Hi all, I'm not sure if this is the right list to send this to, but I think a bunch of knowledgeable people read this one, and hopefully one of you can help me. I'm trying to use Reason and I'm kind of confused about the ssh tunneling part. I start up yt serve on a remote machine (ranger) and I'm told to create a ssh tunnel by executing these commands: ~C -L55185:localhost:55185 My local machine doesn't recognize "~C" and I'm not sure what it is either. I tried creating an ssh tunnel with this: ssh tg803911@ranger.tacc.utexas.edu -L 55185:localhost:55185 I was then able to successfully open the gui in a browser with the url provided, but I am unable to load data (I can view the help though). In order to do the ssh tunnel properly do I need to install something on my local machine? Christine
Hi Christine, On Thu, Apr 21, 2011 at 3:46 PM, Christine Simpson <csimpson@astro.columbia.edu> wrote:
Hi all,
I'm not sure if this is the right list to send this to, but I think a bunch of knowledgeable people read this one, and hopefully one of you can help me. I'm trying to use Reason and I'm kind of confused about the ssh tunneling part. I start up yt serve on a remote machine (ranger) and I'm told to create a ssh tunnel by executing these commands:
~C -L55185:localhost:55185
My local machine doesn't recognize "~C" and I'm not sure what it is either. I tried creating an ssh tunnel with this:
Ah, so ~C should dump you at an ssh console. Sometimes it takes a couple tries, but you should just have to type "~C" right in the window and an ssh prompt should come up, that looks like this: ssh> Then you would type -L ...
ssh tg803911@ranger.tacc.utexas.edu -L 55185:localhost:55185
As a quick note, the ports change randomly for each instance of reason. If someone were to know the hostname, port, AND the UUID of your session, they could run as you. (Not cool. :)
I was then able to successfully open the gui in a browser with the url provided, but I am unable to load data (I can view the help though). In order to do the ssh tunnel properly do I need to install something on my local machine?
It sounds like if you were able to open the GUI, you are already there. Have you been able to "load" any data in the window? For instance, load("/path/to/your/data") and then hitting shift-enter, to send the command? When this is done, it should pop up with a new item in the left hand side. We're still working on the interface, so these comments and issues ARE very useful, particularly if they present stumbling blocks. Best, Matt
Christine
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Okay, I tried ~C a few times and my system keeps telling me "command not found." In the gui I was able to open, I tried loading the data as you suggest with the full path and hitting shift+enter. A red box pops up around the yt input panel and nothing pops up on the left. The yt input panel is then unresponsive (I can't type in it anymore). However, in the terminal where I've set up yt serve, all the text outputs that yt spits out when it loads data are output: yt : [INFO ] 2011-04-21 15:19:13,126 Parameters: current_time = 35.8164705155 yt : [INFO ] 2011-04-21 15:19:13,126 Parameters: domain_dimensions = [128, 128, 128] yt : [INFO ] 2011-04-21 15:19:13,127 Parameters: domain_left_edge = [ 0. 0. