at umich.edu Tue Mar 11 09:00:53 2014 From: mcburton at umich.edu (mcburton) Date: Tue, 11 Mar 2014 14:00:53 +0100 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: ++ for freenode IRC with the webchat. On Tue, Mar 11, 2014 at 1:13 PM, Matthew Turk wrote: > Hi everyone, > > On Mar 11, 2014 4:15 AM, "Matthias BUSSONNIER" > wrote: >> >> Hi all, >> >> Le 11 mars 2014 ? 07:38, Nelle Varoquaux a ?crit : >> >> > I think I already asked Matthias about this, but why not simply use irc? >> > It has all the features you require, and is the default communication channel >> > for software engineers. >> >> The Irc channel still exist, one of the problem is the energy barrier to enter it. >> And for what it is worth, hip chat is really great for beginner user that have no github account >> you just have to follow the link, enter your name and chat. > > We're a smaller project (by a good amount) than IPython, but we've had > good luck with the Freenode web gateway for yt: > > http://webchat.freenode.net/ > > You can have it pre-populate the channels list by adding an argument > to the iframe like ?channels=ipython . Guest accounts work just fine > here. > > -Matt > >> >> I personally don't like hipchat, but there is one big advantage for me, being out of schedule with most of the team, >> @mention me in one of hip chat room send me a text (don't' always have phone data everywhere, or on, so app-things >> is next to useless to me). >> >> Also the migration to hipchat was done more or less at the same time as the "deprecation" of IPython-user list, >> and I think one of the main goal was to simplify things for newcomers. >> >> > There should be one-- and preferably only one --obvious way to do it. >> >> This thing is hipchat. >> >> But it does not prevent you from asking on irc, stack overflow, twitter... >> We should just not tell someone who is in need for help that he need to choose, >> and that he need to install something and learn a new software. >> >> As much as I prefer gitter to hipchat, you can have 100ds of room on gitter, >> which is not friendly for someone not used to the project, and UI is confusing. >> >> Finally I've started to see if we could have a hubot to do hipchat/irc/whatever bridge. >> it seem possible I just miss time to write it. >> >> >> >> >> >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From carl.input at gmail.com Tue Mar 11 10:20:58 2014 From: carl.input at gmail.com (Carl Smith) Date: Tue, 11 Mar 2014 14:20:58 +0000 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: Two Pennies... On surveys, and communication in general, could you not just add a feature to the Notebook that allows it to connect to a central server onload, so it can let users know there's some news and lets them load the content directly? Users should have an obvious, one click toggle to disable the connections, but few would. If you don't use cookies, or really anything that involves auth or privacy, you could do a lot with that easily. Serving Markdown and even pickled [and encrypted] Python objects is all easy if you write the client and the server. On one-- and preferably only one --obvious way, if the buttons are in the app, they're all obvious and are all part of the way it's done. The obvious way should also always be in the Standard Library. -------------- next part -------------- An HTML attachment was scrubbed... URL: From wstein at gmail.com Tue Mar 11 10:43:42 2014 From: wstein at gmail.com (William Stein) Date: Tue, 11 Mar 2014 07:43:42 -0700 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: On Tue, Mar 11, 2014 at 7:20 AM, Carl Smith wrote: > Two Pennies... > > On surveys, and communication in general, could you not just add a feature > to the Notebook that allows it to connect to a central server onload, so it > can let users know there's some news and lets them load the content > directly? Users should have an obvious, one click toggle to disable the > connections, but few would. If you do this sort of "phoning home", it would be good to come up with a plan about how you're going to use the resulting data. You'll have the potential to suddenly know a lot: - how many people are starting IPython, which could be useful data to have (e.g., for grants, etc.). - where they are using it (based on the ip) - what version they are using, so you can tell them to upgrade in the response message. Etc. -- william > > If you don't use cookies, or really anything that involves auth or privacy, > you could do a lot with that easily. Serving Markdown and even pickled [and > encrypted] Python objects is all easy if you write the client and the > server. > > On one-- and preferably only one --obvious way, if the buttons are in the > app, they're all obvious and are all part of the way it's done. The obvious > way should also always be in the Standard Library. > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- William Stein Professor of Mathematics University of Washington http://wstein.org From carl.input at gmail.com Tue Mar 11 11:02:48 2014 From: carl.input at gmail.com (Carl Smith) Date: Tue, 11 Mar 2014 15:02:48 +0000 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: Sorry. Just to be clear, the suggestion was to allow the notebook to /inform/ the user that there's a survey the developers would appreciate them filling out. I didn't mean to collect usage stats silently. Only take what is explicitly given. Now that you mention it though, so long as it was something the user agreed to /each time/ and the data was taken as a snapshot /at that time/, I'd be +1 on that too. -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Tue Mar 11 12:05:25 2014 From: benjaminrk at gmail.com (Min RK) Date: Tue, 11 Mar 2014 09:05:25 -0700 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: <9A084E0D-A9DB-4101-A406-A9B1B49425AA@gmail.com> > On Mar 10, 2014, at 23:38, Nelle Varoquaux wrote: > > I think I already asked Matthias about this, but why not simply use irc? > It has all the features you require, and is the default communication channel > for software engineers. > We started using hipchat specifically because IRC wasn't working well, and it doesn't have features that we require. Among them: - offline notification - history persistence - search -MinRK >> On 11 March 2014 02:39, Aaron Meurer wrote: >> The guest access of HipChat is really what sucks to me. There is no >> chat history, so accidentally reloading the page causes you to lose >> all context. >> >> Gitter does email you when a starred room has activity. I think they >> also email you if you are @mentioned (not positive about that last >> part, but it would be silly if they didn't). According to >> https://gitter.im/apps phone apps are coming soon. >> >> I've been using it since the public beta started in December and I've >> been really impressed with how fast it has improved. >> >> The only concern with gitter is that they are currently in beta, so >> it's not clear what will remain free when they leave it. But seeing as >> you guys are already paying for HipChat, it shouldn't be as big of a >> concern for you. >> >> Aaron Meurer >> >>> On Mon, Mar 10, 2014 at 8:08 PM, MinRK wrote: >>> There are chat rooms for organizations and repos >>> >>> - only members of an organization can join the org room >>> - public repo rooms are public >>> - private repo rooms are private >>> >>> I think there are two things that we use on HipChat that gitter doesn't do: >>> >>> - guest access (GitHub login is required even for public rooms) >>> - push/email notifications for mentions when idle (it supports desktop, but >>> not offline notifications) >>> >>> For the most part, I think Gitter seems a lot better than HipChat, >>> especially for guests (at least those with GitHub accounts). I am not >>> especially fond of HipChat, and the guest experience is pretty horrendous >>> (no native app, no history, frequent crashes, etc.), so I wouldn't mind >>> giving something else a try. >>> >>> -MinRK >>> >>> >>> >>>> On Mon, Mar 10, 2014 at 5:45 PM, Aaron Meurer wrote: >>>> >>>> Gitter does support direct messages. I don't know if it supports private >>>> rooms. >>>> >>>> Aaron Meurer >>>> >>>>> On Mon, Mar 10, 2014 at 3:49 PM, Thomas Kluyver wrote: >>>>>> On 10 March 2014 13:02, Aaron Meurer wrote: >>>>>> >>>>>> I didn't know you guys were dropping 2.6 support so soon. >>>>> >>>>> >>>>> Yep. We've been fairly aggressive with Python version requirements for a >>>>> while - we were one of the first projects to drop 2.5 support as well. >>>>> We >>>>> anticipate we'll get some bug reports about it, but we think it's worth >>>>> it >>>>> to be able to use 2.7 features reliably. It's so easy to forget what 2.6 >>>>> didn't have, which is why we had to release 1.2.1 soon after 1.2.0. >>>>> >>>>>> >>>>>> Regarding hipchat, I hate to say it, but nobody uses it because it >>>>>> sucks. I personally hate it every time I try to use it. The interface >>>>>> is just the worst for people who are not registered. You guys should >>>>>> take a look at gitter.im. It integrates with GitHub, it's free, and it >>>>>> has nice things like backlogs on chats for everyone. >>>>> >>>>> >>>>> We do actually see a couple of questions on Hipchat most days - probably >>>>> more traffic than the mailing list - but I don't think most of the >>>>> people >>>>> using the help chat room would have heard about the survey. Gitter does >>>>> look >>>>> interesting, though. >>>>> >>>>> Brian: I think Gitter has private rooms if you associate it with private >>>>> repos. People who can see the repo see the chat room. I don't know about >>>>> direct messaging. >>>>> >>>>>> >>>>>> I'm a little surprised about the academia vs. industry usage. I >>>>>> suspect there were just more respondents from the academic side than >>>>>> from industry. >>>>> >>>>> >>>>> That's quite possible. The results are probably also biased towards >>>>> English >>>>> speakers, early adopters, and technically proficient users. Getting a >>>>> representative sample is pretty tricky. But I think the results give us >>>>> some >>>>> ballpark figures. >>>>> >>>>> Thomas >>>>> >>>>> _______________________________________________ >>>>> IPython-dev mailing list >>>>> IPython-dev at scipy.org >>>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>>>> >>>> _______________________________________________ >>>> IPython-dev mailing list >>>> IPython-dev at scipy.org >>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >>> >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From takowl at gmail.com Tue Mar 11 12:30:34 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Tue, 11 Mar 2014 09:30:34 -0700 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: On 11 March 2014 05:13, Matthew Turk wrote: > We're a smaller project (by a good amount) than IPython, but we've had > good luck with the Freenode web gateway for yt: > > http://webchat.freenode.net/ > For the record: before we switched to hipchat, we did promote the freenode web version of the #ipython IRC channel with a link on the homepage. In addition to the reasons Min mentioned, Hipchat can handle images inline, which is handy for screenshots as the project has more of a visual component. @mentions are the nicest thing, though - if I see a question in the help room that I can't answer, but I know who can, I can ping them, and most of the time, they log in a few seconds later. > On surveys, and communication in general, could you not just add a feature to the Notebook that allows it to connect to a central server onload, so it can let users know there's some news and lets them load the content directly? Users should have an obvious, one click toggle to disable the connections, but few would. We have actually talked about having some kind of statistics gathering, on a strictly opt in basis. Delivering information would be an interesting extension of that. Thanks, Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From ralf.gommers at gmail.com Tue Mar 11 15:22:35 2014 From: ralf.gommers at gmail.com (Ralf Gommers) Date: Tue, 11 Mar 2014 20:22:35 +0100 Subject: [IPython-dev] 2014 John Hunter Fellowship - Call for Applications Message-ID: Hi all, I'm excited to announce, on behalf of the Numfocus board, that applications for the 2014 John Hunter Technology Fellowship are now being accepted. This is the first fellowship Numfocus is able to offer, which we see as a significant milestone. The John Hunter Technology Fellowship aims to bridge the gap between academia and real-world, open-source scientific computing projects by providing a capstone experience for individuals coming from a scientific, engineering or mathematics background. The program consists of a 6 month project-based training program for postdoctoral scientists or senior graduate students. Fellows work on scientific computing open source projects under the guidance of mentors who are leading scientists and software engineers. The aim of the Fellowship is to enable Fellows to develop the skills needed to contribute to cutting-edge open source software projects while at the same time advancing or supporting the research program they and their mentor are involved in. While proposals in any area of science and engineering are welcome, the following areas are encouraged in particular: - Accessible and reproducible computing - Enabling technology for open access publishing - Infrastructural technology supporting open-source scientific software stacks - Core open-source projects promoted by NumFOCUS Eligible applicants are postdoctoral scientists or senior PhD students, or have equivalent experience in physics, mathematics, engineering, statistics, or a related science. The program is open to applicants from any nationality and can be performed at any university or institute world-wide (US export laws permitting). All applications are due May 15, 2014 by 11:59 p.m. Central Standard Time. For more details on the program see: http://numfocus.org/john_hunter_fellowship_2014.html (this call) http://numfocus.org/fellowships.html (program) And for some background see this blog post: http://numfocus.org/announcing-the-numfocus-technology-fellowship-program.html We're looking forward to receiving your applications! Ralf -------------- next part -------------- An HTML attachment was scrubbed... URL: From patrick.surry at gmail.com Tue Mar 11 16:19:27 2014 From: patrick.surry at gmail.com (Patrick Surry) Date: Tue, 11 Mar 2014 16:19:27 -0400 Subject: [IPython-dev] ERROR: global name 'select_figure_formats' is not defined Message-ID: Works a treat. And I see you've already implemented the answer to my next question before I'd even had time to ask. (Namely, is there a way to find all the "running" notebooks in the tree view? - et voila, a "Running" tab has appeared after my latest git pull). Superior service! Thanks, Patrick >Thomas Kluyver wrote: > >Try this: https://github.com/ipython/ipython/pull/5318 -------------- next part -------------- An HTML attachment was scrubbed... URL: From claresloggett at gmail.com Tue Mar 11 19:39:16 2014 From: claresloggett at gmail.com (Clare Sloggett) Date: Wed, 12 Mar 2014 10:39:16 +1100 Subject: [IPython-dev] 2014 John Hunter Fellowship - Call for Applications In-Reply-To: References: Message-ID: Hi Ralf, Would this fellowship apply to the life sciences as well? Thanks, Clare On 12 March 2014 06:22, Ralf Gommers wrote: > Hi all, > > I'm excited to announce, on behalf of the Numfocus board, that > applications for the > 2014 John Hunter Technology Fellowship are now being accepted. This is the > first > fellowship Numfocus is able to offer, which we see as a significant > milestone. > > The John Hunter Technology Fellowship aims to bridge the gap between > academia and real-world, open-source scientific computing projects by > providing > a capstone experience for individuals coming from a scientific, > engineering or > mathematics background. > > The program consists of a 6 month project-based training program > for postdoctoral scientists or senior graduate students. Fellows work on > scientific computing open source projects under the guidance of mentors > who are > leading scientists and software engineers. The aim of the Fellowship is to > enable Fellows to develop the skills needed to contribute to cutting-edge > open > source software projects while at the same time advancing or supporting the > research program they and their mentor are involved in. > > While proposals in any area of science and engineering are welcome, the > following areas are encouraged in particular: > > - Accessible and reproducible computing > - Enabling technology for open access publishing > - Infrastructural technology supporting open-source scientific software > stacks > - Core open-source projects promoted by NumFOCUS > > Eligible applicants are postdoctoral scientists or senior PhD students, > or have equivalent experience in physics, mathematics, engineering, > statistics, > or a related science. The program is open to applicants from any > nationality > and can be performed at any university or institute world-wide (US export > laws permitting). > > All applications are due May 15, 2014 by 11:59 p.m. Central Standard Time. > > For more details on the program see: > http://numfocus.org/john_hunter_fellowship_2014.html (this call) > http://numfocus.org/fellowships.html (program) > > And for some background see this blog post: > > http://numfocus.org/announcing-the-numfocus-technology-fellowship-program.html > > We're looking forward to receiving your applications! > > Ralf > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ralf.gommers at gmail.com Wed Mar 12 03:08:19 2014 From: ralf.gommers at gmail.com (Ralf Gommers) Date: Wed, 12 Mar 2014 08:08:19 +0100 Subject: [IPython-dev] 2014 John Hunter Fellowship - Call for Applications In-Reply-To: References: Message-ID: Hi Clare On Wed, Mar 12, 2014 at 12:39 AM, Clare Sloggett wrote: > Hi Ralf, > > Would this fellowship apply to the life sciences as well? > Any proposal that falls under "open-source scientific computing" is very welcome. So yes, I see no reason to exclude life sciences. Cheers, Ralf > > Thanks, > Clare > > > On 12 March 2014 06:22, Ralf Gommers wrote: > >> Hi all, >> >> I'm excited to announce, on behalf of the Numfocus board, that >> applications for the >> 2014 John Hunter Technology Fellowship are now being accepted. This is >> the first >> fellowship Numfocus is able to offer, which we see as a significant >> milestone. >> >> The John Hunter Technology Fellowship aims to bridge the gap between >> academia and real-world, open-source scientific computing projects by >> providing >> a capstone experience for individuals coming from a scientific, >> engineering or >> mathematics background. >> >> The program consists of a 6 month project-based training program >> for postdoctoral scientists or senior graduate students. Fellows work on >> scientific computing open source projects under the guidance of mentors >> who are >> leading scientists and software engineers. The aim of the Fellowship is to >> enable Fellows to develop the skills needed to contribute to cutting-edge >> open >> source software projects while at the same time advancing or supporting >> the >> research program they and their mentor are involved in. >> >> While proposals in any area of science and engineering are welcome, the >> following areas are encouraged in particular: >> >> - Accessible and reproducible computing >> - Enabling technology for open access publishing >> - Infrastructural technology supporting open-source scientific software >> stacks >> - Core open-source projects promoted by NumFOCUS >> >> Eligible applicants are postdoctoral scientists or senior PhD students, >> or have equivalent experience in physics, mathematics, engineering, >> statistics, >> or a related science. The program is open to applicants from any >> nationality >> and can be performed at any university or institute world-wide (US export >> laws permitting). >> >> All applications are due May 15, 2014 by 11:59 p.m. Central Standard Time. >> >> For more details on the program see: >> http://numfocus.org/john_hunter_fellowship_2014.html (this call) >> http://numfocus.org/fellowships.html (program) >> >> And for some background see this blog post: >> >> http://numfocus.org/announcing-the-numfocus-technology-fellowship-program.html >> >> We're looking forward to receiving your applications! >> >> Ralf >> >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jr at sun.ac.za Wed Mar 12 10:46:19 2014 From: jr at sun.ac.za (Johann Rohwer) Date: Wed, 12 Mar 2014 16:46:19 +0200 Subject: [IPython-dev] Deploying IPython on Windows campus network Message-ID: <3213443.a10FyKgLBC@biochem433789> We are trying to deploy IPython on our Windows campus network using Microsoft Application Virtualization technology. On our campus network all users have a home directory which is automounted as the network drive H:\ upon login. What is the easiest way to tell IPython to use H:\.ipython or H:\ipython as the configuration directory for IPython? (If I am not mistaken the home drive is set by the environment variable %HOMEDRIVE% under Windows). Unfortunately I am not very familiar with Windows (using Linux myself) and our Windows sysadmin has limited python knowledge. Currently our installation is working, but upon startup the current working directory is set to C:\Windows\system32 which is not writeable for normal users - this is expecially problematic for the notebook which won't start up because of a lack of write permission. Also the IPython config directory defaults to C: \Users\%USERNAME%\.ipython which is machine specific, and we would like it to be under a user's network home directory. We are using the Anaconda python distribution. Any pointers would be appreciated. Regards, Johann The integrity and confidentiality of this email is governed by these terms / Hierdie terme bepaal die integriteit en vertroulikheid van hierdie epos. http://www.sun.ac.za/emaildisclaimer From cfriedline at vcu.edu Wed Mar 12 10:55:21 2014 From: cfriedline at vcu.edu (Chris Friedline) Date: Wed, 12 Mar 2014 10:55:21 -0400 Subject: [IPython-dev] Deploying IPython on Windows campus network In-Reply-To: <3213443.a10FyKgLBC@biochem433789> References: <3213443.a10FyKgLBC@biochem433789> Message-ID: <53207559.6070104@vcu.edu> Hi Johann, That path is controlled by the environment variable, IPYTHONDIR, which your windows admin could set on the boxes via Group Policy. See here for more info: http://ipython.org/ipython-doc/dev/config/intro.html HTH, Chris -- Christopher J. Friedline, Ph.D. NSF Postdoctoral Research Fellow Virginia Commonwealth University Richmond, VA 23284 http://chris.friedline.net Johann Rohwer wrote: > We are trying to deploy IPython on our Windows campus network using Microsoft > Application Virtualization technology. On our campus network all users have a > home directory which is automounted as the network drive H:\ upon login. What > is the easiest way to tell IPython to use H:\.ipython or H:\ipython as the > configuration directory for IPython? (If I am not mistaken the home drive is > set by the environment variable %HOMEDRIVE% under Windows). > > Unfortunately I am not very familiar with Windows (using Linux myself) and our > Windows sysadmin has limited python knowledge. Currently our installation is > working, but upon startup the current working directory is set to > C:\Windows\system32 which is not writeable for normal users - this is > expecially problematic for the notebook which won't start up because of a lack > of write permission. Also the IPython config directory defaults to C: > \Users\%USERNAME%\.ipython which is machine specific, and we would like it to > be under a user's network home directory. > > We are using the Anaconda python distribution. > > Any pointers would be appreciated. > Regards, > Johann > > The integrity and confidentiality of this email is governed by these terms / Hierdie terme bepaal die integriteit en vertroulikheid van hierdie epos. http://www.sun.ac.za/emaildisclaimer > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 801 bytes Desc: OpenPGP digital signature URL: From claresloggett at gmail.com Wed Mar 12 19:59:07 2014 From: claresloggett at gmail.com (Clare Sloggett) Date: Thu, 13 Mar 2014 10:59:07 +1100 Subject: [IPython-dev] 2014 John Hunter Fellowship - Call for Applications In-Reply-To: References: Message-ID: Thanks! On 12/03/2014 6:08 PM, "Ralf Gommers" wrote: > > Hi Clare > > On Wed, Mar 12, 2014 at 12:39 AM, Clare Sloggett wrote: > >> Hi Ralf, >> >> Would this fellowship apply to the life sciences as well? >> > > Any proposal that falls under "open-source scientific computing" is very > welcome. So yes, I see no reason to exclude life sciences. > > Cheers, > Ralf > > > >> >> Thanks, >> Clare >> >> >> On 12 March 2014 06:22, Ralf Gommers wrote: >> >>> Hi all, >>> >>> I'm excited to announce, on behalf of the Numfocus board, that >>> applications for the >>> 2014 John Hunter Technology Fellowship are now being accepted. This is >>> the first >>> fellowship Numfocus is able to offer, which we see as a significant >>> milestone. >>> >>> The John Hunter Technology Fellowship aims to bridge the gap between >>> academia and real-world, open-source scientific computing projects by >>> providing >>> a capstone experience for individuals coming from a scientific, >>> engineering or >>> mathematics background. >>> >>> The program consists of a 6 month project-based training program >>> for postdoctoral scientists or senior graduate students. Fellows work on >>> scientific computing open source projects under the guidance of mentors >>> who are >>> leading scientists and software engineers. The aim of the Fellowship is >>> to >>> enable Fellows to develop the skills needed to contribute to >>> cutting-edge open >>> source software projects while at the same time advancing or supporting >>> the >>> research program they and their mentor are involved in. >>> >>> While proposals in any area of science and engineering are welcome, the >>> following areas are encouraged in particular: >>> >>> - Accessible and reproducible computing >>> - Enabling technology for open access publishing >>> - Infrastructural technology supporting open-source scientific >>> software stacks >>> - Core open-source projects promoted by NumFOCUS >>> >>> Eligible applicants are postdoctoral scientists or senior PhD students, >>> or have equivalent experience in physics, mathematics, engineering, >>> statistics, >>> or a related science. The program is open to applicants from any >>> nationality >>> and can be performed at any university or institute world-wide (US >>> export >>> laws permitting). >>> >>> All applications are due May 15, 2014 by 11:59 p.m. Central Standard >>> Time. >>> >>> For more details on the program see: >>> http://numfocus.org/john_hunter_fellowship_2014.html (this call) >>> http://numfocus.org/fellowships.html (program) >>> >>> And for some background see this blog post: >>> >>> http://numfocus.org/announcing-the-numfocus-technology-fellowship-program.html >>> >>> We're looking forward to receiving your applications! >>> >>> Ralf >>> >>> >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >>> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jason-sage at creativetrax.com Wed Mar 12 22:46:52 2014 From: jason-sage at creativetrax.com (Jason Grout) Date: Wed, 12 Mar 2014 21:46:52 -0500 Subject: [IPython-dev] Link widget Message-ID: <53211C1C.3080209@creativetrax.com> Hi all, I was working on the next level of optimizations for linking widgets together, and came up with a Link widget that links widget values together on the javascript side, to avoid a roundtrip to python. Demo here: http://sagecell.sagemath.org/?q=bhlnao [1] Notice that the first two sliders are lockstep with each other (they are linked with the Link widget), but the third slider has a delay because it necessitates a roundtrip to python. Code is in my ipywidgets repository at https://github.com/jasongrout/ipywidgets The Link widget still has some things to fix (like unregistering if it is destroyed or explicitly unbound, testing to make sure changing the list of bindings has the right effect, etc.), but do you think it is valuable enough of an idea to include in the standard widget set? Thanks, Jason [1] Code at the link is: html("") load('https://rawgithub.com/jasongrout/ipywidgets/master/widgets.py') from IPython.html import widgets a = widgets.FloatSliderWidget(value=30) b = widgets.FloatSliderWidget() c = widgets.FloatSliderWidget() show(a) show(b) show(c) javascript_link=Link(widgets=[[a,'value'], [b,'value']]) show(javascript_link) from IPython.utils.traitlets import link python_link = link([a,'value'], [c, 'value']) print "Notice the delay in the third slider from the roundtrip to python." From sylvain.corlay at gmail.com Thu Mar 13 00:11:43 2014 From: sylvain.corlay at gmail.com (Sylvain Corlay) Date: Thu, 13 Mar 2014 00:11:43 -0400 Subject: [IPython-dev] Link widget In-Reply-To: <53211C1C.3080209@creativetrax.com> References: <53211C1C.3080209@creativetrax.com> Message-ID: Hi Jason, I think this is great. Exactly what we needed for certain performance issues we encountered. Best, Sylvain On Wed, Mar 12, 2014 at 10:46 PM, Jason Grout wrote: > Hi all, > > I was working on the next level of optimizations for linking widgets > together, and came up with a Link widget that links widget values > together on the javascript side, to avoid a roundtrip to python. Demo > here: > > http://sagecell.sagemath.org/?q=bhlnao [1] > > Notice that the first two sliders are lockstep with each other (they are > linked with the Link widget), but the third slider has a delay because > it necessitates a roundtrip to python. > > Code is in my ipywidgets repository at > https://github.com/jasongrout/ipywidgets > > The Link widget still has some things to fix (like unregistering if it > is destroyed or explicitly unbound, testing to make sure changing the > list of bindings has the right effect, etc.), but do you think it is > valuable enough of an idea to include in the standard widget set? > > Thanks, > > Jason > > > [1] Code at the link is: > > html("") > load('https://rawgithub.com/jasongrout/ipywidgets/master/widgets.py') > > from IPython.html import widgets > a = widgets.FloatSliderWidget(value=30) > b = widgets.FloatSliderWidget() > c = widgets.FloatSliderWidget() > show(a) > show(b) > show(c) > javascript_link=Link(widgets=[[a,'value'], [b,'value']]) > show(javascript_link) > > from IPython.utils.traitlets import link > python_link = link([a,'value'], [c, 'value']) > > print "Notice the delay in the third slider from the roundtrip to python." > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jason-sage at creativetrax.com Thu Mar 13 00:23:10 2014 From: jason-sage at creativetrax.com (Jason Grout) Date: Wed, 12 Mar 2014 23:23:10 -0500 Subject: [IPython-dev] Link widget In-Reply-To: References: <53211C1C.3080209@creativetrax.com> Message-ID: <532132AE.8090500@creativetrax.com> On 3/12/14, 23:11, Sylvain Corlay wrote: > Hi Jason, > I think this is great. Exactly what we needed for certain performance > issues we encountered. It also fixed some performance issues with a demo I was working on dealing with texturing a sphere with an image from your webcam. When the image has to make the roundtrip to the server, it's quite a bit slower than just snapping the image onto the sphere all in javascript. Glad to hear it was useful to you too. Thanks, Jason From ellisonbg at gmail.com Thu Mar 13 12:54:57 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Thu, 13 Mar 2014 09:54:57 -0700 Subject: [IPython-dev] Link widget In-Reply-To: <532132AE.8090500@creativetrax.com> References: <53211C1C.3080209@creativetrax.com> <532132AE.8090500@creativetrax.com> Message-ID: I am +1 on this in IPython, but not for 2.0 On Wed, Mar 12, 2014 at 9:23 PM, Jason Grout wrote: > On 3/12/14, 23:11, Sylvain Corlay wrote: >> Hi Jason, >> I think this is great. Exactly what we needed for certain performance >> issues we encountered. > > It also fixed some performance issues with a demo I was working on > dealing with texturing a sphere with an image from your webcam. When > the image has to make the roundtrip to the server, it's quite a bit > slower than just snapping the image onto the sphere all in javascript. > > Glad to hear it was useful to you too. > > Thanks, > > Jason > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From jason-sage at creativetrax.com Thu Mar 13 13:01:31 2014 From: jason-sage at creativetrax.com (Jason Grout) Date: Thu, 13 Mar 2014 12:01:31 -0500 Subject: [IPython-dev] Link widget In-Reply-To: References: <53211C1C.3080209@creativetrax.com> <532132AE.8090500@creativetrax.com> Message-ID: <5321E46B.1080803@creativetrax.com> On 3/13/14, 11:54, Brian Granger wrote: > I am +1 on this in IPython, but not for 2.0 Understood. As I finish it over the next little bit, I'll put in a pull request and you can merge it whenever. I'll keep it in my ipywidgets repository for the time being as well. Thanks, Jason From apratap at sagebase.org Thu Mar 13 13:51:20 2014 From: apratap at sagebase.org (Abhishek Pratap) Date: Thu, 13 Mar 2014 10:51:20 -0700 Subject: [IPython-dev] shortcuts for adding/deleting cells Message-ID: Hey Guys I just updated to master on gtihub v2.0. Just wondering what are the shortcuts for adding or deleting a new cell. Previously I was able to do CTRL-M A/M/D etc. Has anything changed. -Abhi From sylvain.corlay at gmail.com Thu Mar 13 13:59:50 2014 From: sylvain.corlay at gmail.com (Sylvain Corlay) Date: Thu, 13 Mar 2014 13:59:50 -0400 Subject: [IPython-dev] Link widget In-Reply-To: <5321E46B.1080803@creativetrax.com> References: <53211C1C.3080209@creativetrax.com> <532132AE.8090500@creativetrax.com> <5321E46B.1080803@creativetrax.com> Message-ID: Quick question on the API: Does it really have to be a widget, or could it be a function in IPython.html.widgets. from IPython.html import widgets a = widgets.FloatSliderWidget(value=30) b = widgets.FloatSliderWidget() show(a) show(b) widgets.link([[a,'value'], [b,'value']]) # could be called "sync" as well On Thu, Mar 13, 2014 at 1:01 PM, Jason Grout wrote: > > > On 3/13/14, 11:54, Brian Granger wrote: > > I am +1 on this in IPython, but not for 2.0 > > > Understood. As I finish it over the next little bit, I'll put in a pull > request and you can merge it whenever. I'll keep it in my ipywidgets > repository for the time being as well. > > Thanks, > > Jason > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jason-sage at creativetrax.com Thu Mar 13 14:18:08 2014 From: jason-sage at creativetrax.com (Jason Grout) Date: Thu, 13 Mar 2014 13:18:08 -0500 Subject: [IPython-dev] Link widget In-Reply-To: References: <53211C1C.3080209@creativetrax.com> <532132AE.8090500@creativetrax.com> <5321E46B.1080803@creativetrax.com> Message-ID: <5321F660.8080000@creativetrax.com> On 3/13/14, 12:59, Sylvain Corlay wrote: > Quick question on the API: > > Does it really have to be a widget, or could it be a function in > IPython.html.widgets. > > from IPython.html import widgets > a = widgets.FloatSliderWidget(value=30) > b = widgets.FloatSliderWidget() > > show(a) > show(b) > > widgets.link([[a,'value'], [b,'value']]) # could be called "sync" as well > Somehow, the message has to get across to the javascript side that the models should be linked together. We could send a message to, say, the widget manager to do that, but right now we don't have a good way to directly communicate with the widget manager and send it messages. I hesitate to add such a messaging system when it seems even better to have the link object be a separate object we can manipulate by itself. By making the link a separate object (mirroring the link function in the traitlets file), it's easy to change the properties of the link, or abolish the link altogether. That said, we could make a link() function in the widgets.py file that would construct, display, and return the Link object. I chose the name link because that was what we decided to do in the traitlets case: https://github.com/ipython/ipython/pull/5060. Thanks, Jason From takowl at gmail.com Thu Mar 13 14:36:39 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Thu, 13 Mar 2014 11:36:39 -0700 Subject: [IPython-dev] shortcuts for adding/deleting cells In-Reply-To: References: Message-ID: Hi Abhi, There is a big change in the way keyboard shortcuts work. If you click on 'Help > User Interface Tour' in the menus, you should get a walkthrough of how the new UI works. Then have a look at 'Help > Keyboard Shortcuts' for what shortcuts are available. Thomas On 13 March 2014 10:51, Abhishek Pratap wrote: > Hey Guys > > I just updated to master on gtihub v2.0. Just wondering what are the > shortcuts for adding or deleting a new cell. Previously I was able to > do CTRL-M A/M/D etc. > > Has anything changed. > > -Abhi > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From apratap at sagebase.org Thu Mar 13 15:20:16 2014 From: apratap at sagebase.org (Abhishek Pratap) Date: Thu, 13 Mar 2014 12:20:16 -0700 Subject: [IPython-dev] shortcuts for adding/deleting cells In-Reply-To: References: Message-ID: Hi Thomas Ok I did look at them but I don't think from my mac I can run many of the cell related commands in the edit mode. For example: in the (command mode with grey cell) CTRL + a doesnt insert a new cell Also once I convert the cell to a markdown type the it doesn't get displayed as a markdown. example attached PS: I think a video on new features and how to use them would be very helpful. At this moment I feel very lost and suddenly feel very limited in my ability to navigate using IPython v2. -Abhi On Thu, Mar 13, 2014 at 11:36 AM, Thomas Kluyver wrote: > Hi Abhi, > > There is a big change in the way keyboard shortcuts work. If you click on > 'Help > User Interface Tour' in the menus, you should get a walkthrough of > how the new UI works. Then have a look at 'Help > Keyboard Shortcuts' for > what shortcuts are available. > > Thomas > > > On 13 March 2014 10:51, Abhishek Pratap wrote: >> >> Hey Guys >> >> I just updated to master on gtihub v2.0. Just wondering what are the >> shortcuts for adding or deleting a new cell. Previously I was able to >> do CTRL-M A/M/D etc. >> >> Has anything changed. >> >> -Abhi >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- A non-text attachment was scrubbed... Name: Screen Shot 2014-03-13 at 12.19.18 PM.png Type: image/png Size: 17970 bytes Desc: not available URL: From takowl at gmail.com Thu Mar 13 15:27:29 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Thu, 13 Mar 2014 12:27:29 -0700 Subject: [IPython-dev] shortcuts for adding/deleting cells In-Reply-To: References: Message-ID: On 13 March 2014 12:20, Abhishek Pratap wrote: > For example: > in the (command mode with grey cell) > > CTRL + a doesnt insert a new cell > It's not Ctrl-a, it's just a. When you're in command mode, most of the keyboard shortcuts are single keys pressed by themselves, like gmail keyboard shortcuts. The reason we've done this is because almost all Ctrl shortcuts are already used for something by the browser, and we don't want to conflict with them. > Also once I convert the cell to a markdown type the it doesn't get > displayed as a markdown. > example attached Do any of the same things you would do to execute a code cell - press shift-enter or click the run button in the toolbar. This will render your markdown. Again, it's deliberate that you can leave a markdown cell without rendering it. > I think a video on new features and how to use them would be very helpful. Thanks, we'll think about that. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From apratap at sagebase.org Thu Mar 13 16:32:27 2014 From: apratap at sagebase.org (Abhishek Pratap) Date: Thu, 13 Mar 2014 13:32:27 -0700 Subject: [IPython-dev] shortcuts for adding/deleting cells In-Reply-To: References: Message-ID: On Thu, Mar 13, 2014 at 12:27 PM, Thomas Kluyver wrote: > On 13 March 2014 12:20, Abhishek Pratap wrote: >> >> For example: >> in the (command mode with grey cell) >> >> CTRL + a doesnt insert a new cell > > > It's not Ctrl-a, it's just a. When you're in command mode, most of the > keyboard shortcuts are single keys pressed by themselves, like gmail > keyboard shortcuts. The reason we've done this is because almost all Ctrl > shortcuts are already used for something by the browser, and we don't want > to conflict with them. > >> >> Also once I convert the cell to a markdown type the it doesn't get >> displayed as a markdown. >> example attached > > > Do any of the same things you would do to execute a code cell - press > shift-enter or click the run button in the toolbar. This will render your > markdown. Again, it's deliberate that you can leave a markdown cell without > rendering it. > >> I think a video on new features and how to use them would be very > helpful. > > Thanks, we'll think about that. > > Thomas > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > Whole lot easier after few minutes of practice. Also from the help menu keyboard shortcuts it is sort of confusing which commands are not to be paired with ctrl or shift. My assumption was that anything after ctrl+j follows the same convention. It cud just be my naivety. -Abhi From doug.blank at gmail.com Fri Mar 14 02:01:32 2014 From: doug.blank at gmail.com (Doug Blank) Date: Fri, 14 Mar 2014 02:01:32 -0400 Subject: [IPython-dev] Seeing javascript visualization from nbviewer? Message-ID: I am working on migrating a Google visualization example [1] by victor_zverovich to a simpler version shown here: http://nbviewer.ipython.org/urls/bitbucket.org/ipre/calico/raw/master/notebooks/GeoChart.ipynb (users using the regular IPython Python kernel can just change calico.Javascript to Javascript and import it from IPython.display) What do I need to do be able to have the GeoChart render in nbviewer? Is that possible? Currently I get the same error I get before I execute the cells when I first open the notebook: """ Javascript error adding output! ReferenceError: google is not defined """ Any insight appreciated! -Doug [1] - http://zverovich.net/2013/06/27/visualizing-geographical-ampl-data-using-ipython-and-google-charts.html From aaron.oleary at gmail.com Fri Mar 14 03:57:41 2014 From: aaron.oleary at gmail.com (Aaron O'Leary) Date: Fri, 14 Mar 2014 07:57:41 +0000 Subject: [IPython-dev] shortcuts for adding/deleting cells In-Reply-To: References: Message-ID: > it is sort of confusing which commands are not > to be paired with ctrl or shift. My assumption was that anything after > ctrl+j follows the same convention. I didn't figure this out immediately either. I thought it was my window manager intercepting the commands for a while. From bussonniermatthias at gmail.com Fri Mar 14 04:21:39 2014 From: bussonniermatthias at gmail.com (Matthias Bussonnier) Date: Fri, 14 Mar 2014 09:21:39 +0100 Subject: [IPython-dev] shortcuts for adding/deleting cells In-Reply-To: References: Message-ID: <6817148D-30BC-4DE9-8465-55832E0576FB@gmail.com> That one of the reason we avoid modifiers. Most of the ctrl+smth are intercepted by os/window manager/browser ... Etc. Envoy? de mon iPhone Le 14 mars 2014 ? 08:57, "Aaron O'Leary" a ?crit : >> it is sort of confusing which commands are not >> to be paired with ctrl or shift. My assumption was that anything after >> ctrl+j follows the same convention. > > I didn't figure this out immediately either. I thought it was my > window manager intercepting the commands for a while. > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From maidos93 at laposte.net Fri Mar 14 05:39:21 2014 From: maidos93 at laposte.net (thwiouz) Date: Fri, 14 Mar 2014 02:39:21 -0700 (PDT) Subject: [IPython-dev] Question on remote kernels Message-ID: <1394789961867-5050401.post@n6.nabble.com> Hi guys, I would like to launch kernels on distance machines in the IPython notebook web app. Is it possible to configure the different addresses to launch the kernels in some other server than the one which is currently lanuching the webapp? I know a little bit about IPy cluster but I would prefer to integrate it directly in the app. Is that possible? Thanks in advance, -- View this message in context: http://python.6.x6.nabble.com/Question-on-remote-kernels-tp5050401.html Sent from the IPython - Development mailing list archive at Nabble.com. From aaron.oleary at gmail.com Fri Mar 14 05:52:47 2014 From: aaron.oleary at gmail.com (Aaron O'Leary) Date: Fri, 14 Mar 2014 09:52:47 +0000 Subject: [IPython-dev] Question on remote kernels In-Reply-To: <1394789961867-5050401.post@n6.nabble.com> References: <1394789961867-5050401.post@n6.nabble.com> Message-ID: > I would like to launch kernels on distance machines in the IPython notebook > web app. Not exactly what you're asking for but I start a notebook on the remote machine and connect to it with a ssh tunnel. Then I can work with notebooks in the local browser. On linux, using ssh with keys: # create tunnel to remote-machine on port 8898 ssh -X -N -f -L localhost:8898:localhost:8898 remote-machine # command to start notebook with notebook_up='ipython notebook --no-browser --port=8898 --notebook-dir=/path/to/notebooks &' # start notebook on remote-machine ssh -f remote-machine "${notebook_up}" From jr at sun.ac.za Fri Mar 14 06:01:33 2014 From: jr at sun.ac.za (Johann Rohwer) Date: Fri, 14 Mar 2014 12:01:33 +0200 Subject: [IPython-dev] Deploying IPython on Windows campus network In-Reply-To: <53207559.6070104@vcu.edu> References: <3213443.a10FyKgLBC@biochem433789> <53207559.6070104@vcu.edu> Message-ID: <1452920.J6eeq67OQ1@biochem433789> Hi Chris Thanks for this, we got it working in the end. There were other issues to solve with the virtualization, but at least the path is now set where it should be. Johann On Wednesday 12 March 2014 10:55:21 Chris Friedline wrote: > Hi Johann, > > That path is controlled by the environment variable, IPYTHONDIR, which > your windows admin could set on the boxes via Group Policy. > > See here for more info: http://ipython.org/ipython-doc/dev/config/intro.html > > HTH, > Chris > > > We are trying to deploy IPython on our Windows campus network using > > Microsoft Application Virtualization technology. On our campus network > > all users have a home directory which is automounted as the network drive > > H:\ upon login. What is the easiest way to tell IPython to use > > H:\.ipython or H:\ipython as the configuration directory for IPython? (If > > I am not mistaken the home drive is set by the environment variable > > %HOMEDRIVE% under Windows). > > > > Unfortunately I am not very familiar with Windows (using Linux myself) and > > our Windows sysadmin has limited python knowledge. Currently our > > installation is working, but upon startup the current working directory > > is set to > > C:\Windows\system32 which is not writeable for normal users - this is > > expecially problematic for the notebook which won't start up because of a > > lack of write permission. Also the IPython config directory defaults to > > C: \Users\%USERNAME%\.ipython which is machine specific, and we would > > like it to be under a user's network home directory. > > > > We are using the Anaconda python distribution. > > > > Any pointers would be appreciated. > > Regards, > > Johann > > The integrity and confidentiality of this email is governed by these terms / Hierdie terme bepaal die integriteit en vertroulikheid van hierdie epos. http://www.sun.ac.za/emaildisclaimer From maidos93 at laposte.net Fri Mar 14 06:05:33 2014 From: maidos93 at laposte.net (thwiouz) Date: Fri, 14 Mar 2014 03:05:33 -0700 (PDT) Subject: [IPython-dev] Question on remote kernels In-Reply-To: References: <1394789961867-5050401.post@n6.nabble.com> Message-ID: <1394791533704-5050404.post@n6.nabble.com> Thanks, but I would like to put my webapp in a server machine that wouldn't execute the Python code on itself but on other ones. But thanks again, -- View this message in context: http://python.6.x6.nabble.com/Question-on-remote-kernels-tp5050401p5050404.html Sent from the IPython - Development mailing list archive at Nabble.com. From alimanfoo at googlemail.com Fri Mar 14 06:10:18 2014 From: alimanfoo at googlemail.com (Alistair Miles) Date: Fri, 14 Mar 2014 10:10:18 +0000 Subject: [IPython-dev] throttling widget events Message-ID: Hi all, I've spent a couple of days working with IPython 2.0.0-b1 and am having a lot of fun with interact, thank you for this totally cool feature. Pretty much every plotting function I write now I find myself trying to find ways of making it interactive, just so I can play around with it :-) One thing I'm wondering, I remember from one of Brian's talks that the widget framework throttles the events being sent when, e.g., a slider is being moved, so the kernel is not totally swamped. Even still, I have some plotting functions I'd like to interact with which are relatively expensive to compute (e.g., >1s) and so I'd like to throttle the events even further. Is this possible, and if so is it documented anywhere? Thanks, Alistair -- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health The Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Web: http://purl.org/net/aliman Email: alimanfoo at gmail.com Tel: +44 (0)1865 287721 ***new number*** -------------- next part -------------- An HTML attachment was scrubbed... URL: From alimanfoo at googlemail.com Fri Mar 14 06:32:01 2014 From: alimanfoo at googlemail.com (Alistair Miles) Date: Fri, 14 Mar 2014 10:32:01 +0000 Subject: [IPython-dev] feedback on modal UI Message-ID: Hi all, In case this is helpful, here is some feedback on the new modal UI after a couple of days working with IPython 2.0.0-b1. I'm sure I will get used to it (I had become very fluent with the old set of shortcut keys) but for now I am having a lot of accidents from starting to type and forgetting I'm still in command mode. I've lost count of the number of times I've accidentally turned line numbers on or off, randomly added new cells above or below, switched the cell from code to markdown, etc. Often I've got through typing a number of words before realising something is wrong, and so found several unintended things have happened to my notebook. Fortunately there aren't many words in English with a double d, that could be particularly confusing, maybe deleting cells should be double [delete] key? Also having cut as a plain 'x' press is similar, maybe cut/copy/paste/undo should need a control key modifier? I know these problems may only be transient and when I'm fully used to the new UI they'll happen much less. I know this has been mentioned before, but think it might help if the visual distinction between edit mode and command mode is made stronger. My eyes literally can't tell the difference between the green outline and the grey outline on most screens. I'm sure others will have better ideas, but maybe the box outline for edit mode could be a couple of points thicker, or the cell background could be a different colour/shade of grey/white, or maybe lose the outline completely in command mode and indicate the selected cell by some other means? I'm not a visual designer, I'm sure there are better ways. Hth, thanks for the revolutionary work! Alistair -- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health The Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Web: http://purl.org/net/aliman Email: alimanfoo at gmail.com Tel: +44 (0)1865 287721 ***new number*** -------------- next part -------------- An HTML attachment was scrubbed... URL: From alimanfoo at googlemail.com Fri Mar 14 06:44:56 2014 From: alimanfoo at googlemail.com (Alistair Miles) Date: Fri, 14 Mar 2014 10:44:56 +0000 Subject: [IPython-dev] feedback on modal UI In-Reply-To: References: Message-ID: P.S., adding this: div.cell.edit_mode { border: 5px solid red; background-color: #faa; } ...to custom.css is my temporary solution to help me remember to enter edit mode before typing :-) On Fri, Mar 14, 2014 at 10:32 AM, Alistair Miles wrote: > Hi all, > > In case this is helpful, here is some feedback on the new modal UI after a > couple of days working with IPython 2.0.0-b1. > > I'm sure I will get used to it (I had become very fluent with the old set > of shortcut keys) but for now I am having a lot of accidents from starting > to type and forgetting I'm still in command mode. I've lost count of the > number of times I've accidentally turned line numbers on or off, randomly > added new cells above or below, switched the cell from code to markdown, > etc. Often I've got through typing a number of words before realising > something is wrong, and so found several unintended things have happened to > my notebook. Fortunately there aren't many words in English with a double > d, that could be particularly confusing, maybe deleting cells should be > double [delete] key? Also having cut as a plain 'x' press is similar, maybe > cut/copy/paste/undo should need a control key modifier? > > I know these problems may only be transient and when I'm fully used to the > new UI they'll happen much less. > > I know this has been mentioned before, but think it might help if the > visual distinction between edit mode and command mode is made stronger. My > eyes literally can't tell the difference between the green outline and the > grey outline on most screens. I'm sure others will have better ideas, but > maybe the box outline for edit mode could be a couple of points thicker, or > the cell background could be a different colour/shade of grey/white, or > maybe lose the outline completely in command mode and indicate the selected > cell by some other means? I'm not a visual designer, I'm sure there are > better ways. > > Hth, thanks for the revolutionary work! > > Alistair > -- > Alistair Miles > Head of Epidemiological Informatics > Centre for Genomics and Global Health > The Wellcome Trust Centre for Human Genetics > Roosevelt Drive > Oxford > OX3 7BN > United Kingdom > Web: http://purl.org/net/aliman > Email: alimanfoo at gmail.com > Tel: +44 (0)1865 287721 ***new number*** > -- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health The Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Web: http://purl.org/net/aliman Email: alimanfoo at gmail.com Tel: +44 (0)1865 287721 ***new number*** -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Fri Mar 14 08:45:59 2014 From: doug.blank at gmail.com (Doug Blank) Date: Fri, 14 Mar 2014 08:45:59 -0400 Subject: [IPython-dev] Seeing javascript visualization from nbviewer? In-Reply-To: References: Message-ID: I think the problem was with my implementation of Javascript(.., lib=...). I see that the lib is used like this in the Python kernel: $.getScript("https://www.google.com/jsapi", function () { [code] } whereas I had: require["https://www.google.com/jsapi"]; [code]; Which did work, but only when running the cells. Is putting Javascript(lib="") the best method for including external Javascript libraries that are used often/many times? Or is there something I should add to custom.js that would load it at the right time? -Doug On Fri, Mar 14, 2014 at 2:01 AM, Doug Blank wrote: > I am working on migrating a Google visualization example [1] by > victor_zverovich to a simpler version shown here: > > http://nbviewer.ipython.org/urls/bitbucket.org/ipre/calico/raw/master/notebooks/GeoChart.ipynb > > (users using the regular IPython Python kernel can just change > calico.Javascript to Javascript and import it from IPython.display) > > What do I need to do be able to have the GeoChart render in nbviewer? > Is that possible? Currently I get the same error I get before I > execute the cells when I first open the notebook: > > """ > Javascript error adding output! > ReferenceError: google is not defined > """ > > Any insight appreciated! > > -Doug > > > [1] - http://zverovich.net/2013/06/27/visualizing-geographical-ampl-data-using-ipython-and-google-charts.html From alimanfoo at googlemail.com Fri Mar 14 10:58:56 2014 From: alimanfoo at googlemail.com (Alistair Miles) Date: Fri, 14 Mar 2014 14:58:56 +0000 Subject: [IPython-dev] feedback on modal UI In-Reply-To: References: Message-ID: Sorry for spam, in case it's useful for anyone else here's a couple more CSS tweaks which are helping me with the modal UI... div.cell.selected { border: 3px #ababab solid; background-color: #ddd; } div.cell.edit_mode { border: 5px red solid; background-color: #faa; } div.cell.selected div.output_subarea { background-color: #fff; margin: .5em 0 .5em 0; padding: .5em; border: 1px #ababab solid; border-radius: 4px; } div.cell.edit_mode div.output_subarea { border: 1px red solid; border-radius: 4px; } div.cell.edit_mode div.input_area { border: 1px red solid; } On Fri, Mar 14, 2014 at 10:44 AM, Alistair Miles wrote: > P.S., adding this: > > div.cell.edit_mode { > border: 5px solid red; > background-color: #faa; > } > > ...to custom.css is my temporary solution to help me remember to enter > edit mode before typing :-) > > > On Fri, Mar 14, 2014 at 10:32 AM, Alistair Miles > wrote: > >> Hi all, >> >> In case this is helpful, here is some feedback on the new modal UI after >> a couple of days working with IPython 2.0.0-b1. >> >> I'm sure I will get used to it (I had become very fluent with the old set >> of shortcut keys) but for now I am having a lot of accidents from starting >> to type and forgetting I'm still in command mode. I've lost count of the >> number of times I've accidentally turned line numbers on or off, randomly >> added new cells above or below, switched the cell from code to markdown, >> etc. Often I've got through typing a number of words before realising >> something is wrong, and so found several unintended things have happened to >> my notebook. Fortunately there aren't many words in English with a double >> d, that could be particularly confusing, maybe deleting cells should be >> double [delete] key? Also having cut as a plain 'x' press is similar, maybe >> cut/copy/paste/undo should need a control key modifier? >> >> I know these problems may only be transient and when I'm fully used to >> the new UI they'll happen much less. >> >> I know this has been mentioned before, but think it might help if the >> visual distinction between edit mode and command mode is made stronger. My >> eyes literally can't tell the difference between the green outline and the >> grey outline on most screens. I'm sure others will have better ideas, but >> maybe the box outline for edit mode could be a couple of points thicker, or >> the cell background could be a different colour/shade of grey/white, or >> maybe lose the outline completely in command mode and indicate the selected >> cell by some other means? I'm not a visual designer, I'm sure there are >> better ways. >> >> Hth, thanks for the revolutionary work! >> >> Alistair >> -- >> Alistair Miles >> Head of Epidemiological Informatics >> Centre for Genomics and Global Health >> The Wellcome Trust Centre for Human Genetics >> Roosevelt Drive >> Oxford >> OX3 7BN >> United Kingdom >> Web: http://purl.org/net/aliman >> Email: alimanfoo at gmail.com >> Tel: +44 (0)1865 287721 ***new number*** >> > > > > -- > Alistair Miles > Head of Epidemiological Informatics > Centre for Genomics and Global Health > The Wellcome Trust Centre for Human Genetics > Roosevelt Drive > Oxford > OX3 7BN > United Kingdom > Web: http://purl.org/net/aliman > Email: alimanfoo at gmail.com > Tel: +44 (0)1865 287721 ***new number*** > -- Alistair Miles Head of Epidemiological Informatics Centre for Genomics and Global Health The Wellcome Trust Centre for Human Genetics Roosevelt Drive Oxford OX3 7BN United Kingdom Web: http://purl.org/net/aliman Email: alimanfoo at gmail.com Tel: +44 (0)1865 287721 ***new number*** -------------- next part -------------- An HTML attachment was scrubbed... URL: From ellisonbg at gmail.com Fri Mar 14 11:28:51 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Fri, 14 Mar 2014 08:28:51 -0700 Subject: [IPython-dev] feedback on modal UI In-Reply-To: References: Message-ID: Thanks for sharing this. I think we prefer to keep the default styling of this extremely simple and encourage users to customize to suite their needs. Your feedback is helpful though - we will monitor the situation. Cheers, Brian On Fri, Mar 14, 2014 at 7:58 AM, Alistair Miles wrote: > Sorry for spam, in case it's useful for anyone else here's a couple more CSS > tweaks which are helping me with the modal UI... > > div.cell.selected { > border: 3px #ababab solid; > background-color: #ddd; > } > > div.cell.edit_mode { > border: 5px red solid; > background-color: #faa; > } > > div.cell.selected div.output_subarea { > background-color: #fff; > margin: .5em 0 .5em 0; > padding: .5em; > border: 1px #ababab solid; > border-radius: 4px; > } > > div.cell.edit_mode div.output_subarea { > border: 1px red solid; > border-radius: 4px; > } > > div.cell.edit_mode div.input_area { > border: 1px red solid; > } > > > > On Fri, Mar 14, 2014 at 10:44 AM, Alistair Miles > wrote: >> >> P.S., adding this: >> >> div.cell.edit_mode { >> border: 5px solid red; >> background-color: #faa; >> } >> >> ...to custom.css is my temporary solution to help me remember to enter >> edit mode before typing :-) >> >> >> On Fri, Mar 14, 2014 at 10:32 AM, Alistair Miles >> wrote: >>> >>> Hi all, >>> >>> In case this is helpful, here is some feedback on the new modal UI after >>> a couple of days working with IPython 2.0.0-b1. >>> >>> I'm sure I will get used to it (I had become very fluent with the old set >>> of shortcut keys) but for now I am having a lot of accidents from starting >>> to type and forgetting I'm still in command mode. I've lost count of the >>> number of times I've accidentally turned line numbers on or off, randomly >>> added new cells above or below, switched the cell from code to markdown, >>> etc. Often I've got through typing a number of words before realising >>> something is wrong, and so found several unintended things have happened to >>> my notebook. Fortunately there aren't many words in English with a double d, >>> that could be particularly confusing, maybe deleting cells should be double >>> [delete] key? Also having cut as a plain 'x' press is similar, maybe >>> cut/copy/paste/undo should need a control key modifier? >>> >>> I know these problems may only be transient and when I'm fully used to >>> the new UI they'll happen much less. >>> >>> I know this has been mentioned before, but think it might help if the >>> visual distinction between edit mode and command mode is made stronger. My >>> eyes literally can't tell the difference between the green outline and the >>> grey outline on most screens. I'm sure others will have better ideas, but >>> maybe the box outline for edit mode could be a couple of points thicker, or >>> the cell background could be a different colour/shade of grey/white, or >>> maybe lose the outline completely in command mode and indicate the selected >>> cell by some other means? I'm not a visual designer, I'm sure there are >>> better ways. >>> >>> Hth, thanks for the revolutionary work! >>> >>> Alistair >>> -- >>> Alistair Miles >>> Head of Epidemiological Informatics >>> Centre for Genomics and Global Health >>> The Wellcome Trust Centre for Human Genetics >>> Roosevelt Drive >>> Oxford >>> OX3 7BN >>> United Kingdom >>> Web: http://purl.org/net/aliman >>> Email: alimanfoo at gmail.com >>> Tel: +44 (0)1865 287721 ***new number*** >> >> >> >> >> -- >> Alistair Miles >> Head of Epidemiological Informatics >> Centre for Genomics and Global Health >> The Wellcome Trust Centre for Human Genetics >> Roosevelt Drive >> Oxford >> OX3 7BN >> United Kingdom >> Web: http://purl.org/net/aliman >> Email: alimanfoo at gmail.com >> Tel: +44 (0)1865 287721 ***new number*** > > > > > -- > Alistair Miles > Head of Epidemiological Informatics > Centre for Genomics and Global Health > The Wellcome Trust Centre for Human Genetics > Roosevelt Drive > Oxford > OX3 7BN > United Kingdom > Web: http://purl.org/net/aliman > Email: alimanfoo at gmail.com > Tel: +44 (0)1865 287721 ***new number*** > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From asmeurer at gmail.com Fri Mar 14 11:29:47 2014 From: asmeurer at gmail.com (Aaron Meurer) Date: Fri, 14 Mar 2014 10:29:47 -0500 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: SymPy's IRC channel has been mostly dead. I was too often the only core member on there. And if I was not online, there was no easy way for me to be notified about things (I did try setting up a bouncer with Colloquy, but I never got it working). The entry barrier for IRC is way too high for new users. You have to install some client, or use some webclient, none of which are very user-friendly. What has happened all the time is that people have gone on there, asked a question, and no one answered (because I was not online basically). This turns people away from the project. Gitter has backlogs, and is super easy. If someone comes online and asks a question and no one is online, I can answer later, @mentioning the person (which will email them). You can look at the Gitter logs for SymPy vs. our IRC logs for the past month or so, and see a huge difference in usage (a lot of that also because I recently promoted Gitter on our list). I agree that IRC is de-facto among "hackers" and a certain class of software engineers, but quite a few developers for SymPy (and IPython too I'm sure) are not part of this class, and very few of our users (the more important group of people) are. I wanted to point out that Gitter does now allow you to create custom rooms. They are owned either by you or any organization you belong to, are namespaced under the owner, and can be public or private (invitation only). That's a pretty nice feature for something that didn't exist a week ago. Aaron Meurer On Tue, Mar 11, 2014 at 11:30 AM, Thomas Kluyver wrote: > On 11 March 2014 05:13, Matthew Turk wrote: >> >> We're a smaller project (by a good amount) than IPython, but we've had >> good luck with the Freenode web gateway for yt: >> >> http://webchat.freenode.net/ > > > For the record: before we switched to hipchat, we did promote the freenode > web version of the #ipython IRC channel with a link on the homepage. > > In addition to the reasons Min mentioned, Hipchat can handle images inline, > which is handy for screenshots as the project has more of a visual > component. @mentions are the nicest thing, though - if I see a question in > the help room that I can't answer, but I know who can, I can ping them, and > most of the time, they log in a few seconds later. > > >> On surveys, and communication in general, could you not just add a feature >> to the Notebook that allows it to connect to a central server onload, so it >> can let users know there's some news and lets them load the content >> directly? Users should have an obvious, one click toggle to disable the >> connections, but few would. > > We have actually talked about having some kind of statistics gathering, on a > strictly opt in basis. Delivering information would be an interesting > extension of that. > > Thanks, > Thomas > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From ellisonbg at gmail.com Fri Mar 14 11:32:37 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Fri, 14 Mar 2014 08:32:37 -0700 Subject: [IPython-dev] Seeing javascript visualization from nbviewer? In-Reply-To: References: Message-ID: If the library is aware of require.js, you should use that. But the calling syntax for require is similar to `getScript` is similar - the second argument is a function that gets called when the libraries are loaded. These days, many JavaScript libraries have to be loaded with require.js in the notebook. With IPython 2.0, all different ways should work fine on nbviewer/nbconvert though. On Fri, Mar 14, 2014 at 5:45 AM, Doug Blank wrote: > I think the problem was with my implementation of Javascript(.., > lib=...). I see that the lib is used like this in the Python kernel: > > $.getScript("https://www.google.com/jsapi", function () { > [code] > } > > whereas I had: > > require["https://www.google.com/jsapi"]; > [code]; > > Which did work, but only when running the cells. > > Is putting Javascript(lib="") the best method for including external > Javascript libraries that are used often/many times? Or is there > something I should add to custom.js that would load it at the right > time? > > -Doug > > On Fri, Mar 14, 2014 at 2:01 AM, Doug Blank wrote: >> I am working on migrating a Google visualization example [1] by >> victor_zverovich to a simpler version shown here: >> >> http://nbviewer.ipython.org/urls/bitbucket.org/ipre/calico/raw/master/notebooks/GeoChart.ipynb >> >> (users using the regular IPython Python kernel can just change >> calico.Javascript to Javascript and import it from IPython.display) >> >> What do I need to do be able to have the GeoChart render in nbviewer? >> Is that possible? Currently I get the same error I get before I >> execute the cells when I first open the notebook: >> >> """ >> Javascript error adding output! >> ReferenceError: google is not defined >> """ >> >> Any insight appreciated! >> >> -Doug >> >> >> [1] - http://zverovich.net/2013/06/27/visualizing-geographical-ampl-data-using-ipython-and-google-charts.html > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From ellisonbg at gmail.com Fri Mar 14 11:33:21 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Fri, 14 Mar 2014 08:33:21 -0700 Subject: [IPython-dev] Seeing javascript visualization from nbviewer? In-Reply-To: References: Message-ID: I should note that nbviewer 2.0 is not yet deployed - should be in the next few days. These things don't currently work on the nbviewer that is deployed today. On Fri, Mar 14, 2014 at 8:32 AM, Brian Granger wrote: > If the library is aware of require.js, you should use that. But the > calling syntax for require is similar to `getScript` is similar - the > second argument is a function that gets called when the libraries are > loaded. These days, many JavaScript libraries have to be loaded with > require.js in the notebook. With IPython 2.0, all different ways > should work fine on nbviewer/nbconvert though. > > On Fri, Mar 14, 2014 at 5:45 AM, Doug Blank wrote: >> I think the problem was with my implementation of Javascript(.., >> lib=...). I see that the lib is used like this in the Python kernel: >> >> $.getScript("https://www.google.com/jsapi", function () { >> [code] >> } >> >> whereas I had: >> >> require["https://www.google.com/jsapi"]; >> [code]; >> >> Which did work, but only when running the cells. >> >> Is putting Javascript(lib="") the best method for including external >> Javascript libraries that are used often/many times? Or is there >> something I should add to custom.js that would load it at the right >> time? >> >> -Doug >> >> On Fri, Mar 14, 2014 at 2:01 AM, Doug Blank wrote: >>> I am working on migrating a Google visualization example [1] by >>> victor_zverovich to a simpler version shown here: >>> >>> http://nbviewer.ipython.org/urls/bitbucket.org/ipre/calico/raw/master/notebooks/GeoChart.ipynb >>> >>> (users using the regular IPython Python kernel can just change >>> calico.Javascript to Javascript and import it from IPython.display) >>> >>> What do I need to do be able to have the GeoChart render in nbviewer? >>> Is that possible? Currently I get the same error I get before I >>> execute the cells when I first open the notebook: >>> >>> """ >>> Javascript error adding output! >>> ReferenceError: google is not defined >>> """ >>> >>> Any insight appreciated! >>> >>> -Doug >>> >>> >>> [1] - http://zverovich.net/2013/06/27/visualizing-geographical-ampl-data-using-ipython-and-google-charts.html >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > -- > Brian E. Granger > Cal Poly State University, San Luis Obispo > bgranger at calpoly.edu and ellisonbg at gmail.com -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From doug.blank at gmail.com Fri Mar 14 11:42:17 2014 From: doug.blank at gmail.com (Doug Blank) Date: Fri, 14 Mar 2014 11:42:17 -0400 Subject: [IPython-dev] Seeing javascript visualization from nbviewer? In-Reply-To: References: Message-ID: On Fri, Mar 14, 2014 at 11:33 AM, Brian Granger wrote: > I should note that nbviewer 2.0 is not yet deployed - should be in the > next few days. These things don't currently work on the nbviewer that > is deployed today. Ah, yes, that is a key bit of information; thanks! -Doug > On Fri, Mar 14, 2014 at 8:32 AM, Brian Granger wrote: >> If the library is aware of require.js, you should use that. But the >> calling syntax for require is similar to `getScript` is similar - the >> second argument is a function that gets called when the libraries are >> loaded. These days, many JavaScript libraries have to be loaded with >> require.js in the notebook. With IPython 2.0, all different ways >> should work fine on nbviewer/nbconvert though. >> >> On Fri, Mar 14, 2014 at 5:45 AM, Doug Blank wrote: >>> I think the problem was with my implementation of Javascript(.., >>> lib=...). I see that the lib is used like this in the Python kernel: >>> >>> $.getScript("https://www.google.com/jsapi", function () { >>> [code] >>> } >>> >>> whereas I had: >>> >>> require["https://www.google.com/jsapi"]; >>> [code]; >>> >>> Which did work, but only when running the cells. >>> >>> Is putting Javascript(lib="") the best method for including external >>> Javascript libraries that are used often/many times? Or is there >>> something I should add to custom.js that would load it at the right >>> time? >>> >>> -Doug >>> >>> On Fri, Mar 14, 2014 at 2:01 AM, Doug Blank wrote: >>>> I am working on migrating a Google visualization example [1] by >>>> victor_zverovich to a simpler version shown here: >>>> >>>> http://nbviewer.ipython.org/urls/bitbucket.org/ipre/calico/raw/master/notebooks/GeoChart.ipynb >>>> >>>> (users using the regular IPython Python kernel can just change >>>> calico.Javascript to Javascript and import it from IPython.display) >>>> >>>> What do I need to do be able to have the GeoChart render in nbviewer? >>>> Is that possible? Currently I get the same error I get before I >>>> execute the cells when I first open the notebook: >>>> >>>> """ >>>> Javascript error adding output! >>>> ReferenceError: google is not defined >>>> """ >>>> >>>> Any insight appreciated! >>>> >>>> -Doug >>>> >>>> >>>> [1] - http://zverovich.net/2013/06/27/visualizing-geographical-ampl-data-using-ipython-and-google-charts.html >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> >> >> -- >> Brian E. Granger >> Cal Poly State University, San Luis Obispo >> bgranger at calpoly.edu and ellisonbg at gmail.com > > > > -- > Brian E. Granger > Cal Poly State University, San Luis Obispo > bgranger at calpoly.edu and ellisonbg at gmail.com > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From raymond.yee at gmail.com Fri Mar 14 12:14:45 2014 From: raymond.yee at gmail.com (Raymond Yee) Date: Fri, 14 Mar 2014 09:14:45 -0700 Subject: [IPython-dev] How to get syntax highlight code blocks in the markdown cells in an IPython notebook? Message-ID: <53232AF5.5060300@gmail.com> Is there any easy way to get syntax highlighting for code blocks in IPython notebook markdown cells? (I'm writing a midterm in the notebook and would like to write code in code cells and then move code into markdown cells but retain syntax highlighting.) Thanks, -Raymond From bussonniermatthias at gmail.com Fri Mar 14 12:22:51 2014 From: bussonniermatthias at gmail.com (Matthias BUSSONNIER) Date: Fri, 14 Mar 2014 17:22:51 +0100 Subject: [IPython-dev] How to get syntax highlight code blocks in the markdown cells in an IPython notebook? In-Reply-To: <53232AF5.5060300@gmail.com> References: <53232AF5.5060300@gmail.com> Message-ID: Le 14 mars 2014 ? 17:14, Raymond Yee a ?crit : > Is there any easy way to get syntax highlighting for code blocks in > IPython notebook markdown cells? (I'm writing a midterm in the notebook > and would like to write code in code cells and then move code into > markdown cells but retain syntax highlighting.) ```language blah blah.. ``` should work. -- M > > Thanks, > -Raymond > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From carl.input at gmail.com Fri Mar 14 12:30:01 2014 From: carl.input at gmail.com (Carl Smith) Date: Fri, 14 Mar 2014 16:30:01 +0000 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: +1 on Gitter. I've never personally used IRC. I've tried a couple of times just because it sounds cool, but the clients are all a bit gnarly and the actual experience once you get it working is pretty underwhelming. Obviously, it's something you have to get into, but /expecting/ users to 'jump on IRC' is going to cost you a lot on communication. If you're already invested in GitHub, Gitter extends it cleanly, and Gitter's easy to extend with user scripts if you want to tweak parts of it for a particular project's needs. Tip: Setting `.trpChatInputArea` to have roughly `height:450px` and `.trpChatInputBoxTextArea`'s font family to anything monospaced makes it easier to compose more complex chunks of Markdown. From matthewturk at gmail.com Fri Mar 14 12:34:41 2014 From: matthewturk at gmail.com (Matthew Turk) Date: Fri, 14 Mar 2014 12:34:41 -0400 Subject: [IPython-dev] Results of last year's user survey In-Reply-To: References: Message-ID: On Fri, Mar 14, 2014 at 12:30 PM, Carl Smith wrote: > +1 on Gitter. I've never personally used IRC. I've tried a couple of > times just because it sounds cool, but the clients are all a bit > gnarly and the actual experience once you get it working is pretty > underwhelming. Obviously, it's something you have to get into, but > /expecting/ users to 'jump on IRC' is going to cost you a lot on > communication. > > If you're already invested in GitHub, Gitter extends it cleanly, and > Gitter's easy to extend with user scripts if you want to tweak parts > of it for a particular project's needs. > > Tip: Setting `.trpChatInputArea` to have roughly `height:450px` and > `.trpChatInputBoxTextArea`'s font family to anything monospaced makes > it easier to compose more complex chunks of Markdown. So, it seems like the consensus is that IRC is not a good match, but since I was one of the people who chimed in suggesting it, I wanted to just point out that (after checking) most of the people I talk to regularly use Adium to connect to it. That doesn't solve the backtrace issue, though; Adium would need a proxy for that. Some people on our channel use IRCCloud, but I'm not familiar with it. Anyway, the main reason *we* use it is because people can connect from basically any client -- web, etc etc -- and we have a pretty active set of devs with good timezone distribution, so usually people get answers to their questions right away. And, it looks like Gitter supports an IRC bridge now: http://blog.gitter.im/long-live-irc/ so ... -Matt > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From raymond.yee at gmail.com Fri Mar 14 12:35:07 2014 From: raymond.yee at gmail.com (Raymond Yee) Date: Fri, 14 Mar 2014 09:35:07 -0700 Subject: [IPython-dev] How to edit cell-level metadata in the IPython notebook interface? Message-ID: <53232FBB.4020907@gmail.com> Is there a way to edit cell-level metadata using the IPython notebook interface? I don't see one and I'm guessing there isn't one (yet). BTW, is https://github.com/ipython/ipython/wiki/IPEP-20%3A-Informal-structure-of-cell-metadata a pretty good summary of where cell-level metadata is heading? Thanks, -Raymond From j.davidgriffiths at gmail.com Fri Mar 14 12:34:54 2014 From: j.davidgriffiths at gmail.com (John Griffiths) Date: Fri, 14 Mar 2014 16:34:54 +0000 Subject: [IPython-dev] How to get syntax highlight code blocks in the markdown cells in an IPython notebook? In-Reply-To: <53232AF5.5060300@gmail.com> References: <53232AF5.5060300@gmail.com> Message-ID: e.g. ```python daveshouse = '10 Downing Street' def get_dave_arrested(daveshouse): import drugs print 'Get your stash at %s' davesshouse drugs.stash_at(daveshouse) ``` On 14 March 2014 16:14, Raymond Yee wrote: > Is there any easy way to get syntax highlighting for code blocks in > IPython notebook markdown cells? (I'm writing a midterm in the notebook > and would like to write code in code cells and then move code into > markdown cells but retain syntax highlighting.) > > Thanks, > -Raymond > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Mr. John Griffiths, MSc PhD Candidate Centre for Speech, Language, and the Brain Department of Experimental Psychology University of Cambridge, UK -------------- next part -------------- An HTML attachment was scrubbed... URL: From patrick.surry at gmail.com Fri Mar 14 12:35:11 2014 From: patrick.surry at gmail.com (Patrick Surry) Date: Fri, 14 Mar 2014 12:35:11 -0400 Subject: [IPython-dev] Loop over a set of notebook cells? Message-ID: A common pattern for me is to do some data analysis w/ tables and plots via a sequence of cells in a notebook. Then I want to come back and repeat the same analysis for a (small number of) different versions or subsets of the same dataset. There's a related question on stackoverflow I could save the whole notebook and run it from the command-line programmatically, but that breaks the interactive/interactive approach (e.g. when one of my variants ends up going down a slightly different track). Or I could reorganize the notebook so it does everything in one huge cell but that makes multiple plots and tables challenging. Feels like I want some looping magic to loop over the next N cells, or at least some kind of block copy & paste or something? Am I missing something? Cheers, Patrick -------------- next part -------------- An HTML attachment was scrubbed... URL: From ntezak at stanford.edu Fri Mar 14 12:46:29 2014 From: ntezak at stanford.edu (Nikolas Tezak) Date: Fri, 14 Mar 2014 09:46:29 -0700 Subject: [IPython-dev] Loop over a set of notebook cells? In-Reply-To: References: Message-ID: <8235C7D8-41A7-4484-AB5F-091330E2DAD0@stanford.edu> I?ve been thinking about this for a while as well. Although this doesn?t quite address looping, I think one idea might be to use cell tags to create certain ?run sets? of cells and then have a little widget/ui component that allows for selecting and running all cells of a certain tag. So, in that respect a good interface for marking/selecting multiple cells and then tagging them or running them or copying them would be great. Best, Nikolas On Mar 14, 2014, at 9:35 AM, Patrick Surry wrote: > A common pattern for me is to do some data analysis w/ tables and plots via a sequence of cells in a notebook. Then I want to come back and repeat the same analysis for a (small number of) different versions or subsets of the same dataset. > > There's a related question on stackoverflow > > I could save the whole notebook and run it from the command-line programmatically, but that breaks the interactive/interactive approach (e.g. when one of my variants ends up going down a slightly different track). > > Or I could reorganize the notebook so it does everything in one huge cell but that makes multiple plots and tables challenging. > > Feels like I want some looping magic to loop over the next N cells, or at least some kind of block copy & paste or something? > > Am I missing something? > > Cheers, > Patrick > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From benjaminrk at gmail.com Fri Mar 14 13:20:28 2014 From: benjaminrk at gmail.com (MinRK) Date: Fri, 14 Mar 2014 10:20:28 -0700 Subject: [IPython-dev] How to edit cell-level metadata in the IPython notebook interface? In-Reply-To: <53232FBB.4020907@gmail.com> References: <53232FBB.4020907@gmail.com> Message-ID: On Fri, Mar 14, 2014 at 9:35 AM, Raymond Yee wrote: > Is there a way to edit cell-level metadata using the IPython notebook > interface? I don't see one and I'm guessing there isn't one (yet). > Cell Toolbar > Edit Metadata > > BTW, is > > https://github.com/ipython/ipython/wiki/IPEP-20%3A-Informal-structure-of-cell-metadata > a pretty good summary of where cell-level metadata is heading? > > Thanks, > -Raymond > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Fri Mar 14 13:29:31 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Fri, 14 Mar 2014 10:29:31 -0700 Subject: [IPython-dev] throttling widget events In-Reply-To: References: Message-ID: Hi Alistair, On 14 March 2014 03:10, Alistair Miles wrote: > I've spent a couple of days working with IPython 2.0.0-b1 and am having a > lot of fun with interact, thank you for this totally cool feature. Pretty > much every plotting function I write now I find myself trying to find ways > of making it interactive, just so I can play around with it :-) > > One thing I'm wondering, I remember from one of Brian's talks that the > widget framework throttles the events being sent when, e.g., a slider is > being moved, so the kernel is not totally swamped. Even still, I have some > plotting functions I'd like to interact with which are relatively expensive > to compute (e.g., >1s) and so I'd like to throttle the events even further. > Is this possible, and if so is it documented anywhere? > Yes, you can set the w.msg_throttle parameter on your widget. It is the number of messages that the frontend will queue for processing. The default is 3. Turning it down will throttle more aggressively, and turning it up will allow more traffic to be sent. I'm not sure if this is documented anywhere, though. For @interact, the widget object is accessible as func.widget. Or you can use interactive() to make a widget object without displaying it. You may also need to go into the container widget and set throttling on the individual widgets. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Fri Mar 14 13:39:07 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Fri, 14 Mar 2014 10:39:07 -0700 Subject: [IPython-dev] Loop over a set of notebook cells? In-Reply-To: <8235C7D8-41A7-4484-AB5F-091330E2DAD0@stanford.edu> References: <8235C7D8-41A7-4484-AB5F-091330E2DAD0@stanford.edu> Message-ID: On 14 March 2014 09:46, Nikolas Tezak wrote: > I've been thinking about this for a while as well. Although this doesn't > quite address looping, I think one idea might be to use cell tags to create > certain "run sets" of cells and then have a little widget/ui component that > allows for selecting and running all cells of a certain tag. So, in that > respect a good interface for marking/selecting multiple cells and then > tagging them or running them or copying them would be great. Tags on cells is something that we want to do - there are a lot of useful things that could be done with it. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Fri Mar 14 14:16:56 2014 From: doug.blank at gmail.com (Doug Blank) Date: Fri, 14 Mar 2014 14:16:56 -0400 Subject: [IPython-dev] Javascript-based printed output? Message-ID: It looks like there isn't much support in IPython for creating printed versions of notebooks with Javascript-based output... is that correct? * The notebook menu -> File -> Print option doesn't show the Javascript output * Under Chrome, right-click on notebook, select Print...: gives the whole notebook on a single sheet of paper in a scrolled box; seems like a print.css should do better * "ipython nbconvert --to html" doesn't show the output at all * notebook, menu -> File -> Download as... html, doesn't show it I couldn't figure out anything to get a printout of the Javascript output. Maybe the new nbconvert will be the way, when it is released? -Doug From ntezak at stanford.edu Fri Mar 14 14:27:58 2014 From: ntezak at stanford.edu (Nikolas Tezak) Date: Fri, 14 Mar 2014 11:27:58 -0700 Subject: [IPython-dev] Widget View post render hook Message-ID: <76E452D7-1098-4E0D-8DC6-B5E831E6DAD2@stanford.edu> Hi everybody, first of all, I _love_ the widgets, they are extremely useful!! And I?m super impressed by IPython?s crazy fast development progress. I am currently trying to create a custom widget type that embeds a jsplumb circuit editor ( http://jsplumbtoolkit.com/demo/home/jquery.html ) However, when I implement my custom ?render? method for the view I am running into a problem. I first dynamically create the circuit components (currently just absolutely positioned divs within the widget view?s this.$el div. This works fine. I then initialize a jsplumb instance and pass this.$el as the default container. Now, however, when I try to add the connectors and connections with jsplumb within the render method, it fails because jsplumb is trying to read out dynamically the position of the widget view?s div before that div has been inserted into the DOM. So basically, my problem is that I want code within ?render()? that already requires the view to be inserted into the document. Should I use a different approach, i.e., maybe put this initialization code into "update()?, or do you expect that this will be a common enough use case that it would justify adding a special post render hook that gets called after the view?s div has been attached to a cell's widget sub_area? I hope I described this well enough, let me know if you?d like me to share a code example. Thanks, Nik From takowl at gmail.com Fri Mar 14 14:44:51 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Fri, 14 Mar 2014 11:44:51 -0700 Subject: [IPython-dev] Javascript-based printed output? In-Reply-To: References: Message-ID: On 14 March 2014 11:16, Doug Blank wrote: > * The notebook menu -> File -> Print option doesn't show the Javascript > output > > * Under Chrome, right-click on notebook, select Print...: gives the > whole notebook on a single sheet of paper in a scrolled box; seems > like a print.css should do better > > * "ipython nbconvert --to html" doesn't show the output at all > > * notebook, menu -> File -> Download as... html, doesn't show it To save you time when investigating things like this: the notebook's own print view, download as html, and nbconvert at the command line, are all doing essentially the same thing: using nbconvert to produce convert the notebook to static HTML. Best wishes, Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From ellisonbg at gmail.com Fri Mar 14 14:47:58 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Fri, 14 Mar 2014 11:47:58 -0700 Subject: [IPython-dev] Javascript-based printed output? In-Reply-To: References: Message-ID: You should be able to customize nbconvert to show the JS output as plain text using our template system and the display priority. On Fri, Mar 14, 2014 at 11:16 AM, Doug Blank wrote: > It looks like there isn't much support in IPython for creating printed > versions of notebooks with Javascript-based output... is that correct? > > * The notebook menu -> File -> Print option doesn't show the Javascript output > > * Under Chrome, right-click on notebook, select Print...: gives the > whole notebook on a single sheet of paper in a scrolled box; seems > like a print.css should do better > > * "ipython nbconvert --to html" doesn't show the output at all > > * notebook, menu -> File -> Download as... html, doesn't show it > > I couldn't figure out anything to get a printout of the Javascript > output. Maybe the new nbconvert will be the way, when it is released? > > -Doug > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From sylvain.corlay at gmail.com Fri Mar 14 15:40:09 2014 From: sylvain.corlay at gmail.com (Sylvain Corlay) Date: Fri, 14 Mar 2014 15:40:09 -0400 Subject: [IPython-dev] feedback on modal UI In-Reply-To: References: Message-ID: Hello, Some people trying the notebook environment for the first time around me got confused by the modal UI, just because they did not notice right away the green/gray line around the cell. Therefore we also decided to make it a bit thicker. Another 2 cents about the user interface: Some first-time users are confused by the fact that the "run current cell" button only runs one cell rather than the entire notebook. How about using the "step-forward" font-awesome icon for this button? Best, Sylvain Thanks for sharing this. I think we prefer to keep the default styling of this extremely simple and encourage users to customize to suite their needs. Your feedback is helpful though - we will monitor the situation. Cheers, Brian On Fri, Mar 14, 2014 at 7:58 AM, Alistair Miles wrote: > Sorry for spam, in case it's useful for anyone else here's a couple more CSS > tweaks which are helping me with the modal UI... > > div.cell.selected { > border: 3px #ababab solid; > background-color: #ddd; > } > > div.cell.edit_mode { > border: 5px red solid; > background-color: #faa; > } > > div.cell.selected div.output_subarea { > background-color: #fff; > margin: .5em 0 .5em 0; > padding: .5em; > border: 1px #ababab solid; > border-radius: 4px; > } > > div.cell.edit_mode div.output_subarea { > border: 1px red solid; > border-radius: 4px; > } > > div.cell.edit_mode div.input_area { > border: 1px red solid; > } > > > > On Fri, Mar 14, 2014 at 10:44 AM, Alistair Miles > wrote: >> >> P.S., adding this: >> >> div.cell.edit_mode { >> border: 5px solid red; >> background-color: #faa; >> } >> >> ...to custom.css is my temporary solution to help me remember to enter >> edit mode before typing :-) >> >> >> On Fri, Mar 14, 2014 at 10:32 AM, Alistair Miles >> wrote: >>> >>> Hi all, >>> >>> In case this is helpful, here is some feedback on the new modal UI after >>> a couple of days working with IPython 2.0.0-b1. >>> >>> I'm sure I will get used to it (I had become very fluent with the old set >>> of shortcut keys) but for now I am having a lot of accidents from starting >>> to type and forgetting I'm still in command mode. I've lost count of the >>> number of times I've accidentally turned line numbers on or off, randomly >>> added new cells above or below, switched the cell from code to markdown, >>> etc. Often I've got through typing a number of words before realising >>> something is wrong, and so found several unintended things have happened to >>> my notebook. Fortunately there aren't many words in English with a double d, >>> that could be particularly confusing, maybe deleting cells should be double >>> [delete] key? Also having cut as a plain 'x' press is similar, maybe >>> cut/copy/paste/undo should need a control key modifier? >>> >>> I know these problems may only be transient and when I'm fully used to >>> the new UI they'll happen much less. >>> >>> I know this has been mentioned before, but think it might help if the >>> visual distinction between edit mode and command mode is made stronger. My >>> eyes literally can't tell the difference between the green outline and the >>> grey outline on most screens. I'm sure others will have better ideas, but >>> maybe the box outline for edit mode could be a couple of points thicker, or >>> the cell background could be a different colour/shade of grey/white, or >>> maybe lose the outline completely in command mode and indicate the selected >>> cell by some other means? I'm not a visual designer, I'm sure there are >>> better ways. >>> >>> Hth, thanks for the revolutionary work! >>> >>> Alistair >>> -- >>> Alistair Miles >>> Head of Epidemiological Informatics >>> Centre for Genomics and Global Health >>> The Wellcome Trust Centre for Human Genetics >>> Roosevelt Drive >>> Oxford >>> OX3 7BN >>> United Kingdom >>> Web: http://purl.org/net/aliman >>> Email: alimanfoo at gmail.com >>> Tel: +44 (0)1865 287721 ***new number*** >> >> >> >> >> -- >> Alistair Miles >> Head of Epidemiological Informatics >> Centre for Genomics and Global Health >> The Wellcome Trust Centre for Human Genetics >> Roosevelt Drive >> Oxford >> OX3 7BN >> United Kingdom >> Web: http://purl.org/net/aliman >> Email: alimanfoo at gmail.com >> Tel: +44 (0)1865 287721 ***new number*** > > > > > -- > Alistair Miles > Head of Epidemiological Informatics > Centre for Genomics and Global Health > The Wellcome Trust Centre for Human Genetics > Roosevelt Drive > Oxford > OX3 7BN > United Kingdom > Web: http://purl.org/net/aliman > Email: alimanfoo at gmail.com > Tel: +44 (0)1865 287721 ***new number*** > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Fri Mar 14 16:00:59 2014 From: doug.blank at gmail.com (Doug Blank) Date: Fri, 14 Mar 2014 16:00:59 -0400 Subject: [IPython-dev] feedback on modal UI In-Reply-To: References: Message-ID: On Fri, Mar 14, 2014 at 3:40 PM, Sylvain Corlay wrote: > Hello, > > Some people trying the notebook environment for the first time around me got > confused by the modal UI, just because they did not notice right away the > green/gray line around the cell. Therefore we also decided to make it a bit > thicker. +1 ... the only difference I often notice is lack of a cursor! There should a bigger difference. > Another 2 cents about the user interface: Some first-time users are confused > by the fact that the "run current cell" button only runs one cell rather > than the entire notebook. How about using the "step-forward" font-awesome > icon for this button? Another +1 as a good idea, although at least there is a mouse-over for it. -Doug > Best, > Sylvain > > Thanks for sharing this. I think we prefer to keep the default styling > of this extremely simple and encourage users to customize to suite > their needs. Your feedback is helpful though - we will monitor the > situation. > > Cheers, > > Brian > > On Fri, Mar 14, 2014 at 7:58 AM, Alistair Miles > wrote: >> Sorry for spam, in case it's useful for anyone else here's a couple more >> CSS >> tweaks which are helping me with the modal UI... >> >> div.cell.selected { >> border: 3px #ababab solid; >> background-color: #ddd; >> } >> >> div.cell.edit_mode { >> border: 5px red solid; >> background-color: #faa; >> } >> >> div.cell.selected div.output_subarea { >> background-color: #fff; >> margin: .5em 0 .5em 0; >> padding: .5em; >> border: 1px #ababab solid; >> border-radius: 4px; >> } >> >> div.cell.edit_mode div.output_subarea { >> border: 1px red solid; >> border-radius: 4px; >> } >> >> div.cell.edit_mode div.input_area { >> border: 1px red solid; >> } >> >> >> >> On Fri, Mar 14, 2014 at 10:44 AM, Alistair Miles >> >> wrote: >>> >>> P.S., adding this: >>> >>> div.cell.edit_mode { >>> border: 5px solid red; >>> background-color: #faa; >>> } >>> >>> ...to custom.css is my temporary solution to help me remember to enter >>> edit mode before typing :-) >>> >>> >>> On Fri, Mar 14, 2014 at 10:32 AM, Alistair Miles >>> wrote: >>>> >>>> Hi all, >>>> >>>> In case this is helpful, here is some feedback on the new modal UI after >>>> a couple of days working with IPython 2.0.0-b1. >>>> >>>> I'm sure I will get used to it (I had become very fluent with the old >>>> set >>>> of shortcut keys) but for now I am having a lot of accidents from >>>> starting >>>> to type and forgetting I'm still in command mode. I've lost count of the >>>> number of times I've accidentally turned line numbers on or off, >>>> randomly >>>> added new cells above or below, switched the cell from code to markdown, >>>> etc. Often I've got through typing a number of words before realising >>>> something is wrong, and so found several unintended things have happened >>>> to >>>> my notebook. Fortunately there aren't many words in English with a >>>> double d, >>>> that could be particularly confusing, maybe deleting cells should be >>>> double >>>> [delete] key? Also having cut as a plain 'x' press is similar, maybe >>>> cut/copy/paste/undo should need a control key modifier? >>>> >>>> I know these problems may only be transient and when I'm fully used to >>>> the new UI they'll happen much less. >>>> >>>> I know this has been mentioned before, but think it might help if the >>>> visual distinction between edit mode and command mode is made stronger. >>>> My >>>> eyes literally can't tell the difference between the green outline and >>>> the >>>> grey outline on most screens. I'm sure others will have better ideas, >>>> but >>>> maybe the box outline for edit mode could be a couple of points thicker, >>>> or >>>> the cell background could be a different colour/shade of grey/white, or >>>> maybe lose the outline completely in command mode and indicate the >>>> selected >>>> cell by some other means? I'm not a visual designer, I'm sure there are >>>> better ways. >>>> >>>> Hth, thanks for the revolutionary work! >>>> >>>> Alistair >>>> -- >>>> Alistair Miles >>>> Head of Epidemiological Informatics >>>> Centre for Genomics and Global Health >>>> The Wellcome Trust Centre for Human Genetics >>>> Roosevelt Drive >>>> Oxford >>>> OX3 7BN >>>> United Kingdom >>>> Web: http://purl.org/net/aliman >>>> Email: alimanfoo at gmail.com >>>> Tel: +44 (0)1865 287721 ***new number*** >>> >>> >>> >>> >>> -- >>> Alistair Miles >>> Head of Epidemiological Informatics >>> Centre for Genomics and Global Health >>> The Wellcome Trust Centre for Human Genetics >>> Roosevelt Drive >>> Oxford >>> OX3 7BN >>> United Kingdom >>> Web: http://purl.org/net/aliman >>> Email: alimanfoo at gmail.com >>> Tel: +44 (0)1865 287721 ***new number*** >> >> >> >> >> -- >> Alistair Miles >> Head of Epidemiological Informatics >> Centre for Genomics and Global Health >> The Wellcome Trust Centre for Human Genetics >> Roosevelt Drive >> Oxford >> OX3 7BN >> United Kingdom >> Web: http://purl.org/net/aliman >> Email: alimanfoo at gmail.com >> Tel: +44 (0)1865 287721 ***new number*** >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > > > > -- > Brian E. Granger > Cal Poly State University, San Luis Obispo > bgranger at calpoly.edu and ellisonbg at gmail.com > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From claresloggett at gmail.com Sat Mar 15 00:10:24 2014 From: claresloggett at gmail.com (Clare Sloggett) Date: Sat, 15 Mar 2014 15:10:24 +1100 Subject: [IPython-dev] Loop over a set of notebook cells? In-Reply-To: References: <8235C7D8-41A7-4484-AB5F-091330E2DAD0@stanford.edu> Message-ID: On a somewhat related note - I'd really like the ability to execute cells based on the notebook's hierarchical structure, i.e. Heading cells. I've been looking into extending the Table of Contents plugin to add a "run this section" option. That way I could have an Initialisation section, and different sections for different parts of the analysis, and run only the required bits. So this is a bit more limited than what it sounds like you're looking for, since you also want to be able to loop over a set, or to cut and paste. Still, Headings do (conceptually) give you a "set" of cells and these sets can be nested. Being able to cut-and-paste all cells in a section might be a good way to implement what you're aiming at? On 15/03/2014 4:39 AM, "Thomas Kluyver" wrote: > On 14 March 2014 09:46, Nikolas Tezak wrote: > >> I've been thinking about this for a while as well. Although this doesn't >> quite address looping, I think one idea might be to use cell tags to create >> certain "run sets" of cells and then have a little widget/ui component that >> allows for selecting and running all cells of a certain tag. So, in that >> respect a good interface for marking/selecting multiple cells and then >> tagging them or running them or copying them would be great. > > > Tags on cells is something that we want to do - there are a lot of useful > things that could be done with it. > > Thomas > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From patrick.surry at gmail.com Sat Mar 15 07:18:12 2014 From: patrick.surry at gmail.com (Patrick Surry) Date: Sat, 15 Mar 2014 07:18:12 -0400 Subject: [IPython-dev] Loop over a set of notebook cells? Message-ID: What about a "loop delimiter" magic where you insert a matching pair of %magics bracketing a sequence of regular cells, and then it basically unrolls the loop for you? i.e. when you run the magic, on the first loop iteration it executes each cell in place, with the first loop value, and for each subsequent iterations, it appends a copied sequence of new cells after the %end, evaluated with the corresponding loop element. So if you looped over a list of two items, you'd end up with a second copy of all your looped cells that have been executed with the second item in the list. (see example sketch below) So it's doing a kind of repeated block copy & paste (useful in it's own right), unrolling the loop, and leaving you the new results to explore and further tweak. Maybe it would also mark itself as executed somehow (comment itself out?!) so the notebook is stable if you re-execute all cells? You'd rapidly want multi-cell delete I guess :) Perhaps other bracketing magics like %block cut / %endblock / %block paste ?? Patrick Example with cells a, b, c: %begin for item in [1,2] --- a --- b --- c --- %end when executed becomes: # %begin for item in [1, 2] --- a [item=1] --- b [item=1] --- c [item=1] --- # %end --- a [item=2] --- b [item=2] --- c [item=2] --- From: Thomas Kluyver >On 14 March 2014 09:46, Nikolas Tezak wrote: > I've been thinking about this for a while as well. Although this doesn't > quite address looping, I think one idea might be to use cell tags to create > certain "run sets" of cells and then have a little widget/ui component that > allows for selecting and running all cells of a certain tag. So, in that > respect a good interface for marking/selecting multiple cells and then > tagging them or running them or copying them would be great. Tags on cells is something that we want to do - there are a lot of useful things that could be done with it. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From jon.freder at gmail.com Sat Mar 15 16:03:48 2014 From: jon.freder at gmail.com (Jonathan Frederic) Date: Sat, 15 Mar 2014 13:03:48 -0700 Subject: [IPython-dev] Widget View post render hook In-Reply-To: <76E452D7-1098-4E0D-8DC6-B5E831E6DAD2@stanford.edu> References: <76E452D7-1098-4E0D-8DC6-B5E831E6DAD2@stanford.edu> Message-ID: Hi Nik, I'm glad you like the widgets and thank you for the feedback. I encountered the same problem when implementing a D3.js widget (see https://github.com/jdfreder/ipython-d3). As a hack-ish workaround I used a timeout with a 0ms interval ( https://github.com/jdfreder/ipython-d3/blob/master/widget_forcedirectedgraph.js#L13). Right now I don't believe we have a nice way to do this, adding an event or method is probably a good idea. -Jon On Fri, Mar 14, 2014 at 11:27 AM, Nikolas Tezak wrote: > Hi everybody, > > first of all, I _love_ the widgets, they are extremely useful!! And I'm > super impressed by IPython's crazy fast development progress. > I am currently trying to create a custom widget type that embeds a jsplumb > circuit editor > ( http://jsplumbtoolkit.com/demo/home/jquery.html ) > > However, when I implement my custom "render" method for the view I am > running into a problem. > I first dynamically create the circuit components (currently just > absolutely positioned divs within the widget view's this.$el div. This > works fine. > I then initialize a jsplumb instance and pass this.$el as the default > container. > Now, however, when I try to add the connectors and connections with > jsplumb within the render method, it fails because jsplumb is trying to > read out dynamically the position of the widget view's div before that div > has been inserted into the DOM. > > So basically, my problem is that I want code within "render()" that > already requires the view to be inserted into the document. > Should I use a different approach, i.e., maybe put this initialization > code into "update()", or do you expect that this will be a common enough > use case that it would justify adding a special post render hook that gets > called after the view's div has been attached to a cell's widget sub_area? > > > I hope I described this well enough, let me know if you'd like me to share > a code example. > Thanks, > > Nik > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From stefan at sun.ac.za Sun Mar 16 14:56:24 2014 From: stefan at sun.ac.za (=?ISO-8859-1?Q?St=E9fan_van_der_Walt?=) Date: Sun, 16 Mar 2014 20:56:24 +0200 Subject: [IPython-dev] Interesting rendering by nbviewer Message-ID: Hi all, Has anyone seen The Matrix-like renderings of their notebooks recently? http://nbviewer.ipython.org/github/stefanv/teaching/blob/master/2014_ctpug_ipython/ctpug_ipython.ipynb (That does not look normal for me on Google Chrome) The notebook renders fine using (nbconvert from) the latest IPython. St?fan From fperez.net at gmail.com Sun Mar 16 16:47:47 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Sun, 16 Mar 2014 13:47:47 -0700 Subject: [IPython-dev] Interesting rendering by nbviewer In-Reply-To: References: Message-ID: This obviously looks wrong, could you file an nbviewer bug, please? On Sun, Mar 16, 2014 at 11:56 AM, St?fan van der Walt wrote: > Hi all, > > Has anyone seen The Matrix-like renderings of their notebooks recently? > > > http://nbviewer.ipython.org/github/stefanv/teaching/blob/master/2014_ctpug_ipython/ctpug_ipython.ipynb > > (That does not look normal for me on Google Chrome) > > The notebook renders fine using (nbconvert from) the latest IPython. > > St?fan > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From stefan at sun.ac.za Sun Mar 16 16:52:28 2014 From: stefan at sun.ac.za (=?ISO-8859-1?Q?St=E9fan_van_der_Walt?=) Date: Sun, 16 Mar 2014 22:52:28 +0200 Subject: [IPython-dev] Interesting rendering by nbviewer In-Reply-To: References: Message-ID: On Sun, Mar 16, 2014 at 10:47 PM, Fernando Perez wrote: > This obviously looks wrong, could you file an nbviewer bug, please? Done: https://github.com/ipython/nbviewer/issues/219 St?fan From fperez.net at gmail.com Sun Mar 16 16:56:34 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Sun, 16 Mar 2014 13:56:34 -0700 Subject: [IPython-dev] Interesting rendering by nbviewer In-Reply-To: References: Message-ID: thx! On Sun, Mar 16, 2014 at 1:52 PM, St?fan van der Walt wrote: > On Sun, Mar 16, 2014 at 10:47 PM, Fernando Perez > wrote: > > This obviously looks wrong, could you file an nbviewer bug, please? > > Done: https://github.com/ipython/nbviewer/issues/219 > > St?fan > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From kikocorreoso at gmail.com Sun Mar 16 18:18:22 2014 From: kikocorreoso at gmail.com (Kiko) Date: Sun, 16 Mar 2014 23:18:22 +0100 Subject: [IPython-dev] brythonmagic In-Reply-To: References: <20140303171350.GY21162@HbI-OTOH.berkeley.edu> Message-ID: 2014-03-03 23:48 GMT+01:00 Thomas Kluyver : > Hi Kiko, > > On 3 March 2014 14:05, Kiko wrote: > >> Right now my brython.js is a slight modification of the official brython >> release. Right now it is just a proof of concept. If it is interesting (I >> think it is) for the IPython community I can PR my modifications to the >> official brython repo and then you can use brython.js from a cdn or from >> the official brython site (brython.info) so it wouldn't be necessary to >> load the js libs locally. >> > > I think it's interesting, and if you think the changes are likely to be > accepted back into Brython, go for it. > > Hi all. Now, brythonmagic uses the official brython js files from the official Brython repo (my modifications have been merged). I would like to know the best way to load the required js libs. I know IPython uses requirejs but I am not experienced with it. Is it a good way to load the required js libs as I do? See code cell [3] here: http://nbviewer.ipython.org/github/kikocorreoso/brythonmagic/blob/master/Brython%20usage%20in%20the%20IPython%20notebook.ipynb?create=1#brythonmagic-installation, or is there a better and more elegant way? BTW, thanks Paul and Thomas. > Best wishes, > Thomas > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From maidos93 at laposte.net Sun Mar 16 18:51:45 2014 From: maidos93 at laposte.net (thwiouz) Date: Sun, 16 Mar 2014 15:51:45 -0700 (PDT) Subject: [IPython-dev] Question on remote kernels In-Reply-To: <1394791533704-5050404.post@n6.nabble.com> References: <1394789961867-5050401.post@n6.nabble.com> <1394791533704-5050404.post@n6.nabble.com> Message-ID: <1395010305634-5050621.post@n6.nabble.com> Anyone please? -- View this message in context: http://python.6.x6.nabble.com/Question-on-remote-kernels-tp5050401p5050621.html Sent from the IPython - Development mailing list archive at Nabble.com. From takowl at gmail.com Sun Mar 16 19:31:55 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Sun, 16 Mar 2014 16:31:55 -0700 Subject: [IPython-dev] Question on remote kernels In-Reply-To: <1395010305634-5050621.post@n6.nabble.com> References: <1394789961867-5050401.post@n6.nabble.com> <1394791533704-5050404.post@n6.nabble.com> <1395010305634-5050621.post@n6.nabble.com> Message-ID: On 16 March 2014 15:51, thwiouz wrote: > Anyone please? I don't think we currently expose any good way to start notebook kernels on a remote machine. You'll probably have to customise some of the kernel launcher code to make that possible at present. Min has talked about unifying the interactive kernel and the parallel machinery more, though, which could make it possible to use the various launcher classes for e.g. launching over SSH. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Sun Mar 16 21:30:26 2014 From: doug.blank at gmail.com (Doug Blank) Date: Sun, 16 Mar 2014 21:30:26 -0400 Subject: [IPython-dev] Output from background threads in notebook? Message-ID: If you run the following cell: import threading import time class ThreadClass(threading.Thread): def run(self): x = 1 while True: print(x) x += 1 time.sleep(1) t = ThreadClass() t.start() t.join() Then you will see the data for as long as you don't interrupt the cell. If you comment out the t.join(), then you won't see any output (or maybe just the first printed). In the next cell, if you do something such as, say: while True: pass and then interrupt the kernel, you'll see some more of the printouts. It appears that when the execution_reply is received, the notebook doesn't take output messages anymore. Is this all by design? Is there a manner to make sure background-running output can still make its way to the notebook? -Doug From ntezak at stanford.edu Sun Mar 16 23:00:07 2014 From: ntezak at stanford.edu (Nikolas Tezak) Date: Sun, 16 Mar 2014 20:00:07 -0700 Subject: [IPython-dev] Widget View post render hook In-Reply-To: References: <76E452D7-1098-4E0D-8DC6-B5E831E6DAD2@stanford.edu> Message-ID: <2109409B-430E-4622-BF32-F1E2D1AB1423@stanford.edu> Hi Jon, that is pretty much what I ended up using (took me a while cause I?m a total noob at JavaScript). My circuit widget is coming together nicely. I?ve had a couple other ideas that I?ll probably throw together as soon as I get around to them. One example (shamelessly copied from Mathematica?s ?Manipulate[]? command) 1) extending the interactive command by adding a button that prints out a function call with the currently set parameters. 2) like 1) but perhaps an option to save the parameter setting to disk? (and then reload them when re-opening the notebook at a later time). I don?t like this second option very much because it would be nice to make any notebook self-contained. Potentially one could also insert a new input cell with such a function call / stored parameter settings such that they would be available from within python even after restarting the kernel. But I don?t know if that kind of dynamic code inserting clashes with your guys? vision of what should be doable. Anyhow, thanks for your response! And again, my compliments for an awesome new feature in an amazing tool. Nik On Mar 15, 2014, at 1:03 PM, Jonathan Frederic wrote: > Hi Nik, > > I'm glad you like the widgets and thank you for the feedback. I encountered the same problem when implementing a D3.js widget (see https://github.com/jdfreder/ipython-d3). As a hack-ish workaround I used a timeout with a 0ms interval (https://github.com/jdfreder/ipython-d3/blob/master/widget_forcedirectedgraph.js#L13). Right now I don't believe we have a nice way to do this, adding an event or method is probably a good idea. > > -Jon > > > On Fri, Mar 14, 2014 at 11:27 AM, Nikolas Tezak wrote: > Hi everybody, > > first of all, I _love_ the widgets, they are extremely useful!! And I?m super impressed by IPython?s crazy fast development progress. > I am currently trying to create a custom widget type that embeds a jsplumb circuit editor > ( http://jsplumbtoolkit.com/demo/home/jquery.html ) > > However, when I implement my custom ?render? method for the view I am running into a problem. > I first dynamically create the circuit components (currently just absolutely positioned divs within the widget view?s this.$el div. This works fine. > I then initialize a jsplumb instance and pass this.$el as the default container. > Now, however, when I try to add the connectors and connections with jsplumb within the render method, it fails because jsplumb is trying to read out dynamically the position of the widget view?s div before that div has been inserted into the DOM. > > So basically, my problem is that I want code within ?render()? that already requires the view to be inserted into the document. > Should I use a different approach, i.e., maybe put this initialization code into "update()?, or do you expect that this will be a common enough use case that it would justify adding a special post render hook that gets called after the view?s div has been attached to a cell's widget sub_area? > > > I hope I described this well enough, let me know if you?d like me to share a code example. > Thanks, > > Nik > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From maidos93 at laposte.net Mon Mar 17 04:43:21 2014 From: maidos93 at laposte.net (thwiouz) Date: Mon, 17 Mar 2014 01:43:21 -0700 (PDT) Subject: [IPython-dev] Question on remote kernels In-Reply-To: References: <1394789961867-5050401.post@n6.nabble.com> <1394791533704-5050404.post@n6.nabble.com> <1395010305634-5050621.post@n6.nabble.com> Message-ID: <1395045801280-5050663.post@n6.nabble.com> Thanks, I'll try to do that. Cheers, -- View this message in context: http://python.6.x6.nabble.com/Question-on-remote-kernels-tp5050401p5050663.html Sent from the IPython - Development mailing list archive at Nabble.com. From j.davidgriffiths at gmail.com Mon Mar 17 10:01:46 2014 From: j.davidgriffiths at gmail.com (John Griffiths) Date: Mon, 17 Mar 2014 14:01:46 +0000 Subject: [IPython-dev] toggle cell input - changed in dev version? Message-ID: I've been making use of a little piece of javascript in a markdown cell that hides all code cells in the notebook; toggle input (e.g. see here http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html) However, I recently installed the latest ipython dev version, and now this snippet doesn't seem to be working any more. Any obvious reasons for this? Ta, john -- Mr. John Griffiths, MSc PhD Candidate Centre for Speech, Language, and the Brain Department of Experimental Psychology University of Cambridge, UK -------------- next part -------------- An HTML attachment was scrubbed... URL: From clyde.fare at gmail.com Mon Mar 17 10:25:10 2014 From: clyde.fare at gmail.com (Clyde Fare) Date: Mon, 17 Mar 2014 14:25:10 +0000 Subject: [IPython-dev] Loop over a set of notebook cells? In-Reply-To: References: Message-ID: I think MinRK did something like this here: http://nbviewer.ipython.org/gist/minrk/6011986 His NotebookLoader class does loop over the cells. Clyde On 15 March 2014 11:18, Patrick Surry wrote: > What about a "loop delimiter" magic where you insert a matching pair of > %magics bracketing a sequence of regular cells, and then it basically > unrolls the loop for you? > > i.e. when you run the magic, on the first loop iteration it executes each > cell in place, with the first loop value, and for each subsequent > iterations, it appends a copied sequence of new cells after the %end, > evaluated with the corresponding loop element. So if you looped over a > list of two items, you'd end up with a second copy of all your looped cells > that have been executed with the second item in the list. (see example > sketch below) > > So it's doing a kind of repeated block copy & paste (useful in it's own > right), unrolling the loop, and leaving you the new results to explore and > further tweak. Maybe it would also mark itself as executed somehow > (comment itself out?!) so the notebook is stable if you re-execute all > cells? > > You'd rapidly want multi-cell delete I guess :) Perhaps other bracketing > magics like %block cut / %endblock / %block paste ?? > > Patrick > > Example with cells a, b, c: > > %begin for item in [1,2] > --- > a > --- > b > --- > c > --- > %end > > when executed becomes: > > # %begin for item in [1, 2] > --- > a [item=1] > --- > b [item=1] > --- > c [item=1] > --- > # %end > --- > a [item=2] > --- > b [item=2] > --- > c [item=2] > --- > > > From: Thomas Kluyver > > >On 14 March 2014 09:46, Nikolas Tezak wrote: > > > I've been thinking about this for a while as well. Although this doesn't > > quite address looping, I think one idea might be to use cell tags to > create > > certain "run sets" of cells and then have a little widget/ui component > that > > allows for selecting and running all cells of a certain tag. So, in that > > respect a good interface for marking/selecting multiple cells and then > > tagging them or running them or copying them would be great. > > > Tags on cells is something that we want to do - there are a lot of useful > things that could be done with it. > > Thomas > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From damianavila at gmail.com Mon Mar 17 10:29:08 2014 From: damianavila at gmail.com (=?ISO-8859-1?Q?Dami=E1n_Avila?=) Date: Mon, 17 Mar 2014 11:29:08 -0300 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: Message-ID: Working for me in last master... any message in the js console... or in the console starting the notebook? 2014-03-17 11:01 GMT-03:00 John Griffiths : > > I've been making use of a little piece of javascript in a markdown cell > that hides all code cells in the notebook; > > > toggle input > > (e.g. see here > http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html) > > > However, I recently installed the latest ipython dev version, and now this > snippet doesn't seem to be working any more. > > Any obvious reasons for this? > > Ta, > > john > > > > > -- > > Mr. John Griffiths, MSc > > PhD Candidate > > Centre for Speech, Language, and the Brain > > Department of Experimental Psychology > > University of Cambridge, UK > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -- Dami?n Avila Scientific Python Developer Quantitative Finance Analyst Statistics, Biostatistics and Econometrics Consultant Biochemist -------------- next part -------------- An HTML attachment was scrubbed... URL: From j.davidgriffiths at gmail.com Mon Mar 17 10:56:39 2014 From: j.davidgriffiths at gmail.com (John Griffiths) Date: Mon, 17 Mar 2014 14:56:39 +0000 Subject: [IPython-dev] \citep and \citet for markdown latex citations Message-ID: Following up on an discussion re: latex citations here http://stackoverflow.com/questions/22425847/can-other-latex-citation-commands-like-citep-and-citet-be-used-in-notebook-markd?noredirect=1#comment34138366_22425847 I was asking how one might implement \citep and \citet versions of the (Granger, 2013) syntax for markdown citations in the notebook. Seems the natural thing to do would be to be able to write " wrote: > Working for me in last master... any message in the js console... or in > the console starting the notebook? > > > 2014-03-17 11:01 GMT-03:00 John Griffiths : > >> >> I've been making use of a little piece of javascript in a markdown cell >> that hides all code cells in the notebook; >> >> >> toggle input >> >> (e.g. see here >> http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html) >> >> >> However, I recently installed the latest ipython dev version, and now >> this snippet doesn't seem to be working any more. >> >> Any obvious reasons for this? >> >> Ta, >> >> john >> >> >> >> >> -- >> >> Mr. John Griffiths, MSc >> >> PhD Candidate >> >> Centre for Speech, Language, and the Brain >> >> Department of Experimental Psychology >> >> University of Cambridge, UK >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > > -- > Dami?n Avila > Scientific Python Developer > Quantitative Finance Analyst > Statistics, Biostatistics and Econometrics Consultant > Biochemist > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -- Mr. John Griffiths, MSc PhD Candidate Centre for Speech, Language, and the Brain Department of Experimental Psychology University of Cambridge, UK -------------- next part -------------- An HTML attachment was scrubbed... URL: From patrick.surry at gmail.com Mon Mar 17 12:05:51 2014 From: patrick.surry at gmail.com (Patrick Surry) Date: Mon, 17 Mar 2014 12:05:51 -0400 Subject: [IPython-dev] Loop over a set of notebook cells? Message-ID: Yes, I'd seen that, but it seemed to fall into the external/batch category (write a script that post-processes a notebook) rather than something that you could do interactively within the notebook workflow. Cheers, Patrick From: Clyde Fare I think MinRK did something like this here: http://nbviewer.ipython.org/gist/minrk/6011986 His NotebookLoader class does loop over the cells. Clyde On 15 March 2014 11:18, Patrick Surry wrote: > What about a "loop delimiter" magic where you insert a matching pair of > %magics bracketing a sequence of regular cells, and then it basically > unrolls the loop for you? > > i.e. when you run the magic, on the first loop iteration it executes each > cell in place, with the first loop value, and for each subsequent > iterations, it appends a copied sequence of new cells after the %end, > evaluated with the corresponding loop element. So if you looped over a > list of two items, you'd end up with a second copy of all your looped cells > that have been executed with the second item in the list. (see example > sketch below) > > So it's doing a kind of repeated block copy & paste (useful in it's own > right), unrolling the loop, and leaving you the new results to explore and > further tweak. Maybe it would also mark itself as executed somehow > (comment itself out?!) so the notebook is stable if you re-execute all > cells? > > You'd rapidly want multi-cell delete I guess :) Perhaps other bracketing > magics like %block cut / %endblock / %block paste ?? > > Patrick > > Example with cells a, b, c: > > %begin for item in [1,2] > --- > a > --- > b > --- > c > --- > %end > > when executed becomes: > > # %begin for item in [1, 2] > --- > a [item=1] > --- > b [item=1] > --- > c [item=1] > --- > # %end > --- > a [item=2] > --- > b [item=2] > --- > c [item=2] > --- > > > From: Thomas Kluyver > > >On 14 March 2014 09:46, Nikolas Tezak wrote: > > > I've been thinking about this for a while as well. Although this doesn't > > quite address looping, I think one idea might be to use cell tags to > create > > certain "run sets" of cells and then have a little widget/ui component > that > > allows for selecting and running all cells of a certain tag. So, in that > > respect a good interface for marking/selecting multiple cells and then > > tagging them or running them or copying them would be great. > > > Tags on cells is something that we want to do - there are a lot of useful > things that could be done with it. > > Thomas > -------------- next part -------------- An HTML attachment was scrubbed... URL: From damianavila at gmail.com Mon Mar 17 12:10:21 2014 From: damianavila at gmail.com (=?ISO-8859-1?Q?Dami=E1n_Avila?=) Date: Mon, 17 Mar 2014 13:10:21 -0300 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: Message-ID: It seems to be the new Sanitizer (google caja), but first let me know the output of the console where you launch the IPython notebook... 2014-03-17 12:17 GMT-03:00 John Griffiths : > > Ok perhaps some useful info here. When I 'run' (shift-enter) the markdown > cell in google chrome, this pops up: > > HTML Sanitizer script removed Object {change: "removed", tagName: > "script"} security.js?v=e9cf1cd9e6335aeaf3d0c366765ad26d:93 > HTML Sanitizer a.href removed Object {change: "removed", tagName: "a", > attribName: "href", oldValue: "javascript:toggle()", newValue: null} > security.js?v=e9cf1cd9e6335aeaf3d0c366765ad26d:93 > HTML Sanitizer a.target removed Object {change: "removed", tagName: "a", > attribName: "target", oldValue: "_self", newValue: null} > security.js?v=e9cf1cd9e6335aeaf3d0c366765ad26d:93 > > This means very little to me (first time I've ever intentionally opened a > browser javascript console). Any ideas? > > > Ta. > > > > > > On 17 March 2014 14:29, Dami?n Avila wrote: > >> Working for me in last master... any message in the js console... or in >> the console starting the notebook? >> >> >> 2014-03-17 11:01 GMT-03:00 John Griffiths : >> >>> >>> I've been making use of a little piece of javascript in a markdown >>> cell that hides all code cells in the notebook; >>> >>> >>> toggle input >>> >>> (e.g. see here >>> http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html) >>> >>> >>> However, I recently installed the latest ipython dev version, and now >>> this snippet doesn't seem to be working any more. >>> >>> Any obvious reasons for this? >>> >>> Ta, >>> >>> john >>> >>> >>> >>> >>> -- >>> >>> Mr. John Griffiths, MSc >>> >>> PhD Candidate >>> >>> Centre for Speech, Language, and the Brain >>> >>> Department of Experimental Psychology >>> >>> University of Cambridge, UK >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >>> >> >> >> -- >> Dami?n Avila >> Scientific Python Developer >> Quantitative Finance Analyst >> Statistics, Biostatistics and Econometrics Consultant >> Biochemist >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > > -- > > Mr. John Griffiths, MSc > > PhD Candidate > > Centre for Speech, Language, and the Brain > > Department of Experimental Psychology > > University of Cambridge, UK > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -- Dami?n Avila Scientific Python Developer Quantitative Finance Analyst Statistics, Biostatistics and Econometrics Consultant Biochemist -------------- next part -------------- An HTML attachment was scrubbed... URL: From ellisonbg at gmail.com Mon Mar 17 13:14:21 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Mon, 17 Mar 2014 10:14:21 -0700 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: Message-ID: As of 2.0, Markdown cells no longer support script tags. You will need to put this code into HTML/JavaScript output using %%html or %%javascript. This is a security issue - we will have more details on this as 2.0 is released. Cheers, Brian On Mon, Mar 17, 2014 at 7:01 AM, John Griffiths wrote: > > I've been making use of a little piece of javascript in a markdown cell that > hides all code cells in the notebook; > > > toggle input > > (e.g. see here > http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html > ) > > > However, I recently installed the latest ipython dev version, and now this > snippet doesn't seem to be working any more. > > Any obvious reasons for this? > > Ta, > > john > > > > > -- > > Mr. John Griffiths, MSc > > PhD Candidate > > Centre for Speech, Language, and the Brain > > Department of Experimental Psychology > > University of Cambridge, UK > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From damianavila at gmail.com Mon Mar 17 13:28:54 2014 From: damianavila at gmail.com (=?ISO-8859-1?Q?Dami=E1n_Avila?=) Date: Mon, 17 Mar 2014 14:28:54 -0300 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: Message-ID: Uppsss... I forgot to mention I tested your code with the %%HTML magic... sorry to did not mention it before. Thanks Bryan. 2014-03-17 14:14 GMT-03:00 Brian Granger : > As of 2.0, Markdown cells no longer support script tags. You will need > to put this code into HTML/JavaScript output using %%html or > %%javascript. This is a security issue - we will have more details on > this as 2.0 is released. > > Cheers, > > Brian > > On Mon, Mar 17, 2014 at 7:01 AM, John Griffiths > wrote: > > > > I've been making use of a little piece of javascript in a markdown cell > that > > hides all code cells in the notebook; > > > > > > toggle input > > > > (e.g. see here > > > http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html > > ) > > > > > > However, I recently installed the latest ipython dev version, and now > this > > snippet doesn't seem to be working any more. > > > > Any obvious reasons for this? > > > > Ta, > > > > john > > > > > > > > > > -- > > > > Mr. John Griffiths, MSc > > > > PhD Candidate > > > > Centre for Speech, Language, and the Brain > > > > Department of Experimental Psychology > > > > University of Cambridge, UK > > > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > > > -- > Brian E. Granger > Cal Poly State University, San Luis Obispo > bgranger at calpoly.edu and ellisonbg at gmail.com > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Dami?n Avila Scientific Python Developer Quantitative Finance Analyst Statistics, Biostatistics and Econometrics Consultant Biochemist -------------- next part -------------- An HTML attachment was scrubbed... URL: From andrew.gibiansky at gmail.com Mon Mar 17 13:29:52 2014 From: andrew.gibiansky at gmail.com (Andrew Gibiansky) Date: Mon, 17 Mar 2014 10:29:52 -0700 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: Message-ID: My 2-cents: While I understand the security issues, it's a major inconvenience in my personal work... In addition, it's a pretty heavy burden on any non-Python kernel implementors; IHaskell currently has absolutely no notion of magics (though directives are similar, maybe) but certainly no notion of cell magics (and none planned for the time being)... (I have a ton of things to do before implementing that, and haven't seen much use for cell magics, especially given that Haskell has QuasiQuotes which allow similar things). It's be really nice if there were some way to disable this security feature. Do you think this might be possible somehow? -- Andrew On March 17, 2014 at 10:14:27 AM, Brian Granger (ellisonbg at gmail.com) wrote: As of 2.0, Markdown cells no longer support script tags. You will need to put this code into HTML/JavaScript output using %%html or %%javascript. This is a security issue - we will have more details on this as 2.0 is released. Cheers, Brian On Mon, Mar 17, 2014 at 7:01 AM, John Griffiths wrote: > > I've been making use of a little piece of javascript in a markdown cell that > hides all code cells in the notebook; > > > toggle input > > (e.g. see here > http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html > ) > > > However, I recently installed the latest ipython dev version, and now this > snippet doesn't seem to be working any more. > > Any obvious reasons for this? > > Ta, > > john > > > > > -- > > Mr. John Griffiths, MSc > > PhD Candidate > > Centre for Speech, Language, and the Brain > > Department of Experimental Psychology > > University of Cambridge, UK > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: From fperez.net at gmail.com Mon Mar 17 13:38:02 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Mon, 17 Mar 2014 10:38:02 -0700 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: Message-ID: Andrew, is it not easy/possible for you to simply write Haskell functions that return as a value, the necessary JS/HTML? That's all that %%html and %%js do. All we're forbidding is embedded > > toggle input > > > > (e.g. see here > > > http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html > > ) > > > > > > However, I recently installed the latest ipython dev version, and now > this > > snippet doesn't seem to be working any more. > > > > Any obvious reasons for this? > > > > Ta, > > > > john > > > > > > > > > > -- > > > > Mr. John Griffiths, MSc > > > > PhD Candidate > > > > Centre for Speech, Language, and the Brain > > > > Department of Experimental Psychology > > > > University of Cambridge, UK > > > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > > > -- > Brian E. Granger > Cal Poly State University, San Luis Obispo > bgranger at calpoly.edu and ellisonbg at gmail.com > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From ellisonbg at gmail.com Mon Mar 17 13:50:14 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Mon, 17 Mar 2014 10:50:14 -0700 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: Message-ID: Sorry I missed that this was not for the python kernel. As Fernando mentioned, just using the HTML/JS output and display protocol would work fine. I agree with Fernando that this part of our security approach is not something we will make configurable. The reason is that >> > toggle input >> > >> > (e.g. see here >> > >> > http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html >> > ) >> > >> > >> > However, I recently installed the latest ipython dev version, and now >> > this >> > snippet doesn't seem to be working any more. >> > >> > Any obvious reasons for this? >> > >> > Ta, >> > >> > john >> > >> > >> > >> > >> > -- >> > >> > Mr. John Griffiths, MSc >> > >> > PhD Candidate >> > >> > Centre for Speech, Language, and the Brain >> > >> > Department of Experimental Psychology >> > >> > University of Cambridge, UK >> > >> > >> > _______________________________________________ >> > IPython-dev mailing list >> > IPython-dev at scipy.org >> > http://mail.scipy.org/mailman/listinfo/ipython-dev >> > >> >> >> >> -- >> Brian E. Granger >> Cal Poly State University, San Luis Obispo >> bgranger at calpoly.edu and ellisonbg at gmail.com >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > > > > -- > Fernando Perez (@fperez_org; http://fperez.org) > fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) > fernando.perez-at-berkeley: contact me here for any direct mail -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From ellisonbg at gmail.com Mon Mar 17 13:53:45 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Mon, 17 Mar 2014 10:53:45 -0700 Subject: [IPython-dev] \citep and \citet for markdown latex citations In-Reply-To: References: Message-ID: I am +1 on adding support for other common citation patterns. This should be a very simple extension of the code we already have. You can either open a PR if you want to do this yourself, or create an issue and someone else will pick it up. Cheers, Brian On Mon, Mar 17, 2014 at 7:56 AM, John Griffiths wrote: > > Following up on an discussion re: latex citations here > > http://stackoverflow.com/questions/22425847/can-other-latex-citation-commands-like-citep-and-citet-be-used-in-notebook-markd?noredirect=1#comment34138366_22425847 > > I was asking how one might implement \citep and \citet versions of the > > (Granger, 2013) > > syntax for markdown citations in the notebook. > > Seems the natural thing to do would be to be able to write " data-citep=" > and " wrote: > Andrew, is it not easy/possible for you to simply write Haskell functions > that return as a value, the necessary JS/HTML? That's all that %%html and > %%js do. > > All we're forbidding is embedded >> > toggle input >> > >> > (e.g. see here >> > >> > http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html >> > ) >> > >> > >> > However, I recently installed the latest ipython dev version, and now >> > this >> > snippet doesn't seem to be working any more. >> > >> > Any obvious reasons for this? >> > >> > Ta, >> > >> > john >> > >> > >> > >> > >> > -- >> > >> > Mr. John Griffiths, MSc >> > >> > PhD Candidate >> > >> > Centre for Speech, Language, and the Brain >> > >> > Department of Experimental Psychology >> > >> > University of Cambridge, UK >> > >> > >> > _______________________________________________ >> > IPython-dev mailing list >> > IPython-dev at scipy.org >> > http://mail.scipy.org/mailman/listinfo/ipython-dev >> > >> >> >> >> -- >> Brian E. Granger >> Cal Poly State University, San Luis Obispo >> bgranger at calpoly.edu and ellisonbg at gmail.com >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > > > > -- > Fernando Perez (@fperez_org; http://fperez.org) > fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) > fernando.perez-at-berkeley: contact me here for any direct mail -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: From python at elbonia.de Mon Mar 17 14:33:25 2014 From: python at elbonia.de (Juergen Hasch) Date: Mon, 17 Mar 2014 19:33:25 +0100 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: Message-ID: <53273FF5.1060102@elbonia.de> You can easily make a notebook extension out of this. Take a look at: https://github.com/ipython-contrib/IPython-notebook-extensions/blob/master/usability/toggle_codecells.js Am 17.03.2014 15:01, schrieb John Griffiths: > > I've been making use of a little piece of javascript in a markdown cell that hides all code cells in the notebook; > > > toggle input > > (e.g. see here http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html ) > > > However, I recently installed the latest ipython dev version, and now this snippet doesn't seem to be working any more. > > Any obvious reasons for this? > > Ta, > > john > > > > > -- > > Mr. John Griffiths, MSc > > PhD Candidate > > Centre for Speech, Language, and the Brain____ > > Department of Experimental Psychology > > University of Cambridge, UK > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From benjaminrk at gmail.com Mon Mar 17 14:44:07 2014 From: benjaminrk at gmail.com (MinRK) Date: Mon, 17 Mar 2014 11:44:07 -0700 Subject: [IPython-dev] Loop over a set of notebook cells? In-Reply-To: References: Message-ID: You can execute a cell range with a little javascript (e.g. in a %%javascript cell): var start = 2; var stop = 4; for (var i = start; i < stop; i++) { var cell = IPython.notebook.get_cell(i); cell.execute(); } On Mon, Mar 17, 2014 at 9:05 AM, Patrick Surry wrote: > Yes, I'd seen that, but it seemed to fall into the external/batch category > (write a script that post-processes a notebook) rather than something > that you could do interactively within the notebook workflow. > > Cheers, > Patrick > > > From: Clyde Fare > > > I think MinRK did something like this here: > > http://nbviewer.ipython.org/gist/minrk/6011986 > > His NotebookLoader class does loop over the cells. > > Clyde > > > > On 15 March 2014 11:18, Patrick Surry wrote: > > > What about a "loop delimiter" magic where you insert a matching pair of > > %magics bracketing a sequence of regular cells, and then it basically > > unrolls the loop for you? > > > > i.e. when you run the magic, on the first loop iteration it executes each > > cell in place, with the first loop value, and for each subsequent > > iterations, it appends a copied sequence of new cells after the %end, > > evaluated with the corresponding loop element. So if you looped over a > > list of two items, you'd end up with a second copy of all your looped > cells > > that have been executed with the second item in the list. (see example > > sketch below) > > > > So it's doing a kind of repeated block copy & paste (useful in it's own > > right), unrolling the loop, and leaving you the new results to explore > and > > further tweak. Maybe it would also mark itself as executed somehow > > (comment itself out?!) so the notebook is stable if you re-execute all > > cells? > > > > You'd rapidly want multi-cell delete I guess :) Perhaps other bracketing > > magics like %block cut / %endblock / %block paste ?? > > > > Patrick > > > > Example with cells a, b, c: > > > > %begin for item in [1,2] > > --- > > a > > --- > > b > > --- > > c > > --- > > %end > > > > when executed becomes: > > > > # %begin for item in [1, 2] > > --- > > a [item=1] > > --- > > b [item=1] > > --- > > c [item=1] > > --- > > # %end > > --- > > a [item=2] > > --- > > b [item=2] > > --- > > c [item=2] > > --- > > > > > > From: Thomas Kluyver > > > > >On 14 March 2014 09:46, Nikolas Tezak wrote: > > > > > I've been thinking about this for a while as well. Although this > doesn't > > > quite address looping, I think one idea might be to use cell tags to > > create > > > certain "run sets" of cells and then have a little widget/ui component > > that > > > allows for selecting and running all cells of a certain tag. So, in > that > > > respect a good interface for marking/selecting multiple cells and then > > > tagging them or running them or copying them would be great. > > > > > > Tags on cells is something that we want to do - there are a lot of useful > > things that could be done with it. > > > > Thomas > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From j.davidgriffiths at gmail.com Mon Mar 17 15:04:49 2014 From: j.davidgriffiths at gmail.com (John Griffiths) Date: Mon, 17 Mar 2014 19:04:49 +0000 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: <53273FF5.1060102@elbonia.de> References: <53273FF5.1060102@elbonia.de> Message-ID: Thanks everyone for the informative discussion. Running the script with %%HTML is working fine. Juergen - thanks for adding the extension. Perhaps better to call it something like 'hide_input_all.js' though - so it's clearer that it's doing the same thing as 'hide_input.js'. Or, even better, have a single extension that adds buttons for both. Apologies I'm not ofay enough with javascript to do that myself. In any case its useful to have a functional snippet to hand in addition to the toolbar extensions, as tend to move between systems quite a lot. Ta, john On 17 March 2014 18:33, Juergen Hasch wrote: > > You can easily make a notebook extension out of this. Take a look at: > > https://github.com/ipython-contrib/IPython-notebook-extensions/blob/master/usability/toggle_codecells.js > > > Am 17.03.2014 15:01, schrieb John Griffiths: > > > > I've been making use of a little piece of javascript in a markdown cell > that hides all code cells in the notebook; > > > > > > toggle input > > > > (e.g. see here > http://python.6.x6.nabble.com/IPython-User-Hide-code-cells-in-the-notebook-td4997151.html) > > > > > > However, I recently installed the latest ipython dev version, and now > this snippet doesn't seem to be working any more. > > > > Any obvious reasons for this? > > > > Ta, > > > > john > > > > > > > > > > -- > > > > Mr. John Griffiths, MSc > > > > PhD Candidate > > > > Centre for Speech, Language, and the Brain____ > > > > Department of Experimental Psychology > > > > University of Cambridge, UK > > > > > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Mr. John Griffiths, MSc PhD Candidate Centre for Speech, Language, and the Brain Department of Experimental Psychology University of Cambridge, UK -------------- next part -------------- An HTML attachment was scrubbed... URL: From python at elbonia.de Mon Mar 17 15:27:39 2014 From: python at elbonia.de (Juergen Hasch) Date: Mon, 17 Mar 2014 20:27:39 +0100 Subject: [IPython-dev] toggle cell input - changed in dev version? In-Reply-To: References: <53273FF5.1060102@elbonia.de> Message-ID: <53274CAB.9080006@elbonia.de> Am 17.03.2014 20:04, schrieb John Griffiths: > > > Thanks everyone for the informative discussion. > > Running the script with %%HTML is working fine. > > Juergen - thanks for adding the extension. Perhaps better to call it something like 'hide_input_all.js' though - so it's > clearer that it's doing the same thing as 'hide_input.js'. Or, even better, have a single extension that adds buttons > for both. Apologies I'm not ofay enough with javascript to do that myself. In any case its useful to have a functional > snippet to hand in addition to the toolbar extensions, as tend to move between systems quite a lot. > This sounds like a better name for the extension. I will rename it. If you use more than one computer you can add an extra static path on a network drive and place the extensions there. This is at least what I do. Juergen From takowl at gmail.com Mon Mar 17 15:51:19 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Mon, 17 Mar 2014 12:51:19 -0700 Subject: [IPython-dev] \citep and \citet for markdown latex citations In-Reply-To: References: Message-ID: Hi John, On 17 March 2014 07:56, John Griffiths wrote: > @takluyver suggests making a modified the citation filter to do this. > Looking at the nbconvert code this looks like it should be quite simple. > But what I'm not sure about is what the the data-cite attribute is doing > here. Would someone mind explaining this a little? > > In particular, would e.g. a data-citet attribute need to be explicitly > defined somewhere, and would that be a non-trivial exercise? > HTML 5 allows any data-something attributes: there's no need to define it explicitly, and it doesn't have a fixed meaning. data-cite is just the name we chose - you can parse other names just the same way. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From wstein at gmail.com Mon Mar 17 20:37:48 2014 From: wstein at gmail.com (William Stein) Date: Mon, 17 Mar 2014 17:37:48 -0700 Subject: [IPython-dev] [sage-support] Re: Connection beetwen sage and notebook/sagemathcloud In-Reply-To: References: Message-ID: On Mon, Mar 17, 2014 at 5:00 PM, Andrey Novoseltsev wrote: > Is it at all possible to hook into (I)Python to catch traceback and replace > it with a translation? Such an approach would work with all frontends > including SMC. There can be a problem with error messages that are > "constructed from pieces", but, of course, even partial translation is > better than none. > > If things really have to be handled by frontends, then sagenb may not be > worth putting in a lot of effort since it is likely to be replaced and > William is the right person to ask about SMC! And the IPython dev list is a good place to ask about this. I'm sure they've thought about internationalization strategies for IPython (the command line, web frontend, etc.). I don't know which way will be best for SMC. There are many options -- with anything, I usually spend a while (from minutes to *years*!) investigating the possibilities... William > > Best regards, > Andrey > > > On Sunday, 16 March 2014 00:10:15 UTC-6, ?????? ?????? wrote: >> >> Hello! >> >> I decided to make a study of internationalization of error messages >> (python traceback). >> I came to decision there should be next steps: >> 1. to find error messages at code of all functions/modules and systematize >> these messages. >> 2. to translate all messages and create table (maybe, .csv file). >> 3. to learn how notebook or sagemathcloud connect to sage. >> 4. to write function that catch error messages from sage to notebook or >> sagemathcloud and translate these messages. >> >> I tried to understand how notebook connects to sage (I read source code or >> notebook_object.py, run_notebook.py and online documentation and some code >> on twisted site) but not I can't to understand how it's realized. Please >> advise me where I can read about setting and getting information beetwin >> sage and notebook. Because there isn't any way to install sagemathcloud on >> local computer I don't know how I can do such internationalization for >> sagemathcloud. Is there any way to do know how sagemathcloud connect to >> sage? Help me please. >> >> Best regards, >> Andrei Shirshov. > > -- > You received this message because you are subscribed to the Google Groups > "sage-support" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-support+unsubscribe at googlegroups.com. > To post to this group, send email to sage-support at googlegroups.com. > Visit this group at http://groups.google.com/group/sage-support. > For more options, visit https://groups.google.com/d/optout. -- William Stein Professor of Mathematics University of Washington http://wstein.org From fperez.net at gmail.com Mon Mar 17 21:52:11 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Mon, 17 Mar 2014 18:52:11 -0700 Subject: [IPython-dev] [sage-support] Re: Connection beetwen sage and notebook/sagemathcloud In-Reply-To: References: Message-ID: On Mon, Mar 17, 2014 at 5:37 PM, William Stein wrote: > I'm sure they've thought about internationalization strategies for > IPython (the command line, web frontend, etc.). > Ah, you're giving us way too much credit :) Questions on i18n have popped up from time to time, but we've never had any bandwidth to really act on them, and nobody has stepped up to do any actual work on it. So if Andrey wants to look at the problem, he's welcome to, but he's going into 100% uncharted waters, I'm afraid... Cheers, f -------------- next part -------------- An HTML attachment was scrubbed... URL: From wstein at gmail.com Mon Mar 17 22:12:44 2014 From: wstein at gmail.com (William Stein) Date: Mon, 17 Mar 2014 19:12:44 -0700 Subject: [IPython-dev] meeting in Berkeley In-Reply-To: <20140304001827.GB21162@HbI-OTOH.berkeley.edu> References: <20140304001827.GB21162@HbI-OTOH.berkeley.edu> Message-ID: On Mon, Mar 3, 2014 at 4:18 PM, Paul Ivanov wrote: >>> On Sun, Mar 2, 2014 at 6:56 PM, William Stein wrote: >>>> Would any IPython notebook devs be interested in meeting in >>>> **Berkeley** on Friday, March 21? Is there any time that day that >>>> would work for some of you? >>On 2 March 2014 20:15, MinRK wrote: >>> Absolutely! Just about any time should work for me. >>> >>> -MinRK >>> >>> > Thomas Kluyver, on 2014-03-03 10:05, wrote: >> Yep, I'll be here all day as well. > > Ditto for me. If you're feeling adventurous, you can attend the > Python Workers' Party 4-6pm. > > http://python.berkeley.edu/events/2014/01/24/python-workers-party-rally.html Cool -- I have my plane tickets and will be in Berkeley from Wednesday through Sunday. See you guys later this week to randomly chat about stuff! -- William > > cheers, > -- > _ > / \ > A* \^ - > ,./ _.`\\ / \ > / ,--.S \/ \ > / `"~,_ \ \ > __o ? > _ \<,_ /:\ > --(_)/-(_)----.../ | \ > --------------.......J > Paul Ivanov > http://pirsquared.org > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- William Stein Professor of Mathematics University of Washington http://wstein.org From aaron.oleary at gmail.com Tue Mar 18 04:06:00 2014 From: aaron.oleary at gmail.com (Aaron O'Leary) Date: Tue, 18 Mar 2014 08:06:00 +0000 Subject: [IPython-dev] \citep and \citet for markdown latex citations In-Reply-To: References: Message-ID: > Seems the natural thing to do would be to be able to write " data-citep=" > and " References: Message-ID: 2014-03-17 19:44 GMT+01:00 MinRK : > You can execute a cell range with a little javascript (e.g. in a > %%javascript cell): > > var start = 2; > var stop = 4; > > for (var i = start; i < stop; i++) { > var cell = IPython.notebook.get_cell(i); > cell.execute(); > } > > Maybe this: %%javascript IPython.notebook.execute_cell_range(x,y) with x,y the range of the cells. > > On Mon, Mar 17, 2014 at 9:05 AM, Patrick Surry wrote: > >> Yes, I'd seen that, but it seemed to fall into the external/batch >> category (write a script that post-processes a notebook) rather than >> something that you could do interactively within the notebook workflow. >> >> Cheers, >> Patrick >> >> >> From: Clyde Fare >> >> >> I think MinRK did something like this here: >> >> http://nbviewer.ipython.org/gist/minrk/6011986 >> >> His NotebookLoader class does loop over the cells. >> >> Clyde >> >> >> >> On 15 March 2014 11:18, Patrick Surry wrote: >> >> > What about a "loop delimiter" magic where you insert a matching pair of >> > %magics bracketing a sequence of regular cells, and then it basically >> > unrolls the loop for you? >> > >> > i.e. when you run the magic, on the first loop iteration it executes >> each >> > cell in place, with the first loop value, and for each subsequent >> > iterations, it appends a copied sequence of new cells after the %end, >> > evaluated with the corresponding loop element. So if you looped over a >> > list of two items, you'd end up with a second copy of all your looped >> cells >> > that have been executed with the second item in the list. (see example >> > sketch below) >> > >> > So it's doing a kind of repeated block copy & paste (useful in it's own >> > right), unrolling the loop, and leaving you the new results to explore >> and >> > further tweak. Maybe it would also mark itself as executed somehow >> > (comment itself out?!) so the notebook is stable if you re-execute all >> > cells? >> > >> > You'd rapidly want multi-cell delete I guess :) Perhaps other >> bracketing >> > magics like %block cut / %endblock / %block paste ?? >> > >> > Patrick >> > >> > Example with cells a, b, c: >> > >> > %begin for item in [1,2] >> > --- >> > a >> > --- >> > b >> > --- >> > c >> > --- >> > %end >> > >> > when executed becomes: >> > >> > # %begin for item in [1, 2] >> > --- >> > a [item=1] >> > --- >> > b [item=1] >> > --- >> > c [item=1] >> > --- >> > # %end >> > --- >> > a [item=2] >> > --- >> > b [item=2] >> > --- >> > c [item=2] >> > --- >> > >> > >> > From: Thomas Kluyver >> > >> > >On 14 March 2014 09:46, Nikolas Tezak wrote: >> > >> > > I've been thinking about this for a while as well. Although this >> doesn't >> > > quite address looping, I think one idea might be to use cell tags to >> > create >> > > certain "run sets" of cells and then have a little widget/ui component >> > that >> > > allows for selecting and running all cells of a certain tag. So, in >> that >> > > respect a good interface for marking/selecting multiple cells and then >> > > tagging them or running them or copying them would be great. >> > >> > >> > Tags on cells is something that we want to do - there are a lot of >> useful >> > things that could be done with it. >> > >> > Thomas >> > >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From j.davidgriffiths at gmail.com Tue Mar 18 06:42:57 2014 From: j.davidgriffiths at gmail.com (John Griffiths) Date: Tue, 18 Mar 2014 10:42:57 +0000 Subject: [IPython-dev] \citep and \citet for markdown latex citations In-Reply-To: References: Message-ID: So I opened an issue, and in the end did a little hack myself and submitted PR. See links here for the PR and for an example notebook https://github.com/ipython/ipython/issues/5370#issuecomment-37897451 I agree that the html tag is a bit long-winded. Shorter solution would be better if poss. But in terms of rendering they look fine; (Granger, 2013) just gets rendered as 'Granger, 2013'. I've raised a few latex issues over the last year or so on stackoverflow that various ipython peeps have helped with. You might find some of those useful. http://stackoverflow.com/users/1224169/j-grif In the end I've actually opted to use my institution's latex thesis template, with various post-nbconvert few hacks on the .tex files. If I had more time I would have preferred to make an nbconvert template that does the same job; but I don't have that luxury and also am not a latex expert. Repo is here https://github.com/JohnGriffiths/phd-thesis-template ...though I wasn't planning on publicising this yet as it isn't particularly clean and lots of uncommitted changes. You might want to take a look at this notebook though http://nbviewer.ipython.org/github/JohnGriffiths/phd-thesis-template/blob/master/JG_hacks/DoMe.ipynb ...which basically does everything and has a reasonable amount of documentation. Let me know your thoughts. I could easily see an 'nbconvert thesis template' being very popular. Happy to help move a mini-project like that forward. On 18 March 2014 08:06, Aaron O'Leary wrote: > > Seems the natural thing to do would be to be able to write " > data-citep=" > > and " Might be more flexible. > > I feel that writing the explicit html tag is a bit cumbersome and that > we could do with a shorthand. We could use latex \citet, \citep > directly, or we could go with the pandoc @granger2013, [@granger2013]. > The problem with shorthand is that it won't be rendered nicely in the > interactive notebook. > > I've been thinking about this recently as I'm developing my thesis in > the notebook. I've also got some ideas about referencing figures. I'd > be happy to work on a PR with you John. > > aaron > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Mr. John Griffiths, MSc PhD Candidate Centre for Speech, Language, and the Brain Department of Experimental Psychology University of Cambridge, UK -------------- next part -------------- An HTML attachment was scrubbed... URL: From aaron.oleary at gmail.com Tue Mar 18 07:50:02 2014 From: aaron.oleary at gmail.com (Aaron O'Leary) Date: Tue, 18 Mar 2014 11:50:02 +0000 Subject: [IPython-dev] \citep and \citet for markdown latex citations In-Reply-To: References: Message-ID: <20140318115002.GA10229@tk422.wireless.leeds.ac.uk> > I agree that the html tag is a bit long-winded. Shorter solution would > be better if poss. > > But in terms of rendering they look fine; > > (Granger, 2013) > > just gets rendered as 'Granger, 2013'. I meant the rendering of things that aren't html tags. If you want to avoid writing html tags this is important - e.g. if we used `\citet{granger2013}`, this currently gets rendered as-is. The natural way around this would be with a javascript citation parser - same idea as mathjax - but this is way too much work I think. For me, it isn't so bad having raw citation commands in my markdown cells. > In the end I've actually opted to use my institution's latex thesis > template, with various post-nbconvert few hacks on the .tex files. If I had > more time I would have preferred to make an nbconvert template that does > the same job; but I don't have that luxury and also am not a latex expert. > > Repo is here > > https://github.com/JohnGriffiths/phd-thesis-template > ...though I wasn't planning on publicising this yet as it isn't > particularly clean and lots of uncommitted changes. You might want to take > a look at this notebook though > > http://nbviewer.ipython.org/github/JohnGriffiths/phd-thesis-template/blob/master/JG_hacks/DoMe.ipynb cheers, that's excellent. Nice to know there is a fall back if I can't get a nbconvert template going. I have some vague ramblings about this here: https://github.com/aaren/notebook-thesis-template Some sort of generic template would be good - putting outputs into figures and the like. This is always going to be customised to what the institution requires, so you'd just keep it simple. I have a few aims: i) reproducibility, ii) not writing in latex, iii) submitting this summer. Some of these might conflict. I'm doing my analysis in the notebook and I'm aiming for that to just transition into a thesis as I flesh it out. It is very easy to procrastinate by spending time thinking about the notebook template! but this is something I need to sort to achieve i and ii. From andrew.gibiansky at gmail.com Tue Mar 18 15:15:36 2014 From: andrew.gibiansky at gmail.com (Andrew Gibiansky) Date: Tue, 18 Mar 2014 12:15:36 -0700 Subject: [IPython-dev] Widgets for non-Python kernels Message-ID: Hey all, I'm trying to implement widgets in IHaskell, and am having a lot of difficulty understanding what should go into this. I currently have Comms implemented, so things can change the way they are displayed to use JS+comms to be dynamic. However, it looks like widgets are significantly more than just that. At least, they are output to a special widget area of the output cell. I understand the general idea behind widgets but cannot understand the raw underlying comm/networking protocol that gets them set up and working. So, a few questions to start with: How can I make, say, a slider widget appear in the frontend? My backend has no Python in it whatsoever. Which side sends the comm_open (backend or frontend)? Is it that the backend sends a comm_open to the frontend with the proper target name, which then creates the model, and later a display() can create the view? I apologize in advance if this is thoroughly documented somewhere; I read the IPEP for backbone widgets and found that it described the architecture but I could not understand how the backend actually dealt with the messages and how and to whom they were sent. Also, the example notebooks are helpful in terms of understanding the Python support for this, but didn't quite answer my questions. Thanks! Andrew Gibiansky -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Tue Mar 18 16:20:31 2014 From: doug.blank at gmail.com (Doug Blank) Date: Tue, 18 Mar 2014 16:20:31 -0400 Subject: [IPython-dev] Widgets for non-Python kernels In-Reply-To: References: Message-ID: On Tue, Mar 18, 2014 at 3:15 PM, Andrew Gibiansky wrote: > Hey all, > > I'm trying to implement widgets in IHaskell, and am having a lot of > difficulty understanding what should go into this. I currently have Comms > implemented, so things can change the way they are displayed to use JS+comms > to be dynamic. > > However, it looks like widgets are significantly more than just that. At > least, they are output to a special widget area of the output cell. I > understand the general idea behind widgets but cannot understand the raw > underlying comm/networking protocol that gets them set up and working. > > So, a few questions to start with: > > How can I make, say, a slider widget appear in the frontend? My backend has > no Python in it whatsoever. > Which side sends the comm_open (backend or frontend)? Is it that the backend > sends a comm_open to the frontend with the proper target name, which then > creates the model, and later a display() can create the view? > > I apologize in advance if this is thoroughly documented somewhere; I read > the IPEP for backbone widgets and found that it described the architecture > but I could not understand how the backend actually dealt with the messages > and how and to whom they were sent. Also, the example notebooks are helpful > in terms of understanding the Python support for this, but didn't quite > answer my questions. > > Thanks! > > Andrew Gibiansky Andrew, We could attempt to share our ZMQServer.cs between Calico and F#, but your native F# is so clean looking that you may not want to :) You can see what is necessary to implement widgets here: https://bitbucket.org/ipre/calico/src/master/Source/Calico/Widgets.cs https://bitbucket.org/ipre/calico/src/master/Source/Calico/ZMQServer.cs This code is still being cleaned up, but I think we have everything implemented (traitlet-like behaviors, compositional widgets, etc). Hope that helps; if not feel free to ask a question. -Doug > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From mail.francesco.rossi+ml at gmail.com Tue Mar 18 16:47:42 2014 From: mail.francesco.rossi+ml at gmail.com (Francesco Rossi) Date: Tue, 18 Mar 2014 21:47:42 +0100 Subject: [IPython-dev] IPyXMPP, a chat bot serving an IPython shell over XMPP Message-ID: Hello all, I am writing a tool for interacting in real time with the data being processed in some simulation codes (running for days). Of course, I find that the best fit for doing that is embedding IPython in the C simulation codes and use its extraordinary features for data analysis. To provide the easiest possible access (without the need to know which cluster the simulation is running on or to set up ssh tunnels) to simple commands, I have written a small chat bot that serves the (embedded) IPython shell over XMPP. The chat bot is here: https://github.com/redsh/IPyXMPP . Has any of you already worked on something similar? Do you think it could be useful in other fields rather than long-simulation monitoring? Which of the existing IPython shell apps would you suggest me to follow as a track for implementing a multi-user/multi-kernel shell over XMPP? Cheers, Francesco Rossi From mark.voorhies at ucsf.edu Tue Mar 18 16:53:54 2014 From: mark.voorhies at ucsf.edu (Mark Voorhies) Date: Tue, 18 Mar 2014 13:53:54 -0700 Subject: [IPython-dev] IPyXMPP, a chat bot serving an IPython shell over XMPP In-Reply-To: References: Message-ID: <5328B262.1030608@ucsf.edu> On 03/18/2014 01:47 PM, Francesco Rossi wrote: > Hello all, > > I am writing a tool for interacting in real time with the data being > processed in some simulation codes (running for days). Of course, I > find that the best fit for doing that is embedding IPython in the C > simulation codes and use its extraordinary features for data analysis. > > To provide the easiest possible access (without the need to know which > cluster the simulation is running on or to set up ssh tunnels) to > simple commands, I have written a small chat bot that serves the > (embedded) IPython shell over XMPP. > > The chat bot is here: https://github.com/redsh/IPyXMPP . > > Has any of you already worked on something similar? This is a bit tangential, but Joey Hess makes interesting use of XMPP in git-annex: http://git-annex.branchable.com/special_remotes/xmpp/ --Mark From jon.freder at gmail.com Tue Mar 18 17:52:52 2014 From: jon.freder at gmail.com (Jonathan Frederic) Date: Tue, 18 Mar 2014 14:52:52 -0700 Subject: [IPython-dev] Widgets for non-Python kernels In-Reply-To: References: Message-ID: Andrew and I had a chance to talk on HipChat. For those who weren't there, I updated the IPEP because it was very out of date ( https://github.com/ipython/ipython/wiki/IPEP-23%3A-Backbone.js-Widgets). It still needs a lot of work but at least the information is correct now. I think this was the cause of the confusion. I added a small section that summarizes the steps required to display a widget. For those of you who are interested, but do not want to visit the full IPEP, see this excerpt: 1. Open a comm with a target_name of WidgetModel. The widget manager watches for comms with the target_name="WidgetModel" and creates a widget model when that comm is created. The comm and widget model are automatically connected. 2. Send a full state update message on the new comm (see the *update* Back-end to front-end message spec in the IPEP). The full state should include the following: - msg_throttle and _view_name - _css and visible if a DOM widget view is specified for _view_name - The rest of the of the properties of the corresponding widget, i.e. disabled, value, min, etc... 3. Send a display message on the comm (see the *display* Back-end to front-end message spec in the IPEP). Cheers, Jon On Tue, Mar 18, 2014 at 1:20 PM, Doug Blank wrote: > On Tue, Mar 18, 2014 at 3:15 PM, Andrew Gibiansky > wrote: > > Hey all, > > > > I'm trying to implement widgets in IHaskell, and am having a lot of > > difficulty understanding what should go into this. I currently have Comms > > implemented, so things can change the way they are displayed to use > JS+comms > > to be dynamic. > > > > However, it looks like widgets are significantly more than just that. At > > least, they are output to a special widget area of the output cell. I > > understand the general idea behind widgets but cannot understand the raw > > underlying comm/networking protocol that gets them set up and working. > > > > So, a few questions to start with: > > > > How can I make, say, a slider widget appear in the frontend? My backend > has > > no Python in it whatsoever. > > Which side sends the comm_open (backend or frontend)? Is it that the > backend > > sends a comm_open to the frontend with the proper target name, which then > > creates the model, and later a display() can create the view? > > > > I apologize in advance if this is thoroughly documented somewhere; I read > > the IPEP for backbone widgets and found that it described the > architecture > > but I could not understand how the backend actually dealt with the > messages > > and how and to whom they were sent. Also, the example notebooks are > helpful > > in terms of understanding the Python support for this, but didn't quite > > answer my questions. > > > > Thanks! > > > > Andrew Gibiansky > > Andrew, > > We could attempt to share our ZMQServer.cs between Calico and F#, but > your native F# is so clean looking that you may not want to :) You can > see what is necessary to implement widgets here: > > https://bitbucket.org/ipre/calico/src/master/Source/Calico/Widgets.cs > https://bitbucket.org/ipre/calico/src/master/Source/Calico/ZMQServer.cs > > This code is still being cleaned up, but I think we have everything > implemented (traitlet-like behaviors, compositional widgets, etc). > > Hope that helps; if not feel free to ask a question. > > -Doug > > > > > > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From damontallen at gmail.com Tue Mar 18 17:56:18 2014 From: damontallen at gmail.com (Damon Allen) Date: Tue, 18 Mar 2014 17:56:18 -0400 Subject: [IPython-dev] Embedded Videos in Slideshows Message-ID: Hi, Does running a slideshow force embedded YouTube videos to play in flash instead of HTML5? I've found that in while running Firefox in Ubuntu I cannot play YouTube videos in a slideshow. I can watch the video in Firefox if it isn't in a slideshow and if I the slideshow runs in Chrome everything works. This isn't a big deal since most of my students are running Windows and they will be able to use up-to-date versions of the flashplayer, but I am curious if there could be a fix to either reveal or nbconvert that would correct this. As it stands, I need to switch my development from using Firefox to Chrome. I'm still using IPython 1.2.1. Thanks, Damon -------------- next part -------------- An HTML attachment was scrubbed... URL: From damianavila at gmail.com Tue Mar 18 18:13:45 2014 From: damianavila at gmail.com (=?ISO-8859-1?Q?Dami=E1n_Avila?=) Date: Tue, 18 Mar 2014 19:13:45 -0300 Subject: [IPython-dev] Embedded Videos in Slideshows In-Reply-To: References: Message-ID: Have you seen this problem using the slideshow generated with IPython.nbconvert? It can be an issue with reveal.js itself... can you provided me with an example notebook to study the case? Thanks. 2014-03-18 18:56 GMT-03:00 Damon Allen : > Hi, > > Does running a slideshow force embedded YouTube videos to play in flash > instead of HTML5? I've found that in while running Firefox in Ubuntu I > cannot play YouTube videos in a slideshow. I can watch the video in > Firefox if it isn't in a slideshow and if I the slideshow runs in Chrome > everything works. > > This isn't a big deal since most of my students are running Windows and > they will be able to use up-to-date versions of the flashplayer, but I am > curious if there could be a fix to either reveal or nbconvert that would > correct this. > > As it stands, I need to switch my development from using Firefox to Chrome. > > I'm still using IPython 1.2.1. > > Thanks, > > Damon > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -- Dami?n Avila Scientific Python Developer Quantitative Finance Analyst Statistics, Biostatistics and Econometrics Consultant Biochemist -------------- next part -------------- An HTML attachment was scrubbed... URL: From damontallen at gmail.com Tue Mar 18 18:20:27 2014 From: damontallen at gmail.com (Damon Allen) Date: Tue, 18 Mar 2014 18:20:27 -0400 Subject: [IPython-dev] Embedded Videos in Slideshows In-Reply-To: References: Message-ID: The notebook is at: http://nbviewer.ipython.org/github/damontallen/Construction-Lectures/blob/master/Week%200%20-%20Greetings.ipynb The slideshow is at: http://damontallen.github.io/Construction-Lectures/Week%200%20-%20Greetings.slides.html#/ And this problem also occurs when I use ipython nbconvert "Week 0 - Greetings.ipynb" --to slides --post serve Thanks for looking into it. Damon T. Allen Ph.D. Adjunct Professor (352) 234-3266 damontallen at gmail.com 344 Rinker Hall College of Construction Management University of Florida On Tue, Mar 18, 2014 at 6:13 PM, Dami?n Avila wrote: > Have you seen this problem using the slideshow generated with > IPython.nbconvert? > It can be an issue with reveal.js itself... can you provided me with an > example notebook to study the case? > > Thanks. > > > > > 2014-03-18 18:56 GMT-03:00 Damon Allen : > >> Hi, >> >> Does running a slideshow force embedded YouTube videos to play in flash >> instead of HTML5? I've found that in while running Firefox in Ubuntu I >> cannot play YouTube videos in a slideshow. I can watch the video in >> Firefox if it isn't in a slideshow and if I the slideshow runs in Chrome >> everything works. >> >> This isn't a big deal since most of my students are running Windows and >> they will be able to use up-to-date versions of the flashplayer, but I am >> curious if there could be a fix to either reveal or nbconvert that would >> correct this. >> >> As it stands, I need to switch my development from using Firefox to >> Chrome. >> >> I'm still using IPython 1.2.1. >> >> Thanks, >> >> Damon >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > > -- > Dami?n Avila > Scientific Python Developer > Quantitative Finance Analyst > Statistics, Biostatistics and Econometrics Consultant > Biochemist > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From alessandro.gagliardi at glassdoor.com Tue Mar 18 18:24:56 2014 From: alessandro.gagliardi at glassdoor.com (Alessandro Gagliardi) Date: Tue, 18 Mar 2014 22:24:56 +0000 Subject: [IPython-dev] IPython for Education Message-ID: I?ve been impressed by the use of IPython Notebook in tutorials at SciPy and other places and have been using slides generated using nbconvert for my Data Science class at General Assembly and I am interested in taking it to the next level. I feel like there?s a lot of potential, especially with widgets (and someone mentioned XMPP) for bringing interactivity into the classroom, though I?m not sure if I have the knowhow to pull it off myself. One example of something that would be cool is if I (as instructor) could look at what students are working on and easily aggregate their results for presentation to the class. (So, say the students are working on a classifier. Aggregate the results of each student?s classification accuracy and present them back to the class.) This is maybe not the best example, it?s just the first that came to mind. Wondering if anyone is working on anything like this. By ?this? I mean extensions to IPython specifically for the classroom (could be something completely unrelated to my specific example). Thanks, -Alessandro -------------- next part -------------- An HTML attachment was scrubbed... URL: From damianavila at gmail.com Tue Mar 18 20:11:34 2014 From: damianavila at gmail.com (=?ISO-8859-1?Q?Dami=E1n_Avila?=) Date: Tue, 18 Mar 2014 21:11:34 -0300 Subject: [IPython-dev] Embedded Videos in Slideshows In-Reply-To: References: Message-ID: I ask to google... and I found I have already answered myself... ja ja https://github.com/damianavila/live_reveal/issues/4 not a good sign, I think ;-) 2014-03-18 19:20 GMT-03:00 Damon Allen : > The notebook is at: > > > http://nbviewer.ipython.org/github/damontallen/Construction-Lectures/blob/master/Week%200%20-%20Greetings.ipynb > > The slideshow is at: > > > http://damontallen.github.io/Construction-Lectures/Week%200%20-%20Greetings.slides.html#/ > > And this problem also occurs when I use > > ipython nbconvert "Week 0 - Greetings.ipynb" --to slides --post serve > > Thanks for looking into it. > > > > Damon T. Allen Ph.D. > Adjunct Professor > (352) 234-3266 > damontallen at gmail.com > 344 Rinker Hall > College of Construction Management > University of Florida > > > On Tue, Mar 18, 2014 at 6:13 PM, Dami?n Avila wrote: > >> Have you seen this problem using the slideshow generated with >> IPython.nbconvert? >> It can be an issue with reveal.js itself... can you provided me with an >> example notebook to study the case? >> >> Thanks. >> >> >> >> >> 2014-03-18 18:56 GMT-03:00 Damon Allen : >> >>> Hi, >>> >>> Does running a slideshow force embedded YouTube videos to play in flash >>> instead of HTML5? I've found that in while running Firefox in Ubuntu I >>> cannot play YouTube videos in a slideshow. I can watch the video in >>> Firefox if it isn't in a slideshow and if I the slideshow runs in Chrome >>> everything works. >>> >>> This isn't a big deal since most of my students are running Windows and >>> they will be able to use up-to-date versions of the flashplayer, but I am >>> curious if there could be a fix to either reveal or nbconvert that would >>> correct this. >>> >>> As it stands, I need to switch my development from using Firefox to >>> Chrome. >>> >>> I'm still using IPython 1.2.1. >>> >>> Thanks, >>> >>> Damon >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >>> >> >> >> -- >> Dami?n Avila >> Scientific Python Developer >> Quantitative Finance Analyst >> Statistics, Biostatistics and Econometrics Consultant >> Biochemist >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -- Dami?n Avila Scientific Python Developer Quantitative Finance Analyst Statistics, Biostatistics and Econometrics Consultant Biochemist -------------- next part -------------- An HTML attachment was scrubbed... URL: From damontallen at gmail.com Tue Mar 18 20:42:03 2014 From: damontallen at gmail.com (Damon Allen) Date: Tue, 18 Mar 2014 20:42:03 -0400 Subject: [IPython-dev] Embedded Videos in Slideshows In-Reply-To: References: Message-ID: Thanks for answering my question so promptly. I'll tell my students to use Chrome if it is an issue for them. Damon T. Allen Ph.D. Adjunct Professor (352) 234-3266 damontallen at gmail.com 344 Rinker Hall College of Construction Management University of Florida On Tue, Mar 18, 2014 at 8:11 PM, Dami?n Avila wrote: > I ask to google... and I found I have already answered myself... ja ja > > https://github.com/damianavila/live_reveal/issues/4 > > not a good sign, I think ;-) > > > 2014-03-18 19:20 GMT-03:00 Damon Allen : > > The notebook is at: >> >> >> http://nbviewer.ipython.org/github/damontallen/Construction-Lectures/blob/master/Week%200%20-%20Greetings.ipynb >> >> The slideshow is at: >> >> >> http://damontallen.github.io/Construction-Lectures/Week%200%20-%20Greetings.slides.html#/ >> >> And this problem also occurs when I use >> >> ipython nbconvert "Week 0 - Greetings.ipynb" --to slides --post serve >> >> Thanks for looking into it. >> >> >> >> Damon T. Allen Ph.D. >> Adjunct Professor >> (352) 234-3266 >> damontallen at gmail.com >> 344 Rinker Hall >> College of Construction Management >> University of Florida >> >> >> On Tue, Mar 18, 2014 at 6:13 PM, Dami?n Avila wrote: >> >>> Have you seen this problem using the slideshow generated with >>> IPython.nbconvert? >>> It can be an issue with reveal.js itself... can you provided me with an >>> example notebook to study the case? >>> >>> Thanks. >>> >>> >>> >>> >>> 2014-03-18 18:56 GMT-03:00 Damon Allen : >>> >>>> Hi, >>>> >>>> Does running a slideshow force embedded YouTube videos to play in flash >>>> instead of HTML5? I've found that in while running Firefox in Ubuntu I >>>> cannot play YouTube videos in a slideshow. I can watch the video in >>>> Firefox if it isn't in a slideshow and if I the slideshow runs in Chrome >>>> everything works. >>>> >>>> This isn't a big deal since most of my students are running Windows and >>>> they will be able to use up-to-date versions of the flashplayer, but I am >>>> curious if there could be a fix to either reveal or nbconvert that would >>>> correct this. >>>> >>>> As it stands, I need to switch my development from using Firefox to >>>> Chrome. >>>> >>>> I'm still using IPython 1.2.1. >>>> >>>> Thanks, >>>> >>>> Damon >>>> >>>> _______________________________________________ >>>> IPython-dev mailing list >>>> IPython-dev at scipy.org >>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>>> >>>> >>> >>> >>> -- >>> Dami?n Avila >>> Scientific Python Developer >>> Quantitative Finance Analyst >>> Statistics, Biostatistics and Econometrics Consultant >>> Biochemist >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >>> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > > -- > Dami?n Avila > Scientific Python Developer > Quantitative Finance Analyst > Statistics, Biostatistics and Econometrics Consultant > Biochemist > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From austin.bingham at gmail.com Wed Mar 19 06:01:57 2014 From: austin.bingham at gmail.com (Austin Bingham) Date: Wed, 19 Mar 2014 11:01:57 +0100 Subject: [IPython-dev] Reloading compiled modules in a notebook Message-ID: (I apologize if this gets doubly-posted. I sent this earlier but it never seemed to show up on the list.) I've been scratching my head over this for a while now. I've created a "magic" command which compiles Boost.Python C++ into importable extension modules. In a notebook, I use that magic in one cell to compile the module, and in the next cell (a normal Python cell) I import that module. By and large this seems to work just fine. However, if I change the code in the C++ cell and rebuild the extension module, I can't find a way to convince the Python cell to reimport the module. I've tried autoreload, aimport, reload, dreload, and everything else I can think of, but ipython just doesn't seem to want to reload/reimport the module. I've verified that the module is being correctly rebuilt at the right times. That is, I'm able to change the C++ code in the notebook, rerun the cell, and then import the generated module in a separate Python instance. When I do this I see the results I expect. So somehow either IPython isn't able to reload extensions modules, or I'm doing something wrong (entirely possible!) So, is this something I should expect to be able to do? And if so, does anyone have any ideas on what else I should try? On the off-chance that it makes any difference, the notebook and "magic" implementation are here: https://github.com/abingham/boost_python_tutorial Austin From p.f.moore at gmail.com Wed Mar 19 06:28:34 2014 From: p.f.moore at gmail.com (Paul Moore) Date: Wed, 19 Mar 2014 10:28:34 +0000 Subject: [IPython-dev] Reloading compiled modules in a notebook In-Reply-To: References: Message-ID: On 19 March 2014 10:01, Austin Bingham wrote: > So, is this something I should expect to be able to do? And if so, > does anyone have any ideas on what else I should try? I don't believe CPython has the ability to reload an extension module into a running interpreter (i.e., it's not an IPython limitation, but a Python one). I think you'd need to restart your kernel to load a new version of an extension. Paul From austin.bingham at gmail.com Wed Mar 19 07:16:45 2014 From: austin.bingham at gmail.com (Austin Bingham) Date: Wed, 19 Mar 2014 12:16:45 +0100 Subject: [IPython-dev] Reloading compiled modules in a notebook In-Reply-To: References: Message-ID: Ah, it appears you're right. I should have checked this first. Thanks for helping me sort this out. On Mar 19, 2014 11:28 AM, "Paul Moore" wrote: > On 19 March 2014 10:01, Austin Bingham wrote: > > So, is this something I should expect to be able to do? And if so, > > does anyone have any ideas on what else I should try? > > I don't believe CPython has the ability to reload an extension module > into a running interpreter (i.e., it's not an IPython limitation, but > a Python one). I think you'd need to restart your kernel to load a new > version of an extension. > > Paul > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From tra at popgen.net Wed Mar 19 07:26:07 2014 From: tra at popgen.net (Tiago Antao) Date: Wed, 19 Mar 2014 11:26:07 +0000 Subject: [IPython-dev] Reloading compiled modules in a notebook In-Reply-To: References: Message-ID: <20140319112607.11fec2e6@grandao> On Wed, 19 Mar 2014 12:16:45 +0100 Austin Bingham wrote: > Ah, it appears you're right. I should have checked this first. Thanks > for helping me sort this out. I think Python 3 allows you to reload modules: http://docs.python.org/3/library/imp.html#imp.reload and, since 3.4 http://docs.python.org/3/library/importlib.html#importlib.reload Of course 2 might be a completely different affair... Tiago From robert.kern at gmail.com Wed Mar 19 08:00:36 2014 From: robert.kern at gmail.com (Robert Kern) Date: Wed, 19 Mar 2014 12:00:36 +0000 Subject: [IPython-dev] Reloading compiled modules in a notebook In-Reply-To: <20140319112607.11fec2e6@grandao> References: <20140319112607.11fec2e6@grandao> Message-ID: On 2014-03-19 11:26, Tiago Antao wrote: > On Wed, 19 Mar 2014 12:16:45 +0100 > Austin Bingham wrote: > >> Ah, it appears you're right. I should have checked this first. Thanks >> for helping me sort this out. > > I think Python 3 allows you to reload modules: Not extension modules. > http://docs.python.org/3/library/imp.html#imp.reload > > and, since 3.4 > > http://docs.python.org/3/library/importlib.html#importlib.reload "The init function of extension modules is not called a second time." "In many cases, however, extension modules are not designed to be initialized more than once, and may fail in arbitrary ways when reloaded." > Of course 2 might be a completely different affair... It's exactly the same, except for where the reload() function is found: http://docs.python.org/2/library/functions.html#reload -- Robert Kern "I have come to believe that the whole world is an enigma, a harmless enigma that is made terrible by our own mad attempt to interpret it as though it had an underlying truth." -- Umberto Eco From Kevin.D.Winters at erdc.dren.mil Wed Mar 19 10:04:30 2014 From: Kevin.D.Winters at erdc.dren.mil (Winters, Kevin D ERDC-RDE-CHL-MS) Date: Wed, 19 Mar 2014 14:04:30 +0000 Subject: [IPython-dev] Proper use of accordion widget Message-ID: <6F84B4D0-4DA9-4A60-9579-3C7978755D61@erdc.dren.mil> I'm attempting to use an accordion widget in my notebook, however I can't get the contents to display. I assumed that the AccordionWidget should be defined similarly to the TabWidget. Here's my test code: from IPython.html import widgets # Widget definitions from IPython.display import display # Used to display widgets in the notebook acc = widgets.AccordionWidget() #acc = widgets.TabWidget() this = widgets.ContainerWidget() this.children = [widgets.CheckboxWidget(description='this', value=True)] that = widgets.ContainerWidget() that.children = [widgets.CheckboxWidget(description='that', value=False)] acc.children = [this, that] dictPages = {0:'This', 1:"That"} display(acc) acc.set_css({'border':'3px solid black', 'height':'500px', 'width':'500px'}) for page in dictPages: acc.set_title(page, dictPages[page]) for page in acc.children: page.set_css({'height':'150px'}) acc.selected_index = 1 Am I missing something or is there a bug in the accordion display? Thanks, Kevin -------------------- Kevin Winters Research Hydraulic Engineer US Army Corps of Engineers Engineer Research and Development Center 3909 Halls Ferry Road Vicksburg, MS 39180 601.634.2102 -------------- next part -------------- An HTML attachment was scrubbed... URL: From mark.voorhies at ucsf.edu Wed Mar 19 19:24:48 2014 From: mark.voorhies at ucsf.edu (Mark Voorhies) Date: Wed, 19 Mar 2014 16:24:48 -0700 Subject: [IPython-dev] Block indent in CodeMirror Message-ID: <532A2740.5010700@ucsf.edu> Just came across this (apologies if it's common knowledge): highlighting a block of code in the notebook and hitting tab indents the full block -- now I've got one fewer reason to keep hopping my code in and out of emacs =) --Mark From fperez.net at gmail.com Wed Mar 19 22:39:08 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Wed, 19 Mar 2014 19:39:08 -0700 Subject: [IPython-dev] Block indent in CodeMirror In-Reply-To: <532A2740.5010700@ucsf.edu> References: <532A2740.5010700@ucsf.edu> Message-ID: And shift-tab unindents, so it's easy to move code right/left while in the NB. Best On Wed, Mar 19, 2014 at 4:24 PM, Mark Voorhies wrote: > Just came across this (apologies if it's common knowledge): > highlighting a block of code in the notebook and hitting tab > indents the full block -- now I've got one fewer reason to > keep hopping my code in and out of emacs =) > > --Mark > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From ellisonbg at gmail.com Wed Mar 19 23:03:51 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Wed, 19 Mar 2014 20:03:51 -0700 Subject: [IPython-dev] Reloading compiled modules in a notebook In-Reply-To: References: <20140319112607.11fec2e6@grandao> Message-ID: Have a look at the Cython magic, we do this type of thing there. The basic idea is that we build each version with a uuid so each version is a completely separate extension module. https://github.com/ipython/ipython/blob/master/IPython/extensions/cythonmagic.py Cheers, Brian On Wed, Mar 19, 2014 at 5:00 AM, Robert Kern wrote: > On 2014-03-19 11:26, Tiago Antao wrote: >> On Wed, 19 Mar 2014 12:16:45 +0100 >> Austin Bingham wrote: >> >>> Ah, it appears you're right. I should have checked this first. Thanks >>> for helping me sort this out. >> >> I think Python 3 allows you to reload modules: > > Not extension modules. > >> http://docs.python.org/3/library/imp.html#imp.reload >> >> and, since 3.4 >> >> http://docs.python.org/3/library/importlib.html#importlib.reload > > "The init function of extension modules is not called a second time." > > "In many cases, however, extension modules are not designed to be initialized > more than once, and may fail in arbitrary ways when reloaded." > >> Of course 2 might be a completely different affair... > > It's exactly the same, except for where the reload() function is found: > > http://docs.python.org/2/library/functions.html#reload > > -- > Robert Kern > > "I have come to believe that the whole world is an enigma, a harmless enigma > that is made terrible by our own mad attempt to interpret it as though it had > an underlying truth." > -- Umberto Eco > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From ellisonbg at gmail.com Wed Mar 19 23:03:51 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Wed, 19 Mar 2014 20:03:51 -0700 Subject: [IPython-dev] Reloading compiled modules in a notebook In-Reply-To: References: <20140319112607.11fec2e6@grandao> Message-ID: Have a look at the Cython magic, we do this type of thing there. The basic idea is that we build each version with a uuid so each version is a completely separate extension module. https://github.com/ipython/ipython/blob/master/IPython/extensions/cythonmagic.py Cheers, Brian On Wed, Mar 19, 2014 at 5:00 AM, Robert Kern wrote: > On 2014-03-19 11:26, Tiago Antao wrote: >> On Wed, 19 Mar 2014 12:16:45 +0100 >> Austin Bingham wrote: >> >>> Ah, it appears you're right. I should have checked this first. Thanks >>> for helping me sort this out. >> >> I think Python 3 allows you to reload modules: > > Not extension modules. > >> http://docs.python.org/3/library/imp.html#imp.reload >> >> and, since 3.4 >> >> http://docs.python.org/3/library/importlib.html#importlib.reload > > "The init function of extension modules is not called a second time." > > "In many cases, however, extension modules are not designed to be initialized > more than once, and may fail in arbitrary ways when reloaded." > >> Of course 2 might be a completely different affair... > > It's exactly the same, except for where the reload() function is found: > > http://docs.python.org/2/library/functions.html#reload > > -- > Robert Kern > > "I have come to believe that the whole world is an enigma, a harmless enigma > that is made terrible by our own mad attempt to interpret it as though it had > an underlying truth." > -- Umberto Eco > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From doug.blank at gmail.com Thu Mar 20 12:11:02 2014 From: doug.blank at gmail.com (Doug Blank) Date: Thu, 20 Mar 2014 12:11:02 -0400 Subject: [IPython-dev] Proper use of accordion widget In-Reply-To: <6F84B4D0-4DA9-4A60-9579-3C7978755D61@erdc.dren.mil> References: <6F84B4D0-4DA9-4A60-9579-3C7978755D61@erdc.dren.mil> Message-ID: I can confirm that this does not work as expected: from IPython.html.widgets import AccordionWidget, ButtonWidget aw = AccordionWidget(children=[ButtonWidget()]) aw while this does: from IPython.html.widgets import TabWidget, ButtonWidget tw = TabWidget(children=[ButtonWidget()]) tw -Doug On Wed, Mar 19, 2014 at 10:04 AM, Winters, Kevin D ERDC-RDE-CHL-MS wrote: > I'm attempting to use an accordion widget in my notebook, however I can't > get the contents to display. I assumed that the AccordionWidget should be > defined similarly to the TabWidget. Here's my test code: > > from IPython.html import widgets # Widget definitions > from IPython.display import display # Used to display widgets in the > notebook > > acc = widgets.AccordionWidget() > #acc = widgets.TabWidget() > > this = widgets.ContainerWidget() > this.children = [widgets.CheckboxWidget(description='this', value=True)] > that = widgets.ContainerWidget() > that.children = [widgets.CheckboxWidget(description='that', value=False)] > > acc.children = [this, that] > dictPages = {0:'This', 1:"That"} > > display(acc) > acc.set_css({'border':'3px solid black', 'height':'500px', 'width':'500px'}) > for page in dictPages: > acc.set_title(page, dictPages[page]) > for page in acc.children: > page.set_css({'height':'150px'}) > acc.selected_index = 1 > > Am I missing something or is there a bug in the accordion display? > > Thanks, > > Kevin > -------------------- > Kevin Winters > Research Hydraulic Engineer > US Army Corps of Engineers > Engineer Research and Development Center > 3909 Halls Ferry Road > Vicksburg, MS 39180 > 601.634.2102 > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From doug.blank at gmail.com Thu Mar 20 12:17:49 2014 From: doug.blank at gmail.com (Doug Blank) Date: Thu, 20 Mar 2014 12:17:49 -0400 Subject: [IPython-dev] IPython for Education In-Reply-To: References: Message-ID: Alessandro, I think you are right that there are many possibilities for using IPython in the classroom. I think many of these could be explored more easily if the students were logged in via an IPython server... at least that is where I will be concentrating my explorations this year. I could imagine: sharing, commenting, grading, hand-in methods, testing, quizzes, and clicker-style anonymous polling ("are we all at the same place? ready to move on?"), etc. It would be great to make a list of such feature ideas, and prioritize.... -Doug On Tue, Mar 18, 2014 at 6:24 PM, Alessandro Gagliardi wrote: > I've been impressed by the use of IPython Notebook in tutorials at SciPy and > other places and have been using slides generated using nbconvert for my > Data Science class at General Assembly and I am interested in taking it to > the next level. I feel like there's a lot of potential, especially with > widgets (and someone mentioned XMPP) for bringing interactivity into the > classroom, though I'm not sure if I have the knowhow to pull it off myself. > One example of something that would be cool is if I (as instructor) could > look at what students are working on and easily aggregate their results for > presentation to the class. (So, say the students are working on a > classifier. Aggregate the results of each student's classification accuracy > and present them back to the class.) This is maybe not the best example, > it's just the first that came to mind. Wondering if anyone is working on > anything like this. By "this" I mean extensions to IPython specifically for > the classroom (could be something completely unrelated to my specific > example). > > Thanks, > -Alessandro > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From doug.blank at gmail.com Thu Mar 20 12:32:58 2014 From: doug.blank at gmail.com (Doug Blank) Date: Thu, 20 Mar 2014 12:32:58 -0400 Subject: [IPython-dev] Block indent in CodeMirror In-Reply-To: References: <532A2740.5010700@ucsf.edu> Message-ID: On Wed, Mar 19, 2014 at 10:39 PM, Fernando Perez wrote: > And shift-tab unindents, so it's easy to move code right/left while in the > NB. Ah, but you must have the line(s) selected... I was wondering why it wasn't working for me. This can be confusing for new users, as shift+tab is also help. Perhaps control+], control+[ are better manipulators of indents as they work more robustly (eg, even when nothing is selected). -Doug > Best > > > On Wed, Mar 19, 2014 at 4:24 PM, Mark Voorhies > wrote: >> >> Just came across this (apologies if it's common knowledge): >> highlighting a block of code in the notebook and hitting tab >> indents the full block -- now I've got one fewer reason to >> keep hopping my code in and out of emacs =) >> >> --Mark >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > > -- > Fernando Perez (@fperez_org; http://fperez.org) > fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) > fernando.perez-at-berkeley: contact me here for any direct mail > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From jon.freder at gmail.com Thu Mar 20 12:56:25 2014 From: jon.freder at gmail.com (Jonathan Frederic) Date: Thu, 20 Mar 2014 09:56:25 -0700 Subject: [IPython-dev] Proper use of accordion widget In-Reply-To: References: <6F84B4D0-4DA9-4A60-9579-3C7978755D61@erdc.dren.mil> Message-ID: Thanks for reporting this guys. I opened an issue: https://github.com/ipython/ipython/issues/5397 Cheers, Jon On Thu, Mar 20, 2014 at 9:11 AM, Doug Blank wrote: > I can confirm that this does not work as expected: > > from IPython.html.widgets import AccordionWidget, ButtonWidget > aw = AccordionWidget(children=[ButtonWidget()]) > aw > > while this does: > > from IPython.html.widgets import TabWidget, ButtonWidget > tw = TabWidget(children=[ButtonWidget()]) > tw > > -Doug > > On Wed, Mar 19, 2014 at 10:04 AM, Winters, Kevin D ERDC-RDE-CHL-MS > wrote: > > I'm attempting to use an accordion widget in my notebook, however I can't > > get the contents to display. I assumed that the AccordionWidget should be > > defined similarly to the TabWidget. Here's my test code: > > > > from IPython.html import widgets # Widget definitions > > from IPython.display import display # Used to display widgets in the > > notebook > > > > acc = widgets.AccordionWidget() > > #acc = widgets.TabWidget() > > > > this = widgets.ContainerWidget() > > this.children = [widgets.CheckboxWidget(description='this', value=True)] > > that = widgets.ContainerWidget() > > that.children = [widgets.CheckboxWidget(description='that', value=False)] > > > > acc.children = [this, that] > > dictPages = {0:'This', 1:"That"} > > > > display(acc) > > acc.set_css({'border':'3px solid black', 'height':'500px', > 'width':'500px'}) > > for page in dictPages: > > acc.set_title(page, dictPages[page]) > > for page in acc.children: > > page.set_css({'height':'150px'}) > > acc.selected_index = 1 > > > > Am I missing something or is there a bug in the accordion display? > > > > Thanks, > > > > Kevin > > -------------------- > > Kevin Winters > > Research Hydraulic Engineer > > US Army Corps of Engineers > > Engineer Research and Development Center > > 3909 Halls Ferry Road > > Vicksburg, MS 39180 > > 601.634.2102 > > > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From alessandro.gagliardi at glassdoor.com Thu Mar 20 13:06:31 2014 From: alessandro.gagliardi at glassdoor.com (Alessandro Gagliardi) Date: Thu, 20 Mar 2014 17:06:31 +0000 Subject: [IPython-dev] IPython for Education In-Reply-To: Message-ID: I was thinking about that. But my understanding is that when multiple people are logged into the same IPython server, this might cause problems for the server. Particularly if they are using the same notebook, it seems like the kernel could easily get into an inconsistent state. Were you imagining that the students would each be using their own notebook but hosted on a single server? Or are there ways to have multiple users working on the same notebook without it causing problems? Date: Thu, 20 Mar 2014 12:17:49 -0400 From: Doug Blank > Subject: Re: [IPython-dev] IPython for Education To: IPython developers list > Message-ID: > Content-Type: text/plain; charset=ISO-8859-1 Alessandro, I think you are right that there are many possibilities for using IPython in the classroom. I think many of these could be explored more easily if the students were logged in via an IPython server... at least that is where I will be concentrating my explorations this year. I could imagine: sharing, commenting, grading, hand-in methods, testing, quizzes, and clicker-style anonymous polling ("are we all at the same place? ready to move on?"), etc. It would be great to make a list of such feature ideas, and prioritize.... -Doug -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Thu Mar 20 13:30:53 2014 From: doug.blank at gmail.com (Doug Blank) Date: Thu, 20 Mar 2014 13:30:53 -0400 Subject: [IPython-dev] IPython for Education In-Reply-To: References: Message-ID: On Thu, Mar 20, 2014 at 1:06 PM, Alessandro Gagliardi wrote: > I was thinking about that. But my understanding is that when multiple people > are logged into the same IPython server, this might cause problems for the > server. Particularly if they are using the same notebook, it seems like the > kernel could easily get into an inconsistent state. Were you imagining that > the students would each be using their own notebook but hosted on a single > server? Or are there ways to have multiple users working on the same > notebook without it causing problems? I am imagining a Notebook Server where each user logs in with unique ID/password and controls their own kernels/clusters (perhaps similar to how Sage Cloud with IPython works). Students would be members of at least one, maybe more, classes. The server would need to be able to run (or farm out) at least one kernel per student. At that point, one can start to think about sharing, and interacting with students as users. A homework submission could just be a menu option. I guess I am thinking about a multi-user system, perhaps like Drupal/Wordpress built on Tornado for managing users as students/ta's/teachers/graders etc. -Doug > Date: Thu, 20 Mar 2014 12:17:49 -0400 > From: Doug Blank > Subject: Re: [IPython-dev] IPython for Education > To: IPython developers list > Message-ID: > > Content-Type: text/plain; charset=ISO-8859-1 > > Alessandro, > > I think you are right that there are many possibilities for using > IPython in the classroom. I think many of these could be explored more > easily if the students were logged in via an IPython server... at > least that is where I will be concentrating my explorations this year. > I could imagine: sharing, commenting, grading, hand-in methods, > testing, quizzes, and clicker-style anonymous polling ("are we all at > the same place? ready to move on?"), etc. It would be great to make a > list of such feature ideas, and prioritize.... > > -Doug > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From tra at popgen.net Thu Mar 20 13:34:06 2014 From: tra at popgen.net (Tiago Antao) Date: Thu, 20 Mar 2014 17:34:06 +0000 Subject: [IPython-dev] IPython for Education In-Reply-To: References: Message-ID: <20140320173406.145457b9@grandao> On Thu, 20 Mar 2014 17:06:31 +0000 Alessandro Gagliardi wrote: > I was thinking about that. But my understanding is that when multiple > people are logged into the same IPython server, this might cause > problems for the server. Particularly if they are using the same > notebook, it seems like the kernel could easily get into an > inconsistent state. Were you imagining that the students would each > be using their own notebook but hosted on a single server? Or are > there ways to have multiple users working on the same notebook > without it causing problems? In a teaching environment it might be interesting to have a setup where people could copy from other server users (very similar to github forks). Initially maybe having some notebooks from the teacher which everyone could copy and hack on. But then having students sharing contents among themselves via forking. Sorry for the brainstorming, but I am involved in discussion precisely to create notebooks for (self)teaching (biopython) and this seemed relevant... Tiago From doug.blank at gmail.com Thu Mar 20 13:58:16 2014 From: doug.blank at gmail.com (Doug Blank) Date: Thu, 20 Mar 2014 13:58:16 -0400 Subject: [IPython-dev] Kernels and parallel support? Message-ID: After reading: http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html and related pages, I am unsure what the relationship between parallel processing and kernels are. Can external kernels be used in a parallel environment? Or is this part of IPython python-kernel specific? Does it use ZMQ messages? Thanks for any additional data, -Doug From benjaminrk at gmail.com Thu Mar 20 14:19:53 2014 From: benjaminrk at gmail.com (MinRK) Date: Thu, 20 Mar 2014 11:19:53 -0700 Subject: [IPython-dev] Kernels and parallel support? In-Reply-To: References: Message-ID: On Thu, Mar 20, 2014 at 10:58 AM, Doug Blank wrote: > After reading: > > http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html > > and related pages, I am unsure what the relationship between parallel > processing and kernels are. Can external kernels be used in a parallel > environment? Or is this part of IPython python-kernel specific? Does > it use ZMQ messages? > Depends what you mean by 'external'. You can start engines at any time on any machine, and they can join a cluster. Yes, IPython.parallel uses zmq. -MinRK > > Thanks for any additional data, > > -Doug > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Thu Mar 20 16:06:53 2014 From: doug.blank at gmail.com (Doug Blank) Date: Thu, 20 Mar 2014 16:06:53 -0400 Subject: [IPython-dev] Kernels and parallel support? In-Reply-To: References: Message-ID: On Thu, Mar 20, 2014 at 2:19 PM, MinRK wrote: > > > > On Thu, Mar 20, 2014 at 10:58 AM, Doug Blank wrote: >> >> After reading: >> >> http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html >> >> and related pages, I am unsure what the relationship between parallel >> processing and kernels are. Can external kernels be used in a parallel >> environment? Or is this part of IPython python-kernel specific? Does >> it use ZMQ messages? > > > Depends what you mean by 'external'. You can start engines at any time on > any machine, and they can join a cluster. Yes, IPython.parallel uses zmq. If I wanted to explore using a cluster of 3rd-party kernels, such as IHaskell or ICalico, in a parallel environment, where would I begin? Has anyone done this before? I see here: http://ipython.org/ipython-doc/dev/parallel/magics.html#engines-as-kernels that kernels connect to a controller to create an engine. If the standard ipcontrollers can start up external kernels, I guess a place to begin might be implementing what the %px magic does? -Doug > -MinRK > >> >> >> Thanks for any additional data, >> >> -Doug >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From zvoros at gmail.com Thu Mar 20 16:10:35 2014 From: zvoros at gmail.com (=?ISO-8859-1?Q?Zolt=E1n_V=F6r=F6s?=) Date: Thu, 20 Mar 2014 21:10:35 +0100 Subject: [IPython-dev] toggle comment doesn't work Message-ID: <532B4B3B.20906@gmail.com> Hi all, I have just noticed in the most recent version from master that toggle comment (Cntr+/) doesn't work any more, neither in firefox, nor in chrome. Is that a design decision (it is still listed in the help), or a bug? Cheers, Zolt?n -------------- next part -------------- An HTML attachment was scrubbed... URL: From scott.s.burns at vanderbilt.edu Thu Mar 20 16:16:00 2014 From: scott.s.burns at vanderbilt.edu (Scott Burns) Date: Thu, 20 Mar 2014 15:16:00 -0500 Subject: [IPython-dev] Notebook connection timeouts Message-ID: <6D242913-3886-4A15-B065-2DEB963607FA@vanderbilt.edu> Long-running notebooks will often become unresponsive and I see this from the server log: ``` ... WARNING:tornado.general:error on read Traceback (most recent call last): File "/gpfs22/home/burnsss1/env6/lib/python2.7/site-packages/tornado/iostream.py", line 391, in _handle_read if self._read_to_buffer() == 0: File "/gpfs22/home/burnsss1/env6/lib/python2.7/site-packages/tornado/iostream.py", line 447, in _read_to_buffer chunk = self.read_from_fd() File "/gpfs22/home/burnsss1/env6/lib/python2.7/site-packages/tornado/iostream.py", line 911, in read_from_fd chunk = self.socket.read(self.read_chunk_size) File "/usr/local/python2/latest/x86_64/gcc46/nonet/lib/python2.7/ssl.py", line 151, in read return self._sslobj.read(len) error: [Errno 110] Connection timed out ... ``` Is there a config in IPython I can set to help mitigate this? Would 2.0 help? 64-bit linux, python 2.7.2... pyzmq==13.1.0 Jinja2==2.7 ipython==1.1.0 tornado==3.1.1 Thanks for the insight! --Scott -------------- next part -------------- An HTML attachment was scrubbed... URL: From claresloggett at gmail.com Thu Mar 20 22:20:36 2014 From: claresloggett at gmail.com (Clare Sloggett) Date: Fri, 21 Mar 2014 13:20:36 +1100 Subject: [IPython-dev] IPython for Education In-Reply-To: References: Message-ID: Hi Doug, I'm still fairly new to IPython Notebook - is this idea (multiple different logins) just talking about something you'd like to see, or are you talking about technology which exists? I've been thinking about using IPython Notebook for teaching by having each student create a separate notebook, but I'm given pause by a) it's very easy for students to mess with one anothers' notebooks, either accidentally or on purpose b) all students would be based in the same working directory so they could easily corrupt one anothers' data files - this would be even more likely to happen by accident than (a). The best I've come up with so far is to create a set of notebook profiles with different passwords, and launch a server instance per student on different ports. If there is currently a way of handling multiple users in IPython Notebook I'd be interested! Of course the other option is to have students run their notebook on localhost on their own computer, but sometimes we're in a lab situation where I don't have a lot of control over the machines and this is difficult to set up. Servers are easier in that situation. Cheers, Clare On 21 March 2014 04:30, Doug Blank wrote: > On Thu, Mar 20, 2014 at 1:06 PM, Alessandro Gagliardi > wrote: > > I was thinking about that. But my understanding is that when multiple > people > > are logged into the same IPython server, this might cause problems for > the > > server. Particularly if they are using the same notebook, it seems like > the > > kernel could easily get into an inconsistent state. Were you imagining > that > > the students would each be using their own notebook but hosted on a > single > > server? Or are there ways to have multiple users working on the same > > notebook without it causing problems? > > I am imagining a Notebook Server where each user logs in with unique > ID/password and controls their own kernels/clusters (perhaps similar > to how Sage Cloud with IPython works). Students would be members of at > least one, maybe more, classes. The server would need to be able to > run (or farm out) at least one kernel per student. > > At that point, one can start to think about sharing, and interacting > with students as users. A homework submission could just be a menu > option. I guess I am thinking about a multi-user system, perhaps like > Drupal/Wordpress built on Tornado for managing users as > students/ta's/teachers/graders etc. > > -Doug > > > Date: Thu, 20 Mar 2014 12:17:49 -0400 > > From: Doug Blank > > Subject: Re: [IPython-dev] IPython for Education > > To: IPython developers list > > Message-ID: > > > > Content-Type: text/plain; charset=ISO-8859-1 > > > > Alessandro, > > > > I think you are right that there are many possibilities for using > > IPython in the classroom. I think many of these could be explored more > > easily if the students were logged in via an IPython server... at > > least that is where I will be concentrating my explorations this year. > > I could imagine: sharing, commenting, grading, hand-in methods, > > testing, quizzes, and clicker-style anonymous polling ("are we all at > > the same place? ready to move on?"), etc. It would be great to make a > > list of such feature ideas, and prioritize.... > > > > -Doug > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From damontallen at gmail.com Thu Mar 20 23:02:58 2014 From: damontallen at gmail.com (Damon Allen) Date: Thu, 20 Mar 2014 23:02:58 -0400 Subject: [IPython-dev] Embedded Videos in Slideshows In-Reply-To: References: Message-ID: Dami?n, I have one more question on the topic. I noticed if I have an cell at the top of the notebook that is set to be skipped in the presentation, it messes up the audio control's autoplay. Specifically the automatic pause of playback when I move on to the next slide. Has this been noticed elsewhere? I noticed that the top cell was getting switched to a markdown cell from a Heading 1, and so I added a cell at the top to act as a buffer at the top. Once I started the slideshow, the audio controls would start playing correctly but they would keep playing once I advanced to the next slide. However when I removed the skipped cell at the top the problem went away. As for the top cell switching from Heading 1 to markdown, I cannot rule out that the cause was that my mouse accidentally clicking on the drop down menu since I haven't reproduced it. Since I fixed the playback problem, I'm just curious. Thanks, Damon Damon T. Allen Ph.D. Adjunct Professor (352) 234-3266 damontallen at gmail.com 344 Rinker Hall College of Construction Management University of Florida On Tue, Mar 18, 2014 at 8:42 PM, Damon Allen wrote: > Thanks for answering my question so promptly. I'll tell my students to > use Chrome if it is an issue for them. > > > > Damon T. Allen Ph.D. > Adjunct Professor > (352) 234-3266 > damontallen at gmail.com > 344 Rinker Hall > College of Construction Management > University of Florida > > > On Tue, Mar 18, 2014 at 8:11 PM, Dami?n Avila wrote: > >> I ask to google... and I found I have already answered myself... ja ja >> >> https://github.com/damianavila/live_reveal/issues/4 >> >> not a good sign, I think ;-) >> >> >> 2014-03-18 19:20 GMT-03:00 Damon Allen : >> >> The notebook is at: >>> >>> >>> http://nbviewer.ipython.org/github/damontallen/Construction-Lectures/blob/master/Week%200%20-%20Greetings.ipynb >>> >>> The slideshow is at: >>> >>> >>> http://damontallen.github.io/Construction-Lectures/Week%200%20-%20Greetings.slides.html#/ >>> >>> And this problem also occurs when I use >>> >>> ipython nbconvert "Week 0 - Greetings.ipynb" --to slides --post serve >>> >>> Thanks for looking into it. >>> >>> >>> >>> Damon T. Allen Ph.D. >>> Adjunct Professor >>> (352) 234-3266 >>> damontallen at gmail.com >>> 344 Rinker Hall >>> College of Construction Management >>> University of Florida >>> >>> >>> On Tue, Mar 18, 2014 at 6:13 PM, Dami?n Avila wrote: >>> >>>> Have you seen this problem using the slideshow generated with >>>> IPython.nbconvert? >>>> It can be an issue with reveal.js itself... can you provided me with an >>>> example notebook to study the case? >>>> >>>> Thanks. >>>> >>>> >>>> >>>> >>>> 2014-03-18 18:56 GMT-03:00 Damon Allen : >>>> >>>>> Hi, >>>>> >>>>> Does running a slideshow force embedded YouTube videos to play in >>>>> flash instead of HTML5? I've found that in while running Firefox in Ubuntu >>>>> I cannot play YouTube videos in a slideshow. I can watch the video in >>>>> Firefox if it isn't in a slideshow and if I the slideshow runs in Chrome >>>>> everything works. >>>>> >>>>> This isn't a big deal since most of my students are running Windows >>>>> and they will be able to use up-to-date versions of the flashplayer, but I >>>>> am curious if there could be a fix to either reveal or nbconvert that would >>>>> correct this. >>>>> >>>>> As it stands, I need to switch my development from using Firefox to >>>>> Chrome. >>>>> >>>>> I'm still using IPython 1.2.1. >>>>> >>>>> Thanks, >>>>> >>>>> Damon >>>>> >>>>> _______________________________________________ >>>>> IPython-dev mailing list >>>>> IPython-dev at scipy.org >>>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>>>> >>>>> >>>> >>>> >>>> -- >>>> Dami?n Avila >>>> Scientific Python Developer >>>> Quantitative Finance Analyst >>>> Statistics, Biostatistics and Econometrics Consultant >>>> Biochemist >>>> >>>> _______________________________________________ >>>> IPython-dev mailing list >>>> IPython-dev at scipy.org >>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>>> >>>> >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >>> >> >> >> -- >> Dami?n Avila >> Scientific Python Developer >> Quantitative Finance Analyst >> Statistics, Biostatistics and Econometrics Consultant >> Biochemist >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Fri Mar 21 08:36:03 2014 From: doug.blank at gmail.com (Doug Blank) Date: Fri, 21 Mar 2014 08:36:03 -0400 Subject: [IPython-dev] Kernels and parallel support? In-Reply-To: References: Message-ID: On Thu, Mar 20, 2014 at 4:06 PM, Doug Blank wrote: > On Thu, Mar 20, 2014 at 2:19 PM, MinRK wrote: >> >> >> >> On Thu, Mar 20, 2014 at 10:58 AM, Doug Blank wrote: >>> >>> After reading: >>> >>> http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html >>> >>> and related pages, I am unsure what the relationship between parallel >>> processing and kernels are. Can external kernels be used in a parallel >>> environment? Or is this part of IPython python-kernel specific? Does >>> it use ZMQ messages? >> >> >> Depends what you mean by 'external'. You can start engines at any time on >> any machine, and they can join a cluster. Yes, IPython.parallel uses zmq. > > If I wanted to explore using a cluster of 3rd-party kernels, such as > IHaskell or ICalico, in a parallel environment, where would I begin? > Has anyone done this before? > > I see here: > > http://ipython.org/ipython-doc/dev/parallel/magics.html#engines-as-kernels > > that kernels connect to a controller to create an engine. If the > standard ipcontrollers can start up external kernels, I guess a place > to begin might be implementing what the %px magic does? Along with those general questions, a specific one: I would have thought that "ipcluster start --profile calico" would have used my c.KernelManager.kernel_cmd to start my kernels, but I can't get ipcluster to use it. Likewise, if I use the "Cluster" tab in the notebook to start a cluster, I also don't get my kernel. How to start a cluster of third-party kernels? -Doug > -Doug > >> -MinRK >> >>> >>> >>> Thanks for any additional data, >>> >>> -Doug >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> From j.davidgriffiths at gmail.com Fri Mar 21 09:34:12 2014 From: j.davidgriffiths at gmail.com (John Griffiths) Date: Fri, 21 Mar 2014 13:34:12 +0000 Subject: [IPython-dev] IPython for Education In-Reply-To: References: Message-ID: sagemath cloud has some very good infrastructure for these kind of class-level group project type ipy setups. And some example use cases. And now a google chrome extension in beta. Definitely worth checking out for teaching use. On 21 Mar 2014 02:21, "Clare Sloggett" wrote: > > Hi Doug, > > I'm still fairly new to IPython Notebook - is this idea (multiple > different logins) just talking about something you'd like to see, or are > you talking about technology which exists? > > I've been thinking about using IPython Notebook for teaching by having > each student create a separate notebook, but I'm given pause by > a) it's very easy for students to mess with one anothers' notebooks, > either accidentally or on purpose > b) all students would be based in the same working directory so they could > easily corrupt one anothers' data files - this would be even more likely to > happen by accident than (a). > > The best I've come up with so far is to create a set of notebook profiles > with different passwords, and launch a server instance per student on > different ports. If there is currently a way of handling multiple users in > IPython Notebook I'd be interested! > > Of course the other option is to have students run their notebook on > localhost on their own computer, but sometimes we're in a lab situation > where I don't have a lot of control over the machines and this is difficult > to set up. Servers are easier in that situation. > > Cheers, > Clare > > > On 21 March 2014 04:30, Doug Blank wrote: > >> On Thu, Mar 20, 2014 at 1:06 PM, Alessandro Gagliardi >> wrote: >> > I was thinking about that. But my understanding is that when multiple >> people >> > are logged into the same IPython server, this might cause problems for >> the >> > server. Particularly if they are using the same notebook, it seems like >> the >> > kernel could easily get into an inconsistent state. Were you imagining >> that >> > the students would each be using their own notebook but hosted on a >> single >> > server? Or are there ways to have multiple users working on the same >> > notebook without it causing problems? >> >> I am imagining a Notebook Server where each user logs in with unique >> ID/password and controls their own kernels/clusters (perhaps similar >> to how Sage Cloud with IPython works). Students would be members of at >> least one, maybe more, classes. The server would need to be able to >> run (or farm out) at least one kernel per student. >> >> At that point, one can start to think about sharing, and interacting >> with students as users. A homework submission could just be a menu >> option. I guess I am thinking about a multi-user system, perhaps like >> Drupal/Wordpress built on Tornado for managing users as >> students/ta's/teachers/graders etc. >> >> -Doug >> >> > Date: Thu, 20 Mar 2014 12:17:49 -0400 >> > From: Doug Blank >> > Subject: Re: [IPython-dev] IPython for Education >> > To: IPython developers list >> > Message-ID: >> > >> > Content-Type: text/plain; charset=ISO-8859-1 >> > >> > Alessandro, >> > >> > I think you are right that there are many possibilities for using >> > IPython in the classroom. I think many of these could be explored more >> > easily if the students were logged in via an IPython server... at >> > least that is where I will be concentrating my explorations this year. >> > I could imagine: sharing, commenting, grading, hand-in methods, >> > testing, quizzes, and clicker-style anonymous polling ("are we all at >> > the same place? ready to move on?"), etc. It would be great to make a >> > list of such feature ideas, and prioritize.... >> > >> > -Doug >> > >> > _______________________________________________ >> > IPython-dev mailing list >> > IPython-dev at scipy.org >> > http://mail.scipy.org/mailman/listinfo/ipython-dev >> > >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Fri Mar 21 09:42:53 2014 From: doug.blank at gmail.com (Doug Blank) Date: Fri, 21 Mar 2014 09:42:53 -0400 Subject: [IPython-dev] IPython for Education In-Reply-To: References: Message-ID: On Thu, Mar 20, 2014 at 10:20 PM, Clare Sloggett wrote: > > Hi Doug, > > I'm still fairly new to IPython Notebook - is this idea (multiple different > logins) just talking about something you'd like to see, or are you talking > about technology which exists? Doesn't exist yet, something I'd like to see, and something that I can help build. > I've been thinking about using IPython Notebook for teaching by having each > student create a separate notebook, but I'm given pause by > a) it's very easy for students to mess with one anothers' notebooks, either > accidentally or on purpose > b) all students would be based in the same working directory so they could > easily corrupt one anothers' data files - this would be even more likely to > happen by accident than (a). > > The best I've come up with so far is to create a set of notebook profiles > with different passwords, and launch a server instance per student on > different ports. If there is currently a way of handling multiple users in > IPython Notebook I'd be interested! That is probably the best that there is. But it wouldn't be too hard to imagine automating that, and allowing id + password via a server. > Of course the other option is to have students run their notebook on > localhost on their own computer, but sometimes we're in a lab situation > where I don't have a lot of control over the machines and this is difficult > to set up. Servers are easier in that situation. Yes, that will always be an option, but I want to make it really easy for beginning students to get started. I believe that the IPython team is planning to turn their energy to a server this summer. I hope to have something in place for this Fall. -Doug > Cheers, > Clare > > > On 21 March 2014 04:30, Doug Blank wrote: >> >> On Thu, Mar 20, 2014 at 1:06 PM, Alessandro Gagliardi >> wrote: >> > I was thinking about that. But my understanding is that when multiple >> > people >> > are logged into the same IPython server, this might cause problems for >> > the >> > server. Particularly if they are using the same notebook, it seems like >> > the >> > kernel could easily get into an inconsistent state. Were you imagining >> > that >> > the students would each be using their own notebook but hosted on a >> > single >> > server? Or are there ways to have multiple users working on the same >> > notebook without it causing problems? >> >> I am imagining a Notebook Server where each user logs in with unique >> ID/password and controls their own kernels/clusters (perhaps similar >> to how Sage Cloud with IPython works). Students would be members of at >> least one, maybe more, classes. The server would need to be able to >> run (or farm out) at least one kernel per student. >> >> At that point, one can start to think about sharing, and interacting >> with students as users. A homework submission could just be a menu >> option. I guess I am thinking about a multi-user system, perhaps like >> Drupal/Wordpress built on Tornado for managing users as >> students/ta's/teachers/graders etc. >> >> -Doug >> >> > Date: Thu, 20 Mar 2014 12:17:49 -0400 >> > From: Doug Blank >> > Subject: Re: [IPython-dev] IPython for Education >> > To: IPython developers list >> > Message-ID: >> > >> > Content-Type: text/plain; charset=ISO-8859-1 >> > >> > Alessandro, >> > >> > I think you are right that there are many possibilities for using >> > IPython in the classroom. I think many of these could be explored more >> > easily if the students were logged in via an IPython server... at >> > least that is where I will be concentrating my explorations this year. >> > I could imagine: sharing, commenting, grading, hand-in methods, >> > testing, quizzes, and clicker-style anonymous polling ("are we all at >> > the same place? ready to move on?"), etc. It would be great to make a >> > list of such feature ideas, and prioritize.... >> > >> > -Doug >> > >> > _______________________________________________ >> > IPython-dev mailing list >> > IPython-dev at scipy.org >> > http://mail.scipy.org/mailman/listinfo/ipython-dev >> > >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From ellisonbg at gmail.com Fri Mar 21 12:03:37 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Fri, 21 Mar 2014 09:03:37 -0700 Subject: [IPython-dev] IPython for Education In-Reply-To: References: Message-ID: We plan on (and have funding for) building a full multiuser notebook server later this year. One of the main usage cases for this server will be teaching. We will have more details as we get started with the work after 2.0 is released. Cheers, Brian On Fri, Mar 21, 2014 at 6:42 AM, Doug Blank wrote: > On Thu, Mar 20, 2014 at 10:20 PM, Clare Sloggett > wrote: >> >> Hi Doug, >> >> I'm still fairly new to IPython Notebook - is this idea (multiple different >> logins) just talking about something you'd like to see, or are you talking >> about technology which exists? > > Doesn't exist yet, something I'd like to see, and something that I can > help build. > >> I've been thinking about using IPython Notebook for teaching by having each >> student create a separate notebook, but I'm given pause by >> a) it's very easy for students to mess with one anothers' notebooks, either >> accidentally or on purpose >> b) all students would be based in the same working directory so they could >> easily corrupt one anothers' data files - this would be even more likely to >> happen by accident than (a). >> >> The best I've come up with so far is to create a set of notebook profiles >> with different passwords, and launch a server instance per student on >> different ports. If there is currently a way of handling multiple users in >> IPython Notebook I'd be interested! > > That is probably the best that there is. But it wouldn't be too hard > to imagine automating that, and allowing id + password via a server. > >> Of course the other option is to have students run their notebook on >> localhost on their own computer, but sometimes we're in a lab situation >> where I don't have a lot of control over the machines and this is difficult >> to set up. Servers are easier in that situation. > > Yes, that will always be an option, but I want to make it really easy > for beginning students to get started. > > I believe that the IPython team is planning to turn their energy to a > server this summer. I hope to have something in place for this Fall. > > -Doug > >> Cheers, >> Clare >> >> >> On 21 March 2014 04:30, Doug Blank wrote: >>> >>> On Thu, Mar 20, 2014 at 1:06 PM, Alessandro Gagliardi >>> wrote: >>> > I was thinking about that. But my understanding is that when multiple >>> > people >>> > are logged into the same IPython server, this might cause problems for >>> > the >>> > server. Particularly if they are using the same notebook, it seems like >>> > the >>> > kernel could easily get into an inconsistent state. Were you imagining >>> > that >>> > the students would each be using their own notebook but hosted on a >>> > single >>> > server? Or are there ways to have multiple users working on the same >>> > notebook without it causing problems? >>> >>> I am imagining a Notebook Server where each user logs in with unique >>> ID/password and controls their own kernels/clusters (perhaps similar >>> to how Sage Cloud with IPython works). Students would be members of at >>> least one, maybe more, classes. The server would need to be able to >>> run (or farm out) at least one kernel per student. >>> >>> At that point, one can start to think about sharing, and interacting >>> with students as users. A homework submission could just be a menu >>> option. I guess I am thinking about a multi-user system, perhaps like >>> Drupal/Wordpress built on Tornado for managing users as >>> students/ta's/teachers/graders etc. >>> >>> -Doug >>> >>> > Date: Thu, 20 Mar 2014 12:17:49 -0400 >>> > From: Doug Blank >>> > Subject: Re: [IPython-dev] IPython for Education >>> > To: IPython developers list >>> > Message-ID: >>> > >>> > Content-Type: text/plain; charset=ISO-8859-1 >>> > >>> > Alessandro, >>> > >>> > I think you are right that there are many possibilities for using >>> > IPython in the classroom. I think many of these could be explored more >>> > easily if the students were logged in via an IPython server... at >>> > least that is where I will be concentrating my explorations this year. >>> > I could imagine: sharing, commenting, grading, hand-in methods, >>> > testing, quizzes, and clicker-style anonymous polling ("are we all at >>> > the same place? ready to move on?"), etc. It would be great to make a >>> > list of such feature ideas, and prioritize.... >>> > >>> > -Doug >>> > >>> > _______________________________________________ >>> > IPython-dev mailing list >>> > IPython-dev at scipy.org >>> > http://mail.scipy.org/mailman/listinfo/ipython-dev >>> > >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From benjaminrk at gmail.com Fri Mar 21 12:10:43 2014 From: benjaminrk at gmail.com (Min RK) Date: Fri, 21 Mar 2014 09:10:43 -0700 Subject: [IPython-dev] Kernels and parallel support? In-Reply-To: References: Message-ID: <7F5E1466-D1A0-452F-8C1A-C48B054F7E61@gmail.com> > On Mar 21, 2014, at 5:36, Doug Blank wrote: > >> On Thu, Mar 20, 2014 at 4:06 PM, Doug Blank wrote: >>> On Thu, Mar 20, 2014 at 2:19 PM, MinRK wrote: >>> >>> >>> >>>> On Thu, Mar 20, 2014 at 10:58 AM, Doug Blank wrote: >>>> >>>> After reading: >>>> >>>> http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html >>>> >>>> and related pages, I am unsure what the relationship between parallel >>>> processing and kernels are. Can external kernels be used in a parallel >>>> environment? Or is this part of IPython python-kernel specific? Does >>>> it use ZMQ messages? >>> >>> >>> Depends what you mean by 'external'. You can start engines at any time on >>> any machine, and they can join a cluster. Yes, IPython.parallel uses zmq. >> >> If I wanted to explore using a cluster of 3rd-party kernels, such as >> IHaskell or ICalico, in a parallel environment, where would I begin? >> Has anyone done this before? >> >> I see here: >> >> http://ipython.org/ipython-doc/dev/parallel/magics.html#engines-as-kernels >> >> that kernels connect to a controller to create an engine. If the >> standard ipcontrollers can start up external kernels, I guess a place >> to begin might be implementing what the %px magic does? > > Along with those general questions, a specific one: > > I would have thought that "ipcluster start --profile calico" would > have used my c.KernelManager.kernel_cmd to start my kernels, but I > can't get ipcluster to use it. Likewise, if I use the "Cluster" tab in > the notebook to start a cluster, I also don't get my kernel. > > How to start a cluster of third-party kernels? It's not that simple. KernelManagers are not used in ipcluster. Normally, all engines are started manually (ipcluster is basically: for i in range(n): ipengine). If you want to turn your kernel into an engine, you are going to have to teach it about the ipcontroller connection files, and connecting instead of binding. > > -Doug > >> -Doug >> >>> -MinRK >>> >>>> >>>> >>>> Thanks for any additional data, >>>> >>>> -Doug >>>> _______________________________________________ >>>> IPython-dev mailing list >>>> IPython-dev at scipy.org >>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> >>> >>> >>> _______________________________________________ >>> IPython-dev mailing list >>> IPython-dev at scipy.org >>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>> > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From zvoros at gmail.com Fri Mar 21 13:00:22 2014 From: zvoros at gmail.com (=?ISO-8859-1?Q?Zolt=E1n_V=F6r=F6s?=) Date: Fri, 21 Mar 2014 18:00:22 +0100 Subject: [IPython-dev] generating input cells dynamically Message-ID: <532C7026.1080105@gmail.com> Hi all, I have a project, where I would have to plot hundreds of data sets, and for various reasons, I would like to place each plot in separate cell in the notebook. Is there a way to do this automatically? I know what the input cells should contain, it is only that I don't want to execute each cell by hand, for each plot takes 1-2 minutes to calculate. The only thing I could come up with for now is that I generate a notebook with the required cells on the command line, load the notebook in the browser, and then hit "Run All", but it would be great, if it was possible in a somewhat more elegant way. Any comments? Thanks, Zolt?n -------------- next part -------------- An HTML attachment was scrubbed... URL: From alessandro.gagliardi at glassdoor.com Fri Mar 21 13:40:28 2014 From: alessandro.gagliardi at glassdoor.com (Alessandro Gagliardi) Date: Fri, 21 Mar 2014 17:40:28 +0000 Subject: [IPython-dev] IPython-dev Digest, Vol 122, Issue 46 In-Reply-To: Message-ID: I?m personally of the opinion that, all else being equal, it?s best to have the students run IPython Notebook locally. It does take time, which may be perceived as ?wasted? by students, but getting your tools working on your computer is a prerequisite to doing anything, and better that students suffer through that with the help of teachers and colleagues to help them debug than on their own. That said, if you don?t want to deal with such issues, Wakari might work for you. The main thing I?m interested in is getting the notebooks to talk to each other (or at least, to the instructor?s and TAs?) which might mean hosting them all on one server. Here?s a thought: I wonder if it would be possible to use IPython.parallel to get a direct view into the student?s namespaces. We could tell the students to use certain variable names and then use the DirectView.gather to get the results and see how they compare. The objects could then be anything pickle-able. Date: Fri, 21 Mar 2014 13:20:36 +1100 From: Clare Sloggett > Subject: Re: [IPython-dev] IPython for Education To: IPython developers list > Message-ID: > Content-Type: text/plain; charset="iso-8859-1" Hi Doug, I'm still fairly new to IPython Notebook - is this idea (multiple different logins) just talking about something you'd like to see, or are you talking about technology which exists? I've been thinking about using IPython Notebook for teaching by having each student create a separate notebook, but I'm given pause by a) it's very easy for students to mess with one anothers' notebooks, either accidentally or on purpose b) all students would be based in the same working directory so they could easily corrupt one anothers' data files - this would be even more likely to happen by accident than (a). The best I've come up with so far is to create a set of notebook profiles with different passwords, and launch a server instance per student on different ports. If there is currently a way of handling multiple users in IPython Notebook I'd be interested! Of course the other option is to have students run their notebook on localhost on their own computer, but sometimes we're in a lab situation where I don't have a lot of control over the machines and this is difficult to set up. Servers are easier in that situation. Cheers, Clare -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Fri Mar 21 13:56:30 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Fri, 21 Mar 2014 10:56:30 -0700 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: <532C7026.1080105@gmail.com> References: <532C7026.1080105@gmail.com> Message-ID: Hi Zolt?n, On 21 March 2014 10:00, Zolt?n V?r?s wrote: > I have a project, where I would have to plot hundreds of data sets, and > for various reasons, I would like to place each plot in separate cell in > the notebook. Is there a way to do this automatically? I know what the > input cells should contain, it is only that I don't want to execute each > cell by hand, for each plot takes 1-2 minutes to calculate. The only thing > I could come up with for now is that I generate a notebook with the > required cells on the command line, load the notebook in the browser, and > then hit "Run All", but it would be great, if it was possible in a somewhat > more elegant way. Once you've generated the notebook, you could use runipy to run it without having to go through the noteboko UI: https://github.com/paulgb/runipy/ Best wishes, Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From jon.freder at gmail.com Fri Mar 21 14:12:36 2014 From: jon.freder at gmail.com (Jonathan Frederic) Date: Fri, 21 Mar 2014 11:12:36 -0700 Subject: [IPython-dev] Widget View post render hook In-Reply-To: <2109409B-430E-4622-BF32-F1E2D1AB1423@stanford.edu> References: <76E452D7-1098-4E0D-8DC6-B5E831E6DAD2@stanford.edu> <2109409B-430E-4622-BF32-F1E2D1AB1423@stanford.edu> Message-ID: Hi Nik, I added a `displayed` event to the widget model in https://github.com/ipython/ipython/pull/5404. We will have to wait to see if the others like it. Cheers, Jon On Sun, Mar 16, 2014 at 8:00 PM, Nikolas Tezak wrote: > Hi Jon, > > that is pretty much what I ended up using (took me a while cause I'm a > total noob at JavaScript). > My circuit widget is coming together nicely. I've had a couple other ideas > that I'll probably throw together as soon as I get around to them. One > example (shamelessly copied from Mathematica's "Manipulate[]" command) > > 1) extending the interactive command by adding a button that prints out a > function call with the currently set parameters. > 2) like 1) but perhaps an option to save the parameter setting to disk? > (and then reload them when re-opening the notebook at a later time). I > don't like this second option very much because it would be nice to make > any notebook self-contained. > > Potentially one could also insert a new input cell with such a function > call / stored parameter settings such that they would be available from > within python even after restarting the kernel. > But I don't know if that kind of dynamic code inserting clashes with your > guys' vision of what should be doable. > > Anyhow, thanks for your response! And again, my compliments for an awesome > new feature in an amazing tool. > > Nik > > > > > On Mar 15, 2014, at 1:03 PM, Jonathan Frederic > wrote: > > > Hi Nik, > > > > I'm glad you like the widgets and thank you for the feedback. I > encountered the same problem when implementing a D3.js widget (see > https://github.com/jdfreder/ipython-d3). As a hack-ish workaround I used > a timeout with a 0ms interval ( > https://github.com/jdfreder/ipython-d3/blob/master/widget_forcedirectedgraph.js#L13). > Right now I don't believe we have a nice way to do this, adding an event > or method is probably a good idea. > > > > -Jon > > > > > > On Fri, Mar 14, 2014 at 11:27 AM, Nikolas Tezak > wrote: > > Hi everybody, > > > > first of all, I _love_ the widgets, they are extremely useful!! And I'm > super impressed by IPython's crazy fast development progress. > > I am currently trying to create a custom widget type that embeds a > jsplumb circuit editor > > ( http://jsplumbtoolkit.com/demo/home/jquery.html ) > > > > However, when I implement my custom "render" method for the view I am > running into a problem. > > I first dynamically create the circuit components (currently just > absolutely positioned divs within the widget view's this.$el div. This > works fine. > > I then initialize a jsplumb instance and pass this.$el as the default > container. > > Now, however, when I try to add the connectors and connections with > jsplumb within the render method, it fails because jsplumb is trying to > read out dynamically the position of the widget view's div before that div > has been inserted into the DOM. > > > > So basically, my problem is that I want code within "render()" that > already requires the view to be inserted into the document. > > Should I use a different approach, i.e., maybe put this initialization > code into "update()", or do you expect that this will be a common enough > use case that it would justify adding a special post render hook that gets > called after the view's div has been attached to a cell's widget sub_area? > > > > > > I hope I described this well enough, let me know if you'd like me to > share a code example. > > Thanks, > > > > Nik > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From aaron.oleary at gmail.com Fri Mar 21 15:13:23 2014 From: aaron.oleary at gmail.com (Aaron O'Leary) Date: Fri, 21 Mar 2014 19:13:23 +0000 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: References: <532C7026.1080105@gmail.com> Message-ID: <20140321191323.GA20739@tk422.wireless.leeds.ac.uk> > Once you've generated the notebook, you could use runipy to run it without > having to go through the noteboko UI: https://github.com/paulgb/runipy/ Or, if you're using IPython 2.0 you can do ipython -c "%run your_notebook.ipynb" from the command line. This is the same as opening it and clicking "run all cells". aaron From zvoros at gmail.com Fri Mar 21 15:21:38 2014 From: zvoros at gmail.com (=?ISO-8859-1?Q?Zolt=E1n_V=F6r=F6s?=) Date: Fri, 21 Mar 2014 20:21:38 +0100 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: References: <532C7026.1080105@gmail.com> Message-ID: <532C9142.1040108@gmail.com> Hi Thomas, On 21/03/14 18:56, Thomas Kluyver wrote: > Hi Zolt?n, > > On 21 March 2014 10:00, Zolt?n V?r?s > wrote: > > I have a project, where I would have to plot hundreds of data > sets, and for various reasons, I would like to place each plot in > separate cell in the notebook. Is there a way to do this > automatically? I know what the input cells should contain, it is > only that I don't want to execute each cell by hand, for each plot > takes 1-2 minutes to calculate. The only thing I could come up > with for now is that I generate a notebook with the required cells > on the command line, load the notebook in the browser, and then > hit "Run All", but it would be great, if it was possible in a > somewhat more elegant way. > > > Once you've generated the notebook, you could use runipy to run it > without having to go through the noteboko UI: > https://github.com/paulgb/runipy/ > Thanks a lot! This is a very interesting project, and very new, too. Indeed, had I asked the question yesterday, you wouldn't have been able to point to this repository:) But I definitely foresee myself using this a lot. In the meantime, I have been trying to generate the code in the notebook itself. I believe, this (or something very similar) should work: %%javascript for(var i=0; i < 3; i++) { IPython.notebook.insert_cell_below('code') var cell = IPython.notebook.get_selected_cell() cell.set_text('plot(' + i + '*sin(x))') cell.select() } Now, the four cells are created (it should be 3, but what the heck!), and inserted in the notebook, yet, only 'plot(2*sin(x))' is displayed, and only in the first cell. Is this the intended behaviour? Cheers, Zolt?n -------------- next part -------------- An HTML attachment was scrubbed... URL: From zvoros at gmail.com Fri Mar 21 15:30:11 2014 From: zvoros at gmail.com (=?ISO-8859-1?Q?Zolt=E1n_V=F6r=F6s?=) Date: Fri, 21 Mar 2014 20:30:11 +0100 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: <20140321191323.GA20739@tk422.wireless.leeds.ac.uk> References: <532C7026.1080105@gmail.com> <20140321191323.GA20739@tk422.wireless.leeds.ac.uk> Message-ID: <532C9343.3030501@gmail.com> Hi Araron, Thanks for the suggestion! But the main problem was not how to run the notebook, but how to generate the code in the first place. Basically, I would like to have a notebook that writes itself. Once it's written, it can be run in many ways, as you pointed out. Cheers, Zolt?n On 21/03/14 20:13, Aaron O'Leary wrote: >> Once you've generated the notebook, you could use runipy to run it without >> having to go through the noteboko UI: https://github.com/paulgb/runipy/ > Or, if you're using IPython 2.0 you can do > > ipython -c "%run your_notebook.ipynb" > > from the command line. This is the same as opening it and clicking "run > all cells". > > aaron > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Fri Mar 21 15:30:25 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Fri, 21 Mar 2014 12:30:25 -0700 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: <532C9142.1040108@gmail.com> References: <532C7026.1080105@gmail.com> <532C9142.1040108@gmail.com> Message-ID: On 21 March 2014 12:21, Zolt?n V?r?s wrote: > Thanks a lot! This is a very interesting project, and very new, too. > Indeed, had I asked the question yesterday, you wouldn't have been able to > point to this repository:) But I definitely foresee myself using this a > lot. It's not quite *that* new - the history goes back a few months: https://github.com/paulgb/runipy/commits/master -------------- next part -------------- An HTML attachment was scrubbed... URL: From fperez.net at gmail.com Fri Mar 21 16:03:04 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Fri, 21 Mar 2014 13:03:04 -0700 Subject: [IPython-dev] IPython for Education In-Reply-To: References: Message-ID: Hi all, for those interested, the notes from our dev meeting when we worked out our plans for the next few months are, as everything else in ipython, publicly available: https://hackpad.com/IPython-Winter-2014-Development-Meeting-fKrExqKCWmC Getting 2.0 fleshed out and released has taken a little longer than we'd anticipated, but we're almost there and will shift energy into this. As usual, our in-depth discussions take place during our public weekly dev meetings, which folks can follow on youtube. And once code begins to materialized, we'll continue doing everything as always, with code review in public on github and other discussions here + weekly meetings. Cheers f On Fri, Mar 21, 2014 at 9:03 AM, Brian Granger wrote: > We plan on (and have funding for) building a full multiuser notebook > server later this year. One of the main usage cases for this server > will be teaching. We will have more details as we get started with the > work after 2.0 is released. > > Cheers, > > Brian > > On Fri, Mar 21, 2014 at 6:42 AM, Doug Blank wrote: > > On Thu, Mar 20, 2014 at 10:20 PM, Clare Sloggett > > wrote: > >> > >> Hi Doug, > >> > >> I'm still fairly new to IPython Notebook - is this idea (multiple > different > >> logins) just talking about something you'd like to see, or are you > talking > >> about technology which exists? > > > > Doesn't exist yet, something I'd like to see, and something that I can > > help build. > > > >> I've been thinking about using IPython Notebook for teaching by having > each > >> student create a separate notebook, but I'm given pause by > >> a) it's very easy for students to mess with one anothers' notebooks, > either > >> accidentally or on purpose > >> b) all students would be based in the same working directory so they > could > >> easily corrupt one anothers' data files - this would be even more > likely to > >> happen by accident than (a). > >> > >> The best I've come up with so far is to create a set of notebook > profiles > >> with different passwords, and launch a server instance per student on > >> different ports. If there is currently a way of handling multiple users > in > >> IPython Notebook I'd be interested! > > > > That is probably the best that there is. But it wouldn't be too hard > > to imagine automating that, and allowing id + password via a server. > > > >> Of course the other option is to have students run their notebook on > >> localhost on their own computer, but sometimes we're in a lab situation > >> where I don't have a lot of control over the machines and this is > difficult > >> to set up. Servers are easier in that situation. > > > > Yes, that will always be an option, but I want to make it really easy > > for beginning students to get started. > > > > I believe that the IPython team is planning to turn their energy to a > > server this summer. I hope to have something in place for this Fall. > > > > -Doug > > > >> Cheers, > >> Clare > >> > >> > >> On 21 March 2014 04:30, Doug Blank wrote: > >>> > >>> On Thu, Mar 20, 2014 at 1:06 PM, Alessandro Gagliardi > >>> wrote: > >>> > I was thinking about that. But my understanding is that when multiple > >>> > people > >>> > are logged into the same IPython server, this might cause problems > for > >>> > the > >>> > server. Particularly if they are using the same notebook, it seems > like > >>> > the > >>> > kernel could easily get into an inconsistent state. Were you > imagining > >>> > that > >>> > the students would each be using their own notebook but hosted on a > >>> > single > >>> > server? Or are there ways to have multiple users working on the same > >>> > notebook without it causing problems? > >>> > >>> I am imagining a Notebook Server where each user logs in with unique > >>> ID/password and controls their own kernels/clusters (perhaps similar > >>> to how Sage Cloud with IPython works). Students would be members of at > >>> least one, maybe more, classes. The server would need to be able to > >>> run (or farm out) at least one kernel per student. > >>> > >>> At that point, one can start to think about sharing, and interacting > >>> with students as users. A homework submission could just be a menu > >>> option. I guess I am thinking about a multi-user system, perhaps like > >>> Drupal/Wordpress built on Tornado for managing users as > >>> students/ta's/teachers/graders etc. > >>> > >>> -Doug > >>> > >>> > Date: Thu, 20 Mar 2014 12:17:49 -0400 > >>> > From: Doug Blank > >>> > Subject: Re: [IPython-dev] IPython for Education > >>> > To: IPython developers list > >>> > Message-ID: > >>> > > >>> > Content-Type: text/plain; charset=ISO-8859-1 > >>> > > >>> > Alessandro, > >>> > > >>> > I think you are right that there are many possibilities for using > >>> > IPython in the classroom. I think many of these could be explored > more > >>> > easily if the students were logged in via an IPython server... at > >>> > least that is where I will be concentrating my explorations this > year. > >>> > I could imagine: sharing, commenting, grading, hand-in methods, > >>> > testing, quizzes, and clicker-style anonymous polling ("are we all at > >>> > the same place? ready to move on?"), etc. It would be great to make a > >>> > list of such feature ideas, and prioritize.... > >>> > > >>> > -Doug > >>> > > >>> > _______________________________________________ > >>> > IPython-dev mailing list > >>> > IPython-dev at scipy.org > >>> > http://mail.scipy.org/mailman/listinfo/ipython-dev > >>> > > >>> _______________________________________________ > >>> IPython-dev mailing list > >>> IPython-dev at scipy.org > >>> http://mail.scipy.org/mailman/listinfo/ipython-dev > >> > >> > >> > >> _______________________________________________ > >> IPython-dev mailing list > >> IPython-dev at scipy.org > >> http://mail.scipy.org/mailman/listinfo/ipython-dev > >> > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > -- > Brian E. Granger > Cal Poly State University, San Luis Obispo > bgranger at calpoly.edu and ellisonbg at gmail.com > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From pmhobson at gmail.com Fri Mar 21 19:28:42 2014 From: pmhobson at gmail.com (Paul Hobson) Date: Fri, 21 Mar 2014 16:28:42 -0700 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: <532C9343.3030501@gmail.com> References: <532C7026.1080105@gmail.com> <20140321191323.GA20739@tk422.wireless.leeds.ac.uk> <532C9343.3030501@gmail.com> Message-ID: At this point, I have to wonder if it makes more sense to just write a small utility to build a LaTeX document for yourself. -paul On Fri, Mar 21, 2014 at 12:30 PM, Zolt?n V?r?s wrote: > Hi Araron, > > Thanks for the suggestion! But the main problem was not how to run the > notebook, but how to generate the code in the first place. Basically, I > would like to have a notebook that writes itself. Once it's written, it can > be run in many ways, as you pointed out. > > Cheers, > Zolt?n > > On 21/03/14 20:13, Aaron O'Leary wrote: > > Once you've generated the notebook, you could use runipy to run it without > having to go through the noteboko UI: https://github.com/paulgb/runipy/ > > Or, if you're using IPython 2.0 you can do > > ipython -c "%run your_notebook.ipynb" > > from the command line. This is the same as opening it and clicking "run > all cells". > > aaron > _______________________________________________ > IPython-dev mailing listIPython-dev at scipy.orghttp://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Fri Mar 21 19:44:51 2014 From: benjaminrk at gmail.com (MinRK) Date: Fri, 21 Mar 2014 16:44:51 -0700 Subject: [IPython-dev] Output from background threads in notebook? In-Reply-To: References: Message-ID: On Sun, Mar 16, 2014 at 6:30 PM, Doug Blank wrote: > If you run the following cell: > > import threading > import time > class ThreadClass(threading.Thread): > def run(self): > x = 1 > while True: > print(x) > x += 1 > time.sleep(1) > t = ThreadClass() > t.start() > t.join() > > Then you will see the data for as long as you don't interrupt the > cell. If you comment out the t.join(), then you won't see any output > (or maybe just the first printed). > > In the next cell, if you do something such as, say: > > while True: > pass > > and then interrupt the kernel, you'll see some more of the printouts. > > It appears that when the execution_reply is received, the notebook > doesn't take output messages anymore. Is this all by design? Is there > a manner to make sure background-running output can still make its way > to the notebook? > zmq sockets are not threadsafe. To protect against segfaults, background threads are not allowed to write directly to the IOPub socket. What they do instead is write to a buffer, which the main thread checks periodically. That's why async output from a thread doesn't show up until there is some interaction in the main thread. It would be easy enough to hook up that checking / flushing to the main eventloop, in which case it would still be threadsafe, but still allow publication while the kernel is idle. But that's now how it works right now. -MinRK > > -Doug > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From zvoros at gmail.com Sat Mar 22 04:11:26 2014 From: zvoros at gmail.com (=?ISO-8859-1?Q?Zolt=E1n_V=F6r=F6s?=) Date: Sat, 22 Mar 2014 09:11:26 +0100 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: References: <532C7026.1080105@gmail.com> <20140321191323.GA20739@tk422.wireless.leeds.ac.uk> <532C9343.3030501@gmail.com> Message-ID: <532D45AE.7000505@gmail.com> But LaTeX has nothing to do with the problem. The question was, given an empty notebook, how does one *generate* the following: In [1]: plot(sin(1*x)) In [2]: plot(sin(2*x)) In [3]: plot(sin(3*x)) . . . In [200]: plot(sin(200*x)) The 200 plots have to be in separate cells, so that their order can be changed afterwards, e.g. I want to have an IPython notebook as we understand it, but I would like to generate its content (executable code) dynamically. Metaprogramming, if you wish. Zolt?n On 22/03/14 00:28, Paul Hobson wrote: > At this point, I have to wonder if it makes more sense to just write a > small utility to build a LaTeX document for yourself. > -paul > > > On Fri, Mar 21, 2014 at 12:30 PM, Zolt?n V?r?s > wrote: > > Hi Araron, > > Thanks for the suggestion! But the main problem was not how to run > the notebook, but how to generate the code in the first place. > Basically, I would like to have a notebook that writes itself. > Once it's written, it can be run in many ways, as you pointed out. > > Cheers, > Zolt?n > > On 21/03/14 20:13, Aaron O'Leary wrote: >>> Once you've generated the notebook, you could use runipy to run it without >>> having to go through the noteboko UI:https://github.com/paulgb/runipy/ >> Or, if you're using IPython 2.0 you can do >> >> ipython -c "%run your_notebook.ipynb" >> >> from the command line. This is the same as opening it and clicking "run >> all cells". >> >> aaron >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Sat Mar 22 09:47:55 2014 From: doug.blank at gmail.com (Doug Blank) Date: Sat, 22 Mar 2014 09:47:55 -0400 Subject: [IPython-dev] Kernels and parallel support? In-Reply-To: <7F5E1466-D1A0-452F-8C1A-C48B054F7E61@gmail.com> References: <7F5E1466-D1A0-452F-8C1A-C48B054F7E61@gmail.com> Message-ID: On Fri, Mar 21, 2014 at 12:10 PM, Min RK wrote: > >> On Mar 21, 2014, at 5:36, Doug Blank wrote: >> >>> On Thu, Mar 20, 2014 at 4:06 PM, Doug Blank wrote: >>>> On Thu, Mar 20, 2014 at 2:19 PM, MinRK wrote: >>>> >>>> >>>> >>>>> On Thu, Mar 20, 2014 at 10:58 AM, Doug Blank wrote: >>>>> >>>>> After reading: >>>>> >>>>> http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html >>>>> >>>>> and related pages, I am unsure what the relationship between parallel >>>>> processing and kernels are. Can external kernels be used in a parallel >>>>> environment? Or is this part of IPython python-kernel specific? Does >>>>> it use ZMQ messages? >>>> >>>> >>>> Depends what you mean by 'external'. You can start engines at any time on >>>> any machine, and they can join a cluster. Yes, IPython.parallel uses zmq. >>> >>> If I wanted to explore using a cluster of 3rd-party kernels, such as >>> IHaskell or ICalico, in a parallel environment, where would I begin? >>> Has anyone done this before? >>> >>> I see here: >>> >>> http://ipython.org/ipython-doc/dev/parallel/magics.html#engines-as-kernels >>> >>> that kernels connect to a controller to create an engine. If the >>> standard ipcontrollers can start up external kernels, I guess a place >>> to begin might be implementing what the %px magic does? >> >> Along with those general questions, a specific one: >> >> I would have thought that "ipcluster start --profile calico" would >> have used my c.KernelManager.kernel_cmd to start my kernels, but I >> can't get ipcluster to use it. Likewise, if I use the "Cluster" tab in >> the notebook to start a cluster, I also don't get my kernel. >> >> How to start a cluster of third-party kernels? > > It's not that simple. KernelManagers are not used in ipcluster. Normally, all engines are started manually (ipcluster is basically: for i in range(n): ipengine). If you want to turn your kernel into an engine, you are going to have to teach it about the ipcontroller connection files, and connecting instead of binding. First, the idea of allowing non-Python kernels to be able to easily become parallelized is a fantastic possibility! There are many languages/systems that could benefit from this. I have gone through the anatomy of the communication between a hub, engines, and a notebook here: http://wiki.roboteducation.org/IPython_Parallel_API (corrections of misunderstandings welcome). Two questions: 1) In addition to adding the ipcontroller communications needed, it appears that pickling and eval are used to execute remote code (even though pack/unpack settings in ipcontroller_*.py files are marked as "json"). Of course, this won't work for non-Python kernels. Could a real json representation be used for values, getting rid of the need for pickle? 2) Knowing that "Full, integrated support for non-Python kernels" is a line item in the 3.0 roadmap [1], and this is mentioned in IPython/parallel/engine/engine.py: # FIXME: This is a hack until IPKernelApp and IPEngineApp can be fully merged app = IPKernelApp(parent=self, shell=self.kernel.shell, kernel=self.kernel, log=self.log) is there a chance that Kernels and Engines will be merged this years so that parallel processing for non-Python kernels would be possible? Thanks for any additional information! -Doug [1] - https://hackpad.com/IPython-Winter-2014-Development-Meeting-fKrExqKCWmC > >> >> -Doug >> >>> -Doug >>> >>>> -MinRK >>>> >>>>> >>>>> >>>>> Thanks for any additional data, >>>>> >>>>> -Doug >>>>> _______________________________________________ >>>>> IPython-dev mailing list >>>>> IPython-dev at scipy.org >>>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>>> >>>> >>>> >>>> _______________________________________________ >>>> IPython-dev mailing list >>>> IPython-dev at scipy.org >>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >>>> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From doug.blank at gmail.com Sat Mar 22 10:01:58 2014 From: doug.blank at gmail.com (Doug Blank) Date: Sat, 22 Mar 2014 10:01:58 -0400 Subject: [IPython-dev] Output from background threads in notebook? In-Reply-To: References: Message-ID: On Fri, Mar 21, 2014 at 7:44 PM, MinRK wrote: > > > > On Sun, Mar 16, 2014 at 6:30 PM, Doug Blank wrote: >> >> If you run the following cell: >> >> import threading >> import time >> class ThreadClass(threading.Thread): >> def run(self): >> x = 1 >> while True: >> print(x) >> x += 1 >> time.sleep(1) >> t = ThreadClass() >> t.start() >> t.join() >> >> Then you will see the data for as long as you don't interrupt the >> cell. If you comment out the t.join(), then you won't see any output >> (or maybe just the first printed). >> >> In the next cell, if you do something such as, say: >> >> while True: >> pass >> >> and then interrupt the kernel, you'll see some more of the printouts. >> >> It appears that when the execution_reply is received, the notebook >> doesn't take output messages anymore. Is this all by design? Is there >> a manner to make sure background-running output can still make its way >> to the notebook? > > > zmq sockets are not threadsafe. To protect against segfaults, background > threads are not allowed to write directly to the IOPub socket. What they do > instead is write to a buffer, which the main thread checks periodically. > That's why async output from a thread doesn't show up until there is some > interaction in the main thread. > > It would be easy enough to hook up that checking / flushing to the main > eventloop, in which case it would still be threadsafe, but still allow > publication while the kernel is idle. But that's now how it works right now. Ok, thanks for the feedback. I've opened issues #5408 to request this issue: https://github.com/ipython/ipython/issues/5408 -Doug > -MinRK > >> >> >> -Doug >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From benjaminrk at gmail.com Sat Mar 22 13:59:50 2014 From: benjaminrk at gmail.com (MinRK) Date: Sat, 22 Mar 2014 10:59:50 -0700 Subject: [IPython-dev] Kernels and parallel support? In-Reply-To: References: <7F5E1466-D1A0-452F-8C1A-C48B054F7E61@gmail.com> Message-ID: On Sat, Mar 22, 2014 at 6:47 AM, Doug Blank wrote: > On Fri, Mar 21, 2014 at 12:10 PM, Min RK wrote: > > > >> On Mar 21, 2014, at 5:36, Doug Blank wrote: > >> > >>> On Thu, Mar 20, 2014 at 4:06 PM, Doug Blank > wrote: > >>>> On Thu, Mar 20, 2014 at 2:19 PM, MinRK wrote: > >>>> > >>>> > >>>> > >>>>> On Thu, Mar 20, 2014 at 10:58 AM, Doug Blank > wrote: > >>>>> > >>>>> After reading: > >>>>> > >>>>> http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html > >>>>> > >>>>> and related pages, I am unsure what the relationship between parallel > >>>>> processing and kernels are. Can external kernels be used in a > parallel > >>>>> environment? Or is this part of IPython python-kernel specific? Does > >>>>> it use ZMQ messages? > >>>> > >>>> > >>>> Depends what you mean by 'external'. You can start engines at any > time on > >>>> any machine, and they can join a cluster. Yes, IPython.parallel uses > zmq. > >>> > >>> If I wanted to explore using a cluster of 3rd-party kernels, such as > >>> IHaskell or ICalico, in a parallel environment, where would I begin? > >>> Has anyone done this before? > >>> > >>> I see here: > >>> > >>> > http://ipython.org/ipython-doc/dev/parallel/magics.html#engines-as-kernels > >>> > >>> that kernels connect to a controller to create an engine. If the > >>> standard ipcontrollers can start up external kernels, I guess a place > >>> to begin might be implementing what the %px magic does? > >> > >> Along with those general questions, a specific one: > >> > >> I would have thought that "ipcluster start --profile calico" would > >> have used my c.KernelManager.kernel_cmd to start my kernels, but I > >> can't get ipcluster to use it. Likewise, if I use the "Cluster" tab in > >> the notebook to start a cluster, I also don't get my kernel. > >> > >> How to start a cluster of third-party kernels? > > > > It's not that simple. KernelManagers are not used in ipcluster. > Normally, all engines are started manually (ipcluster is basically: for i > in range(n): ipengine). If you want to turn your kernel into an engine, > you are going to have to teach it about the ipcontroller connection files, > and connecting instead of binding. > > First, the idea of allowing non-Python kernels to be able to easily > become parallelized is a fantastic possibility! There are many > languages/systems that could benefit from this. > > I have gone through the anatomy of the communication between a hub, > engines, and a notebook here: > > http://wiki.roboteducation.org/IPython_Parallel_API > > (corrections of misunderstandings welcome). Two questions: > > 1) In addition to adding the ipcontroller communications needed, it > appears that pickling and eval are used to execute remote code (even > though pack/unpack settings in ipcontroller_*.py files are marked as > "json"). Of course, this won't work for non-Python kernels. Could a > real json representation be used for values, getting rid of the need > for pickle? > JSON *is* used for message serialization. However, the main message used in IPython.parallel is `apply`, which sends native functions and data, and thus uses native serialization (pickle on Python). There are no plans to change this. Code-as-text execution (%px, view.execute), however, uses the same execute_request message that is used for execution in the notebook, etc., and should work fine on a non-Python kernel. > > 2) Knowing that "Full, integrated support for non-Python kernels" is a > line item in the 3.0 roadmap [1], and this is mentioned in > IPython/parallel/engine/engine.py: > > # FIXME: This is a hack until IPKernelApp and IPEngineApp can be fully > merged > app = IPKernelApp(parent=self, shell=self.kernel.shell, > kernel=self.kernel, log=self.log) > > is there a chance that Kernels and Engines will be merged this year > so that parallel processing for non-Python kernels would be possible? > Yes, this is my hope. > > Thanks for any additional information! > > -Doug > > [1] - > https://hackpad.com/IPython-Winter-2014-Development-Meeting-fKrExqKCWmC > > > > >> > >> -Doug > >> > >>> -Doug > >>> > >>>> -MinRK > >>>> > >>>>> > >>>>> > >>>>> Thanks for any additional data, > >>>>> > >>>>> -Doug > >>>>> _______________________________________________ > >>>>> IPython-dev mailing list > >>>>> IPython-dev at scipy.org > >>>>> http://mail.scipy.org/mailman/listinfo/ipython-dev > >>>> > >>>> > >>>> > >>>> _______________________________________________ > >>>> IPython-dev mailing list > >>>> IPython-dev at scipy.org > >>>> http://mail.scipy.org/mailman/listinfo/ipython-dev > >>>> > >> _______________________________________________ > >> IPython-dev mailing list > >> IPython-dev at scipy.org > >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From doug.blank at gmail.com Sat Mar 22 14:30:37 2014 From: doug.blank at gmail.com (Doug Blank) Date: Sat, 22 Mar 2014 14:30:37 -0400 Subject: [IPython-dev] Kernels and parallel support? In-Reply-To: References: <7F5E1466-D1A0-452F-8C1A-C48B054F7E61@gmail.com> Message-ID: On Sat, Mar 22, 2014 at 1:59 PM, MinRK wrote: > On Sat, Mar 22, 2014 at 6:47 AM, Doug Blank wrote: >> >> On Fri, Mar 21, 2014 at 12:10 PM, Min RK wrote: >> > >> >> On Mar 21, 2014, at 5:36, Doug Blank wrote: >> >> >> >>> On Thu, Mar 20, 2014 at 4:06 PM, Doug Blank >> >>> wrote: >> >>>> On Thu, Mar 20, 2014 at 2:19 PM, MinRK wrote: >> >>>> >> >>>> >> >>>> >> >>>>> On Thu, Mar 20, 2014 at 10:58 AM, Doug Blank >> >>>>> wrote: >> >>>>> >> >>>>> After reading: >> >>>>> >> >>>>> http://ipython.org/ipython-doc/dev/parallel/parallel_intro.html >> >>>>> >> >>>>> and related pages, I am unsure what the relationship between >> >>>>> parallel >> >>>>> processing and kernels are. Can external kernels be used in a >> >>>>> parallel >> >>>>> environment? Or is this part of IPython python-kernel specific? Does >> >>>>> it use ZMQ messages? >> >>>> >> >>>> >> >>>> Depends what you mean by 'external'. You can start engines at any >> >>>> time on >> >>>> any machine, and they can join a cluster. Yes, IPython.parallel uses >> >>>> zmq. >> >>> >> >>> If I wanted to explore using a cluster of 3rd-party kernels, such as >> >>> IHaskell or ICalico, in a parallel environment, where would I begin? >> >>> Has anyone done this before? >> >>> >> >>> I see here: >> >>> >> >>> >> >>> http://ipython.org/ipython-doc/dev/parallel/magics.html#engines-as-kernels >> >>> >> >>> that kernels connect to a controller to create an engine. If the >> >>> standard ipcontrollers can start up external kernels, I guess a place >> >>> to begin might be implementing what the %px magic does? >> >> >> >> Along with those general questions, a specific one: >> >> >> >> I would have thought that "ipcluster start --profile calico" would >> >> have used my c.KernelManager.kernel_cmd to start my kernels, but I >> >> can't get ipcluster to use it. Likewise, if I use the "Cluster" tab in >> >> the notebook to start a cluster, I also don't get my kernel. >> >> >> >> How to start a cluster of third-party kernels? >> > >> > It's not that simple. KernelManagers are not used in ipcluster. >> > Normally, all engines are started manually (ipcluster is basically: for i in >> > range(n): ipengine). If you want to turn your kernel into an engine, you >> > are going to have to teach it about the ipcontroller connection files, and >> > connecting instead of binding. >> >> First, the idea of allowing non-Python kernels to be able to easily >> become parallelized is a fantastic possibility! There are many >> languages/systems that could benefit from this. >> >> I have gone through the anatomy of the communication between a hub, >> engines, and a notebook here: >> >> http://wiki.roboteducation.org/IPython_Parallel_API >> >> (corrections of misunderstandings welcome). Two questions: >> >> 1) In addition to adding the ipcontroller communications needed, it >> appears that pickling and eval are used to execute remote code (even >> though pack/unpack settings in ipcontroller_*.py files are marked as >> "json"). Of course, this won't work for non-Python kernels. Could a >> real json representation be used for values, getting rid of the need >> for pickle? > > > JSON *is* used for message serialization. However, the main message used in > IPython.parallel is `apply`, which sends native functions and data, and thus > uses native serialization (pickle on Python). There are no plans to change > this. Code-as-text execution (%px, view.execute), however, uses the same > execute_request message that is used for execution in the notebook, etc., > and should work fine on a non-Python kernel. Ok, no need to change anything as long as there is a method to evaluate code-as-text. Sounds perfect! >> >> >> 2) Knowing that "Full, integrated support for non-Python kernels" is a >> line item in the 3.0 roadmap [1], and this is mentioned in >> IPython/parallel/engine/engine.py: >> >> # FIXME: This is a hack until IPKernelApp and IPEngineApp can be fully >> merged >> app = IPKernelApp(parent=self, shell=self.kernel.shell, >> kernel=self.kernel, log=self.log) >> >> is there a chance that Kernels and Engines will be merged this year >> so that parallel processing for non-Python kernels would be possible? > > > Yes, this is my hope. Excellent! As you know, I've started issue #5409 to track this request. Thanks again! -Doug https://github.com/ipython/ipython/issues/5409 >> >> >> Thanks for any additional information! >> >> -Doug >> >> [1] - >> https://hackpad.com/IPython-Winter-2014-Development-Meeting-fKrExqKCWmC >> >> > >> >> >> >> -Doug >> >> >> >>> -Doug >> >>> >> >>>> -MinRK >> >>>> >> >>>>> >> >>>>> >> >>>>> Thanks for any additional data, >> >>>>> >> >>>>> -Doug >> >>>>> _______________________________________________ >> >>>>> IPython-dev mailing list >> >>>>> IPython-dev at scipy.org >> >>>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >>>> >> >>>> >> >>>> >> >>>> _______________________________________________ >> >>>> IPython-dev mailing list >> >>>> IPython-dev at scipy.org >> >>>> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >>>> >> >> _______________________________________________ >> >> IPython-dev mailing list >> >> IPython-dev at scipy.org >> >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> > _______________________________________________ >> > IPython-dev mailing list >> > IPython-dev at scipy.org >> > http://mail.scipy.org/mailman/listinfo/ipython-dev >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > From fperez.net at gmail.com Sat Mar 22 19:56:40 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Sat, 22 Mar 2014 16:56:40 -0700 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: <532D45AE.7000505@gmail.com> References: <532C7026.1080105@gmail.com> <20140321191323.GA20739@tk422.wireless.leeds.ac.uk> <532C9343.3030501@gmail.com> <532D45AE.7000505@gmail.com> Message-ID: Here: http://nbviewer.ipython.org/gist/fperez/9716279 Cheers, f On Sat, Mar 22, 2014 at 1:11 AM, Zolt?n V?r?s wrote: > But LaTeX has nothing to do with the problem. The question was, given an > empty notebook, how does one *generate* the following: > > In [1]: plot(sin(1*x)) > > In [2]: plot(sin(2*x)) > > In [3]: plot(sin(3*x)) > . > . > . > > In [200]: plot(sin(200*x)) > > The 200 plots have to be in separate cells, so that their order can be > changed afterwards, e.g. I want to have an IPython notebook as we > understand it, but I would like to generate its content (executable code) > dynamically. Metaprogramming, if you wish. > > Zolt?n > > > On 22/03/14 00:28, Paul Hobson wrote: > > At this point, I have to wonder if it makes more sense to just write a > small utility to build a LaTeX document for yourself. > -paul > > > On Fri, Mar 21, 2014 at 12:30 PM, Zolt?n V?r?s wrote: > >> Hi Araron, >> >> Thanks for the suggestion! But the main problem was not how to run the >> notebook, but how to generate the code in the first place. Basically, I >> would like to have a notebook that writes itself. Once it's written, it can >> be run in many ways, as you pointed out. >> >> Cheers, >> Zolt?n >> >> On 21/03/14 20:13, Aaron O'Leary wrote: >> >> Once you've generated the notebook, you could use runipy to run it without >> having to go through the noteboko UI: https://github.com/paulgb/runipy/ >> >> Or, if you're using IPython 2.0 you can do >> >> ipython -c "%run your_notebook.ipynb" >> >> from the command line. This is the same as opening it and clicking "run >> all cells". >> >> aaron >> _______________________________________________ >> IPython-dev mailing listIPython-dev at scipy.orghttp://mail.scipy.org/mailman/listinfo/ipython-dev >> >> >> >> _______________________________________________ >> IPython-dev mailing list >> IPython-dev at scipy.org >> http://mail.scipy.org/mailman/listinfo/ipython-dev >> >> > > > _______________________________________________ > IPython-dev mailing listIPython-dev at scipy.orghttp://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From patrick.surry at gmail.com Sat Mar 22 20:10:10 2014 From: patrick.surry at gmail.com (Patrick Surry) Date: Sat, 22 Mar 2014 20:10:10 -0400 Subject: [IPython-dev] generating input cells dynamically Message-ID: That's cool! Looks like you are creating a completely new notebook in that case. I'm following this question with interest since it seems closely related to what I was trying to do in "unrolling a loop" over a sequence of cells, so I can do (almost) the same thing a bunch of times in an interactive notebook and still go back and poke around. The suggestion there (below) was to use javascript API to execute cells. Is there also a way to insert/modify cells into the current notebook via the javascript API? (Is the API documented somewhere, I haven't been able to find, although some good hints here http://www.peterbouda.eu/ipython-and-javascript-interaction.html ?) That would probably solve both of our problems? Cheers, Patrick From: Kiko Subject: Re: [IPython-dev] Loop over a set of notebook cells? To: IPython developers list Message-ID: Content-Type: text/plain; charset="iso-8859-1" 2014-03-17 19:44 GMT+01:00 MinRK : > You can execute a cell range with a little javascript (e.g. in a > %%javascript cell): > > var start = 2; > var stop = 4; > > for (var i = start; i < stop; i++) { > var cell = IPython.notebook.get_cell(i); > cell.execute(); > } > > Maybe this: %%javascript IPython.notebook.execute_cell_range(x,y) with x,y the range of the cells. From: Fernando Perez > Subject: Re: [IPython-dev] generating input cells dynamically > To: IPython developers list > Message-ID: > w9fB8OK5SHvUdYWO0SZB7A at mail.gmail.com> > Content-Type: text/plain; charset="iso-8859-1" > > Here: > > http://nbviewer.ipython.org/gist/fperez/9716279 > > Cheers, > > f > > > On Sat, Mar 22, 2014 at 1:11 AM, Zolt?n V?r?s wrote: > > > But LaTeX has nothing to do with the problem. The question was, given an > > empty notebook, how does one *generate* the following: > > > > In [1]: plot(sin(1*x)) > > > > In [2]: plot(sin(2*x)) > > > > In [3]: plot(sin(3*x)) > > . > > . > > . > > > > In [200]: plot(sin(200*x)) > > > > The 200 plots have to be in separate cells, so that their order can be > > changed afterwards, e.g. I want to have an IPython notebook as we > > understand it, but I would like to generate its content (executable code) > > dynamically. Metaprogramming, if you wish. > > > > Zolt?n > > > > > > On 22/03/14 00:28, Paul Hobson wrote: > > > > At this point, I have to wonder if it makes more sense to just write a > > small utility to build a LaTeX document for yourself. > > -paul > > > > > > On Fri, Mar 21, 2014 at 12:30 PM, Zolt?n V?r?s wrote: > > > >> Hi Araron, > >> > >> Thanks for the suggestion! But the main problem was not how to run the > >> notebook, but how to generate the code in the first place. Basically, I > >> would like to have a notebook that writes itself. Once it's written, it > can > >> be run in many ways, as you pointed out. > >> > >> Cheers, > >> Zolt?n > >> > >> On 21/03/14 20:13, Aaron O'Leary wrote: > >> > >> Once you've generated the notebook, you could use runipy to run it > without > >> having to go through the noteboko UI: https://github.com/paulgb/runipy/ > >> > >> Or, if you're using IPython 2.0 you can do > >> > >> ipython -c "%run your_notebook.ipynb" > >> > >> from the command line. This is the same as opening it and clicking "run > >> all cells". > >> > >> aaron > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From zvoros at gmail.com Sun Mar 23 03:59:51 2014 From: zvoros at gmail.com (=?ISO-8859-1?Q?Zolt=E1n_V=F6r=F6s?=) Date: Sun, 23 Mar 2014 08:59:51 +0100 Subject: [IPython-dev] generating input cells dynamically In-Reply-To: References: <532C7026.1080105@gmail.com> <20140321191323.GA20739@tk422.wireless.leeds.ac.uk> <532C9343.3030501@gmail.com> <532D45AE.7000505@gmail.com> Message-ID: <532E9477.70505@gmail.com> Hi Fernando, Many thanks, this is absolutely cool! I haven't known about the interface for reading/writing notebook files. Up to now, I have been reading/writing notebooks via simplejson, but that is not too safe, I guess. Cheers, Zolt?n On 23/03/14 00:56, Fernando Perez wrote: > Here: > > http://nbviewer.ipython.org/gist/fperez/9716279 > > Cheers, > > f > > > On Sat, Mar 22, 2014 at 1:11 AM, Zolt?n V?r?s > wrote: > > But LaTeX has nothing to do with the problem. The question was, > given an empty notebook, how does one *generate* the following: > > In [1]: plot(sin(1*x)) > > In [2]: plot(sin(2*x)) > > In [3]: plot(sin(3*x)) > . > . > . > > In [200]: plot(sin(200*x)) > > The 200 plots have to be in separate cells, so that their order > can be changed afterwards, e.g. I want to have an IPython notebook > as we understand it, but I would like to generate its content > (executable code) dynamically. Metaprogramming, if you wish. > > Zolt?n > -------------- next part -------------- An HTML attachment was scrubbed... URL: From bussonniermatthias at gmail.com Sun Mar 23 13:26:36 2014 From: bussonniermatthias at gmail.com (Matthias BUSSONNIER) Date: Sun, 23 Mar 2014 18:26:36 +0100 Subject: [IPython-dev] IPyXMPP, a chat bot serving an IPython shell over XMPP In-Reply-To: References: Message-ID: <3A5CE663-E67B-4833-9966-C46D24318356@gmail.com> Le 18 mars 2014 ? 21:47, Francesco Rossi a ?crit : > Hello all, > > I am writing a tool for interacting in real time with the data being > processed in some simulation codes (running for days). Of course, I > find that the best fit for doing that is embedding IPython in the C > simulation codes and use its extraordinary features for data analysis. > > To provide the easiest possible access (without the need to know which > cluster the simulation is running on or to set up ssh tunnels) to > simple commands, I have written a small chat bot that serves the > (embedded) IPython shell over XMPP. > > The chat bot is here: https://github.com/redsh/IPyXMPP . That's Cool. Have you considered adding the project to https://github.com/ipython/ipython/wiki/Projects-using-IPython And tag it with IPython framework when releasing on Pypi ! > > Has any of you already worked on something similar? > Do you think it could be useful in other fields rather than > long-simulation monitoring? That would be cool to have a plugin for a chat bot ! > Which of the existing IPython shell apps would you suggest me to > follow as a track for implementing a multi-user/multi-kernel shell > over XMPP? We will get to multi-user/multi-kernel for the notebook (multi kernel in the sense you will be apple to restart with a different kernel). -- M From ntezak at stanford.edu Sun Mar 23 23:51:08 2014 From: ntezak at stanford.edu (Nikolas Tezak) Date: Sun, 23 Mar 2014 20:51:08 -0700 Subject: [IPython-dev] Widget View post render hook In-Reply-To: References: <76E452D7-1098-4E0D-8DC6-B5E831E6DAD2@stanford.edu> <2109409B-430E-4622-BF32-F1E2D1AB1423@stanford.edu> Message-ID: <12FC8A76-5E18-40C1-86F8-FD29096B10BE@stanford.edu> Hey Jon, awesome, thanks! Nik On Mar 21, 2014, at 11:12 AM, Jonathan Frederic wrote: > Hi Nik, > > I added a `displayed` event to the widget model in https://github.com/ipython/ipython/pull/5404. We will have to wait to see if the others like it. > > Cheers, > Jon > > > On Sun, Mar 16, 2014 at 8:00 PM, Nikolas Tezak wrote: > Hi Jon, > > that is pretty much what I ended up using (took me a while cause I?m a total noob at JavaScript). > My circuit widget is coming together nicely. I?ve had a couple other ideas that I?ll probably throw together as soon as I get around to them. One example (shamelessly copied from Mathematica?s ?Manipulate[]? command) > > 1) extending the interactive command by adding a button that prints out a function call with the currently set parameters. > 2) like 1) but perhaps an option to save the parameter setting to disk? (and then reload them when re-opening the notebook at a later time). I don?t like this second option very much because it would be nice to make any notebook self-contained. > > Potentially one could also insert a new input cell with such a function call / stored parameter settings such that they would be available from within python even after restarting the kernel. > But I don?t know if that kind of dynamic code inserting clashes with your guys? vision of what should be doable. > > Anyhow, thanks for your response! And again, my compliments for an awesome new feature in an amazing tool. > > Nik > > > > > On Mar 15, 2014, at 1:03 PM, Jonathan Frederic wrote: > > > Hi Nik, > > > > I'm glad you like the widgets and thank you for the feedback. I encountered the same problem when implementing a D3.js widget (see https://github.com/jdfreder/ipython-d3). As a hack-ish workaround I used a timeout with a 0ms interval (https://github.com/jdfreder/ipython-d3/blob/master/widget_forcedirectedgraph.js#L13). Right now I don't believe we have a nice way to do this, adding an event or method is probably a good idea. > > > > -Jon > > > > > > On Fri, Mar 14, 2014 at 11:27 AM, Nikolas Tezak wrote: > > Hi everybody, > > > > first of all, I _love_ the widgets, they are extremely useful!! And I?m super impressed by IPython?s crazy fast development progress. > > I am currently trying to create a custom widget type that embeds a jsplumb circuit editor > > ( http://jsplumbtoolkit.com/demo/home/jquery.html ) > > > > However, when I implement my custom ?render? method for the view I am running into a problem. > > I first dynamically create the circuit components (currently just absolutely positioned divs within the widget view?s this.$el div. This works fine. > > I then initialize a jsplumb instance and pass this.$el as the default container. > > Now, however, when I try to add the connectors and connections with jsplumb within the render method, it fails because jsplumb is trying to read out dynamically the position of the widget view?s div before that div has been inserted into the DOM. > > > > So basically, my problem is that I want code within ?render()? that already requires the view to be inserted into the document. > > Should I use a different approach, i.e., maybe put this initialization code into "update()?, or do you expect that this will be a common enough use case that it would justify adding a special post render hook that gets called after the view?s div has been attached to a cell's widget sub_area? > > > > > > I hope I described this well enough, let me know if you?d like me to share a code example. > > Thanks, > > > > Nik > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From benjaminrk at gmail.com Tue Mar 25 23:31:26 2014 From: benjaminrk at gmail.com (MinRK) Date: Tue, 25 Mar 2014 20:31:26 -0700 Subject: [IPython-dev] [ANNOUNCE] IPython 2.0.0 release candidate Message-ID: The first release candidate for IPython 2.0.0 is now available: http://archive.ipython.org/testing/2.0.0 Take it for a spin, and let us know if you find new issues. We still have some documentation updates to do, but barring anything unexpected, we should have our first IPython 2.0 release in a week or so. You can install the release candidate with pip: pip install --pre --find-links http://archive.ipython.org/testing/2.0.0ipython -------------- next part -------------- An HTML attachment was scrubbed... URL: From david at wolever.net Wed Mar 26 00:55:03 2014 From: david at wolever.net (David Wolever) Date: Wed, 26 Mar 2014 00:55:03 -0400 Subject: [IPython-dev] [ANNOUNCE] IPython 2.0.0 release candidate In-Reply-To: References: Message-ID: <2D255EDB-AC05-411E-A28B-45963D3D8C1B@wolever.net> Awesome! The only issue I've run into, which Min helped me diagnose: My PYTHONSTARTUP file sets up tab completion as per http://docs.python.org/2/tutorial/interactive.html , which breaks IPython's tab completion. I'm not entirely sure off the top of my head how this could be sniffed + warned, but it was an annoying issue. On 2014-03-25, at 11:31 PM, MinRK wrote: > The first release candidate for IPython 2.0.0 is now available: http://archive.ipython.org/testing/2.0.0 > > Take it for a spin, and let us know if you find new issues. We still have some documentation updates to do, but barring anything unexpected, we should have our first IPython 2.0 release in a week or so. > > You can install the release candidate with pip: > > pip install --pre --find-links http://archive.ipython.org/testing/2.0.0 ipython > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev -- phone: (416) 906-0403 pgp: B230230D -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 203 bytes Desc: Message signed with OpenPGP using GPGMail URL: From j.davidgriffiths at gmail.com Wed Mar 26 04:51:25 2014 From: j.davidgriffiths at gmail.com (John Griffiths) Date: Wed, 26 Mar 2014 08:51:25 +0000 Subject: [IPython-dev] pivot table widgets? Message-ID: Does anyone know of any notebook implementations of pivot table type things that can 'filter' e.g. a pandas dataframe based on button press or text input? Feel like I've seen some stuff like this before. Ideally this would not use the kernel and would be usable on e.g. nbviewer. (If this doesn't yet exist I think it would be a very worthy of an ipython widget treatment) Ta, john Mr. John Griffiths, MSc PhD Candidate Centre for Speech, Language, and the Brain Department of Experimental Psychology University of Cambridge, UK -------------- next part -------------- An HTML attachment was scrubbed... URL: From kikocorreoso at gmail.com Wed Mar 26 15:11:12 2014 From: kikocorreoso at gmail.com (Kiko) Date: Wed, 26 Mar 2014 20:11:12 +0100 Subject: [IPython-dev] [ANNOUNCE] IPython 2.0.0 release candidate In-Reply-To: <2D255EDB-AC05-411E-A28B-45963D3D8C1B@wolever.net> References: <2D255EDB-AC05-411E-A28B-45963D3D8C1B@wolever.net> Message-ID: > > The first release candidate for IPython 2.0.0 is now available: > http://archive.ipython.org/testing/2.0.0 > > Take it for a spin, and let us know if you find new issues. We still have > some documentation updates to do, but barring anything unexpected, we > should have our first IPython 2.0 release in a week or so. > > You can install the release candidate with pip: > > pip install --pre --find-links > http://archive.ipython.org/testing/2.0.0 ipython > > Hi all I've found an issue using OpenLayers. I tested that on versions: 1.2.1, 2.0.0dev 2.0.0rc1 The tiles are misaligned on versions 2.0.0xxx You can test that using the following code in a cell: %%html
Best regards and thanks for the hard work. -------------- next part -------------- An HTML attachment was scrubbed... URL: From damontallen at gmail.com Wed Mar 26 18:57:37 2014 From: damontallen at gmail.com (Damon Allen) Date: Wed, 26 Mar 2014 18:57:37 -0400 Subject: [IPython-dev] User interface Message-ID: Is this the interface of version 2.0? http://i.imgur.com/ExVBycn.png I don't see anything about printing or really any other changes mentioned in the docs except the version number. I've cleared my cache and that did not affect anything. Rebooting didn't have an effect either. What do I need to do to display / access the updates? Thanks, Damon -------------- next part -------------- An HTML attachment was scrubbed... URL: From ellisonbg at gmail.com Wed Mar 26 19:06:08 2014 From: ellisonbg at gmail.com (Brian Granger) Date: Wed, 26 Mar 2014 16:06:08 -0700 Subject: [IPython-dev] User interface In-Reply-To: References: Message-ID: Nope, that is the UI for 1.x. On Wed, Mar 26, 2014 at 3:57 PM, Damon Allen wrote: > Is this the interface of version 2.0? > > http://i.imgur.com/ExVBycn.png > > I don't see anything about printing or really any other changes mentioned in > the docs except the version number. > > I've cleared my cache and that did not affect anything. Rebooting didn't > have an effect either. What do I need to do to display / access the > updates? > > Thanks, > > Damon > > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -- Brian E. Granger Cal Poly State University, San Luis Obispo bgranger at calpoly.edu and ellisonbg at gmail.com From damontallen at gmail.com Thu Mar 27 06:33:38 2014 From: damontallen at gmail.com (Damon Allen) Date: Thu, 27 Mar 2014 06:33:38 -0400 Subject: [IPython-dev] User interface In-Reply-To: References: Message-ID: Got it! The interface looks great! I really like the tour option. Thanks for all your hard work. Damon T. Allen Ph.D. Adjunct Professor (352) 234-3266 damontallen at gmail.com 344 Rinker Hall College of Construction Management University of Florida On Wed, Mar 26, 2014 at 7:06 PM, Brian Granger wrote: > Nope, that is the UI for 1.x. > > On Wed, Mar 26, 2014 at 3:57 PM, Damon Allen > wrote: > > Is this the interface of version 2.0? > > > > http://i.imgur.com/ExVBycn.png > > > > I don't see anything about printing or really any other changes > mentioned in > > the docs except the version number. > > > > I've cleared my cache and that did not affect anything. Rebooting didn't > > have an effect either. What do I need to do to display / access the > > updates? > > > > Thanks, > > > > Damon > > > > > > _______________________________________________ > > IPython-dev mailing list > > IPython-dev at scipy.org > > http://mail.scipy.org/mailman/listinfo/ipython-dev > > > > > > -- > Brian E. Granger > Cal Poly State University, San Luis Obispo > bgranger at calpoly.edu and ellisonbg at gmail.com > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From qclists at inventati.org Thu Mar 27 14:12:29 2014 From: qclists at inventati.org (Robert Cross) Date: Thu, 27 Mar 2014 14:12:29 -0400 Subject: [IPython-dev] Proof of concept HTML5 notification extension Message-ID: <53346A0D.8080005@inventati.org> Hi guys, I made a little proof of concept extension for IPython to implement HTML5 desktop notifications in Chrome and Firefox. I'm not an experienced IPython dev nor do I know much about the proper way of doing certain things, so feel free to criticize. Here's a link to the file on Gist: https://gist.github.com/robcross/9776646 I'll just paste the important bits of my comment block: > Notifyme! - By Robert Cross (rcross at inventati.org) > Tested on IPython 0.13.1 and current master branch as of date above. > === > -I got tired of listening to my laptop's fan speed to determine when my > 20 minute calculations were done, so I created this simple extension. > -This is my first ever open source contribution so feel free to fix, add, > and improve upon my concept. I would love to see this built into IPython > Notebook itself! > -I left a few things incomplete such as directory > sanitation and file not found errors in changing the notification sound, > as I'm not sure of the "proper" way of doing it > -If you decide to modify and improve this extension, please send me an > e-mail letting me know what improvements you are making and why (fo r > my own learning purposes). Thank you! > === > Wishlist > -Proper directory sanitization > -Proper file not found error handling > -Detection of browser notification capabilities on load > -Detection of notification permissions on load, and display button if > no t given -------------- next part -------------- An HTML attachment was scrubbed... URL: From hathaway56 at myfairpoint.net Thu Mar 27 19:01:28 2014 From: hathaway56 at myfairpoint.net (Allen Hathaway) Date: Thu, 27 Mar 2014 19:01:28 -0400 Subject: [IPython-dev] IPython Notebook input boxes Message-ID: <000001cf4a10$7dd1d730$79758590$@net> I installed IPython using Anaconda. I was running IPython Notebook on Python 2.7.6 using Firefox as the browser. I had updated all the relevant packages before starting. Everything ran beautiful for about 6 days. Then yesterday I came home from work and started it up. The input boxes appeared about 3 times taller than usual and would not accept any input. In fact, every keystroke and mouse-click made them get another line taller. I tried deleting the browsing history - no effect. I tried using other browsers - it won't talk to MS Internet Explorer, but Google Chrome behaved exactly like Firefox. Has anyone seen this behavior before? Is there some corrupted cache or file that can be fixed? Thanks for your help. Allen -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Thu Mar 27 19:27:52 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Thu, 27 Mar 2014 16:27:52 -0700 Subject: [IPython-dev] IPython Notebook input boxes In-Reply-To: <000001cf4a10$7dd1d730$79758590$@net> References: <000001cf4a10$7dd1d730$79758590$@net> Message-ID: On 27 March 2014 16:01, Allen Hathaway wrote: > I installed IPython using Anaconda. I was running IPython Notebook on > Python 2.7.6 using Firefox as the browser. I had updated all the relevant > packages before starting. > > > > Everything ran beautiful for about 6 days. Then yesterday I came home > from work and started it up. The input boxes appeared about 3 times taller > than usual and would not accept any input. In fact, every keystroke and > mouse-click made them get another line taller. > > > > I tried deleting the browsing history - no effect. I tried using other > browsers - it won't talk to MS Internet Explorer, but Google Chrome behaved > exactly like Firefox. > > > > Has anyone seen this behavior before? Is there some corrupted cache or > file that can be fixed? > I don't think we've seen that or heard about it before now. It's especially weird that you see the same in different browsers, so it can't be something with the cache or a broken browser update. The only thing I can think of to try is removing IPython and installing it again. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From ntezak at stanford.edu Thu Mar 27 19:29:49 2014 From: ntezak at stanford.edu (Nikolas Tezak) Date: Thu, 27 Mar 2014 16:29:49 -0700 Subject: [IPython-dev] IPython Notebook input boxes In-Reply-To: References: <000001cf4a10$7dd1d730$79758590$@net> Message-ID: <44CB7068-60C1-453D-BE52-CD1351DEC336@stanford.edu> Or maybe you could try just removing the ipython configuration directory? On Mar 27, 2014, at 4:27 PM, Thomas Kluyver wrote: > On 27 March 2014 16:01, Allen Hathaway wrote: > I installed IPython using Anaconda. I was running IPython Notebook on Python 2.7.6 using Firefox as the browser. I had updated all the relevant packages before starting. > > > > Everything ran beautiful for about 6 days. Then yesterday I came home from work and started it up. The input boxes appeared about 3 times taller than usual and would not accept any input. In fact, every keystroke and mouse-click made them get another line taller. > > > > I tried deleting the browsing history ? no effect. I tried using other browsers ? it won?t talk to MS Internet Explorer, but Google Chrome behaved exactly like Firefox. > > > > Has anyone seen this behavior before? Is there some corrupted cache or file that can be fixed? > > > I don't think we've seen that or heard about it before now. It's especially weird that you see the same in different browsers, so it can't be something with the cache or a broken browser update. The only thing I can think of to try is removing IPython and installing it again. > > Thomas > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev From hathaway56 at myfairpoint.net Thu Mar 27 21:45:02 2014 From: hathaway56 at myfairpoint.net (Allen Hathaway) Date: Thu, 27 Mar 2014 21:45:02 -0400 Subject: [IPython-dev] IPython Notebook input boxes In-Reply-To: <44CB7068-60C1-453D-BE52-CD1351DEC336@stanford.edu> References: <000001cf4a10$7dd1d730$79758590$@net> <44CB7068-60C1-453D-BE52-CD1351DEC336@stanford.edu> Message-ID: <000001cf4a27$570f6380$052e2a80$@net> I did the full un-install and re-install, and it seems to have worked for the moment. I don't think it is a very robust solution long term. I don't want to do this every week or so. I don't want to be locked into not installing updates periodically. Any better answers out there? Allen -----Original Message----- From: ipython-dev-bounces at scipy.org [mailto:ipython-dev-bounces at scipy.org] On Behalf Of Nikolas Tezak Sent: Thursday, March 27, 2014 7:30 PM To: IPython developers list Subject: Re: [IPython-dev] IPython Notebook input boxes Or maybe you could try just removing the ipython configuration directory? On Mar 27, 2014, at 4:27 PM, Thomas Kluyver wrote: > On 27 March 2014 16:01, Allen Hathaway wrote: > I installed IPython using Anaconda. I was running IPython Notebook on Python 2.7.6 using Firefox as the browser. I had updated all the relevant packages before starting. > > > > Everything ran beautiful for about 6 days. Then yesterday I came home from work and started it up. The input boxes appeared about 3 times taller than usual and would not accept any input. In fact, every keystroke and mouse-click made them get another line taller. > > > > I tried deleting the browsing history - no effect. I tried using other browsers - it won't talk to MS Internet Explorer, but Google Chrome behaved exactly like Firefox. > > > > Has anyone seen this behavior before? Is there some corrupted cache or file that can be fixed? > > > I don't think we've seen that or heard about it before now. It's especially weird that you see the same in different browsers, so it can't be something with the cache or a broken browser update. The only thing I can think of to try is removing IPython and installing it again. > > Thomas > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev _______________________________________________ IPython-dev mailing list IPython-dev at scipy.org http://mail.scipy.org/mailman/listinfo/ipython-dev From takowl at gmail.com Thu Mar 27 22:06:16 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Thu, 27 Mar 2014 19:06:16 -0700 Subject: [IPython-dev] IPython Notebook input boxes In-Reply-To: <000001cf4a27$570f6380$052e2a80$@net> References: <000001cf4a10$7dd1d730$79758590$@net> <44CB7068-60C1-453D-BE52-CD1351DEC336@stanford.edu> <000001cf4a27$570f6380$052e2a80$@net> Message-ID: On 27 March 2014 18:45, Allen Hathaway wrote: > I did the full un-install and re-install, and it seems to have worked for > the moment. I don't think it is a very robust solution long term. I don't > want to do this every week or so. I don't want to be locked into not > installing updates periodically. > I didn't mean it as a real solution, but we haven't seen the problem before and it's not obvious what could be causing it, so there's not much else I can suggest. It may have been some random one off. If you see the problem again, we can try to work out what the pattern is. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From vasco+python at tenner.nl Fri Mar 28 07:03:01 2014 From: vasco+python at tenner.nl (Vasco Tenner) Date: Fri, 28 Mar 2014 12:03:01 +0100 Subject: [IPython-dev] changing plotting backend Message-ID: <533556E5.9060903@tenner.nl> Hi, when I open a new ipython shell (master), withouth any flags, I can set the plotting backend by: > %matplotlib wx Or > %matplotlib gtk However, I cannot change it halfway my session: > %matplotlib wx > %matplotlib gtk Warning: Cannot change to a different GUI toolkit: qt. Using wx instead. If I start the ipython notebook server, I can only use tk. In the first cell of a new notebook: %matplotlib wx Warning: Cannot change to a different GUI toolkit: wx. Using tk instead. Is there a way to change the plotting backend in an ipython notebook to something else then tk or inline? Kind regards, Vasco From damontallen at gmail.com Fri Mar 28 10:43:46 2014 From: damontallen at gmail.com (Damon Allen) Date: Fri, 28 Mar 2014 10:43:46 -0400 Subject: [IPython-dev] Tab in markdown cells Message-ID: Hello, Great work, yet again! I have a question about using Tab to indent in the markdown cells of Notebooks using IPython 2.0rc. Tab still preforms the same in code cells, but in markdown cells, all it does is change the focus of the browser. I was wondering, since lists use indenting to differentiate levels, and code blocks are denoted with indenting in markdown, is there a keyboard shortcut that will indent in markdown cells? I've tried Tab and Ctrl-] to do indenting, but neither seems to work in markdown cells. I can add four spaces to indent, but I was hoping that there was still a single keystroke available. Thanks again for all the hard work that has gone into version 2.0, Damon -------------- next part -------------- An HTML attachment was scrubbed... URL: From kathleen.m.tacina at nasa.gov Fri Mar 28 11:03:08 2014 From: kathleen.m.tacina at nasa.gov (Kathleen Tacina) Date: Fri, 28 Mar 2014 11:03:08 -0400 Subject: [IPython-dev] IPython 2: disable automatically-added closing parentheses, brackets, and quotes Message-ID: <53358F2C.5050609@nasa.gov> Hi, I've been using IPython notebook at lot in my work. It's a great tool, and I like the edit/command modes adding in version 2. Unfortunately, though, the automatically-added closing parentheses, brackets, and quotes are making it harder for me to edit code. (The visual noise messes up my internal count of what I need to add.) How would I turn this feature off? Thanks for your help with this and for making a great product! Best regards, Kathleen Tacina From takowl at gmail.com Fri Mar 28 11:18:29 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Fri, 28 Mar 2014 08:18:29 -0700 Subject: [IPython-dev] changing plotting backend In-Reply-To: <533556E5.9060903@tenner.nl> References: <533556E5.9060903@tenner.nl> Message-ID: Hi Vasco, That sounds like there's something in your notebook config file which is initialising matplotlib. Are any values set in that file? Thomas On 28 Mar 2014 04:03, "Vasco Tenner" wrote: > Hi, > when I open a new ipython shell (master), withouth any flags, I can set > the plotting backend by: > > %matplotlib wx > Or > > %matplotlib gtk > > However, I cannot change it halfway my session: > > %matplotlib wx > > > %matplotlib gtk > Warning: Cannot change to a different GUI toolkit: qt. Using wx instead. > > If I start the ipython notebook server, I can only use tk. In the first > cell of a new notebook: > > %matplotlib wx > Warning: Cannot change to a different GUI toolkit: wx. Using tk instead. > > Is there a way to change the plotting backend in an ipython notebook to > something else then tk or inline? > > Kind regards, > Vasco > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From benjaminrk at gmail.com Fri Mar 28 14:27:47 2014 From: benjaminrk at gmail.com (MinRK) Date: Fri, 28 Mar 2014 11:27:47 -0700 Subject: [IPython-dev] IPython 2: disable automatically-added closing parentheses, brackets, and quotes In-Reply-To: <53358F2C.5050609@nasa.gov> References: <53358F2C.5050609@nasa.gov> Message-ID: Make sure you have an IPython profile: ipython profile create edit ~/.ipython/profile_default/static/custom/custom.js, adding the following: if (IPython.CodeCell) { IPython.CodeCell.options_default.cm_config.autoCloseBrackets = false; } -MinRK On Fri, Mar 28, 2014 at 8:03 AM, Kathleen Tacina wrote: > Hi, > > I've been using IPython notebook at lot in my work. It's a great tool, > and I like the edit/command modes adding in version 2. > > Unfortunately, though, the automatically-added closing parentheses, > brackets, and quotes are making it harder for me to edit code. (The > visual noise messes up my internal count of what I need to add.) How > would I turn this feature off? > > Thanks for your help with this and for making a great product! > > Best regards, > Kathleen Tacina > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From rmcgibbo at gmail.com Fri Mar 28 21:50:37 2014 From: rmcgibbo at gmail.com (Robert McGibbon) Date: Fri, 28 Mar 2014 18:50:37 -0700 Subject: [IPython-dev] Adding emacs-style keybindings for Ctrl-A and Ctrl-E in notebook Message-ID: Hey, I'm not sure if there's any more principled way to do this. There was some discussion on github a while ago about making the key bindings more configurable, but I'm not sure exactly what got merged. Here's the hacky way I came up with to add a couple of CodeMirror's emacs keybindings to the notebook using custom.js, in case anyone else ways to use it. Just sending it here for googling / feedback. -Robert ------------------ // custom.js "using strict"; $([IPython.events]).on('notebook_loaded.Notebook', function(){ console.log('Adding emacs-style keybindings'); var extraKeys = {'Ctrl-A': 'goLineStart', 'Ctrl-E': "goLineEnd"}; for (var k in extraKeys) { IPython.CodeCell.options_default.cm_config.extraKeys[k] = extraKeys[k]; } var cells = IPython.notebook.get_cells(); var numCells = cells.length; for (var i = 0; i < numCells; i++) { cells[i].code_mirror.setOption('extraKeys', IPython.CodeCell.options_default.cm_config.extraKeys); } }); -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Sat Mar 29 10:52:20 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Sat, 29 Mar 2014 07:52:20 -0700 Subject: [IPython-dev] Fwd: ShiningPanda CI, clap de fin In-Reply-To: References: Message-ID: Shiningpanda is shutting down. That's unfortunate, but its limitations were becoming increasingly apparent - for instance, there was no good way to install Casper. We've still got a month, but next week I'll start looking for a new CI system to complement Travis. Any suggestions or recommendations are welcome. Thanks, Thomas ---------- Forwarded message ---------- From: "Alexis Tabary" Date: 29 Mar 2014 05:15 Subject: ShiningPanda CI, clap de fin To: "Alexis Tabary" , "Olivier Mansion" < olivier.mansion at shiningpanda.com> Cc: Dear all, Today brings some sad news: we are shutting down ShiningPanda CI. Jenkins instances will remain available until April 30th, at which point they will all go offline. This decision became inevitable as we failed to keep ShiningPanda CI relevant in an increasingly competitive environment: our user base has been eroding for a rather long while, so we decided to move forward and simply halt the service. We are incredibly thankful to all the people who have been relying on ShiningPanda for integration and deployment, so we want to ensure as smooth a transition as possible. All users who wish to download Jenkins configuration files, or build history, should contact us. We might also be able to help with migrations to other services, in-house or hosted. Just send us an email. As for us - the team behind ShiningPanda CI - for now we are focusing on requires.io, a web service to keep track of dependencies on pypi. Thank you all for supporting us for so long. Olivier & Alexis -- Alexis Tabary ShiningPanda Founding Partner (+81) 90-6064-3886 Requires.io Monitor your Python dependencies ShiningPanda Consulting Services Build and Release Management -------------- next part -------------- An HTML attachment was scrubbed... URL: From fperez.net at gmail.com Sat Mar 29 16:28:53 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Sat, 29 Mar 2014 13:28:53 -0700 Subject: [IPython-dev] Fwd: ShiningPanda CI, clap de fin In-Reply-To: References: Message-ID: Thanks for the heads-up, I just saw a similar post from the yt team. Good thing we didn't purchase too much in one shot, it seems we actually got to spend most of the credit we paid for. Could you spec out what exactly are the needs we'd like to see covered that Travis doesn't provide? Having that clearly spelled out will be useful in making any future decision. Cheers f On Sat, Mar 29, 2014 at 7:52 AM, Thomas Kluyver wrote: > Shiningpanda is shutting down. That's unfortunate, but its limitations > were becoming increasingly apparent - for instance, there was no good way > to install Casper. We've still got a month, but next week I'll start > looking for a new CI system to complement Travis. Any suggestions or > recommendations are welcome. > > Thanks, > Thomas > ---------- Forwarded message ---------- > From: "Alexis Tabary" > Date: 29 Mar 2014 05:15 > Subject: ShiningPanda CI, clap de fin > To: "Alexis Tabary" , "Olivier Mansion" < > olivier.mansion at shiningpanda.com> > Cc: > > Dear all, > > Today brings some sad news: we are shutting down ShiningPanda CI. > Jenkins instances will remain available until April 30th, at which > point they will all go offline. > > This decision became inevitable as we failed to keep ShiningPanda CI > relevant in an increasingly competitive environment: our user base has > been eroding for a rather long while, so we decided to move forward > and simply halt the service. > > We are incredibly thankful to all the people who have been relying on > ShiningPanda for integration and deployment, so we want to ensure as > smooth a transition as possible. All users who wish to download > Jenkins configuration files, or build history, should contact us. We > might also be able to help with migrations to other services, in-house > or hosted. Just send us an email. > > As for us - the team behind ShiningPanda CI - for now we are focusing > on requires.io, a web service to keep track of dependencies on pypi. > > Thank you all for supporting us for so long. > > Olivier & Alexis > > -- > Alexis Tabary > ShiningPanda Founding Partner > (+81) 90-6064-3886 > > Requires.io > Monitor your Python dependencies > > ShiningPanda Consulting Services > Build and Release Management > > _______________________________________________ > IPython-dev mailing list > IPython-dev at scipy.org > http://mail.scipy.org/mailman/listinfo/ipython-dev > > -- Fernando Perez (@fperez_org; http://fperez.org) fperez.net-at-gmail: mailing lists only (I ignore this when swamped!) fernando.perez-at-berkeley: contact me here for any direct mail -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Sat Mar 29 19:00:06 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Sat, 29 Mar 2014 16:00:06 -0700 Subject: [IPython-dev] Fwd: ShiningPanda CI, clap de fin In-Reply-To: References: Message-ID: On 29 March 2014 13:28, Fernando Perez wrote: > Could you spec out what exactly are the needs we'd like to see covered > that Travis doesn't provide? Having that clearly spelled out will be useful > in making any future decision. > Testing on Windows is the biggest thing - that's what we started paying ShiningPanda for. Travis has had an issue open for multi OS support for a couple of years, and I believe they now have OS X, though they don't advertise it. However, it's also proved useful to have some extra test jobs - like the docs build, or creating a tarball and installing from it before running the tests - that we don't want to wait for on every pull request, but it's good to have something checking every day to every few days. We're also planning to make our JS tests run on Gecko using slimerjs, which we'll probably want to handle in a similar way. I've just done a quick search of the available options, and I could only see one hosted CI system with Windows support (Appveyor), and it doesn't appear to support Python. Previously, J?rgen set his machine up to run Windows builds, but I don't think he had time to maintain it. We may have to set up a VM on AWS/Azure/Google for Windows testing. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From fperez.net at gmail.com Sat Mar 29 19:26:15 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Sat, 29 Mar 2014 16:26:15 -0700 Subject: [IPython-dev] Fwd: ShiningPanda CI, clap de fin In-Reply-To: References: Message-ID: On Sat, Mar 29, 2014 at 4:00 PM, Thomas Kluyver wrote: > On 29 March 2014 13:28, Fernando Perez wrote: > >> Could you spec out what exactly are the needs we'd like to see covered >> that Travis doesn't provide? Having that clearly spelled out will be useful >> in making any future decision. >> > > Testing on Windows is the biggest thing - that's what we started paying > ShiningPanda for. Travis has had an issue open for multi OS support for a > couple of years, and I believe they now have OS X, though they don't > advertise it. However, it's also proved useful to have some extra test jobs > - like the docs build, or creating a tarball and installing from it before > running the tests - that we don't want to wait for on every pull request, > but it's good to have something checking every day to every few days. We're > also planning to make our JS tests run on Gecko using slimerjs, which we'll > probably want to handle in a similar way. > > I've just done a quick search of the available options, and I could only > see one hosted CI system with Windows support (Appveyor), and it doesn't > appear to support Python. Previously, J?rgen set his machine up to run > Windows builds, but I don't think he had time to maintain it. We may have > to set up a VM on AWS/Azure/Google for Windows testing. > Getting a VM on Azure sounds like a good solution. We can have several folks with access to it. Do you see any downsides with that approach? It can probably be spun up only on demand... Cheers f -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Sat Mar 29 19:29:43 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Sat, 29 Mar 2014 16:29:43 -0700 Subject: [IPython-dev] Fwd: ShiningPanda CI, clap de fin In-Reply-To: References: Message-ID: On 29 March 2014 16:26, Fernando Perez wrote: > Getting a VM on Azure sounds like a good solution. We can have several > folks with access to it. Do you see any downsides with that approach? It > can probably be spun up only on demand... The downside is simply that it's more work to set up than a dedicated CI service like ShiningPanda. However, I'm happy to spend some time on it, because I'd like a chance to practice working with a cloud platform. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From fperez.net at gmail.com Sat Mar 29 19:33:56 2014 From: fperez.net at gmail.com (Fernando Perez) Date: Sat, 29 Mar 2014 16:33:56 -0700 Subject: [IPython-dev] Fwd: ShiningPanda CI, clap de fin In-Reply-To: References: Message-ID: On Sat, Mar 29, 2014 at 4:29 PM, Thomas Kluyver wrote: > The downside is simply that it's more work to set up than a dedicated CI > service like ShiningPanda. However, I'm happy to spend some time on it, > because I'd like a chance to practice working with a cloud platform. > Do you think you've exhausted the search on available out-of-the box solutions? I wonder if someone has a Jenkins VM for windows that's ready to go and that could do the trick for us with minimal effort. We really want to have Windows CI up and running all the time, the last thing we need is to find about a serious Windows problem that crept up months earlier and went unnoticed (since few of the core team use Windows on a daily basis). But I also don't want this to become a time black hole... f -------------- next part -------------- An HTML attachment was scrubbed... URL: From takowl at gmail.com Sat Mar 29 19:58:44 2014 From: takowl at gmail.com (Thomas Kluyver) Date: Sat, 29 Mar 2014 16:58:44 -0700 Subject: [IPython-dev] Fwd: ShiningPanda CI, clap de fin In-Reply-To: References: Message-ID: On 29 March 2014 16:33, Fernando Perez wrote: > Do you think you've exhausted the search on available out-of-the box > solutions? I wonder if someone has a Jenkins VM for windows that's ready to > go and that could do the trick for us with minimal effort. We really want > to have Windows CI up and running all the time, the last thing we need is > to find about a serious Windows problem that crept up months earlier and > went unnoticed (since few of the core team use Windows on a daily basis). My impression from the searches I've done is that most out-of-the-box CI systems either target webapp type projects for which a Linux environment is a given, or .Net/C Windows apps. If we want to set up something ourself, the major options are buildbot and Jenkins. buildbot is in the classic open source style - scripts and config files - while Jenkins is a big Java web app that you configure through the web UI. Given those descriptions, I would lean towards buildbot, except that ShiningPanda uses Jenkins, so it's already familiar, and they've offered to export Jenkins config if we want it. I had a quick search for Jenkins/buildbot Windows VM images - there were some instructions for creating one, but none prepackaged. We probably want to install our own things (Python etc.) in the image anyway. ShiningPanda is keeping things running until the end of April, so we have a few weeks to work this out before our Windows builds stop. Thomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From Zach.Jones at netapp.com Sat Mar 29 20:27:46 2014 From: Zach.Jones at netapp.com (Jones, Zach H) Date: Sun, 30 Mar 2014 00:27:46 +0000 Subject: [IPython-dev] Problems using nbconvert with custom templates in IPython2.0-rc1 In-Reply-To: <8CE266D3F25EA84789937471DEF62F6927723157@SACEXCMBX02-PRD.hq.netapp.com> References: <8CE266D3F25EA84789937471DEF62F6927723157@SACEXCMBX02-PRD.hq.netapp.com> Message-ID: <8CE266D3F25EA84789937471DEF62F6927723186@SACEXCMBX02-PRD.hq.netapp.com> All, I have been using nbconvert to generate HTML views of notebooks with IPython1.x. I based my work on this: http://nbviewer.ipython.org/urls/raw.githubusercontent.com/ipython/ipython-in-depth/master/notebooks/07%20-%20NbConvert%20Python%20library.ipynb?create=1#Programatically-make-templates This work great with IPython1.x but is now not working with IPython2.0-rc. I am hoping you can help me sort through my errors. The code looks like this: extra_tpls = DictLoader({'na_full': ..., 'na_report', ...}) c = Config() c.HTMLExporter.default_template = 'na_report' exportHTML = HTMLExporter(config=c, extra_loaders=[extra_tpls]) exportHTML.from_notebook_node(nb) My DictLoader contains two templates: * na_full extends html_basic.tpl * na_report extends display_priority.tpl I get errors with either template, but each template is different. For na_full, it looks like html_basic.tpl cannot be found: --------------------------------------------------------------------------- TemplateNotFound Traceback (most recent call last) in () ----> 1 exportHTML.from_notebook_node(nb) /x/eng/rtpperf/python2.7-Debian-unstable-x86_64-tst/local/lib/python2.7/site-packages/IPython/nbconvert/exporters/templateexporter.pyc in from_notebook_node(self, nb, resources, **kw) 213 214 if self.template is not None: --> 215 output = self.template.render(nb=nb_copy, resources=resources) 216 else: 217 raise IOError('template file "%s" could not be found' % self.template_file) /x/eng/rtpperf/python2.7-Debian-unstable-x86_64-tst/local/lib/python2.7/site-packages/jinja2/environment.pyc in render(self, *args, **kwargs) 967 except Exception: 968 exc_info = sys.exc_info() --> 969 return self.environment.handle_exception(exc_info, True) 970 971 def stream(self, *args, **kwargs): /x/eng/rtpperf/python2.7-Debian-unstable-x86_64-tst/local/lib/python2.7/site-packages/jinja2/environment.pyc in handle_exception(self, exc_info, rendered, source_hint) 740 self.exception_handler(traceback) 741 exc_type, exc_value, tb = traceback.standard_exc_info --> 742 reraise(exc_type, exc_value, tb) 743 744 def join_path(self, template, parent):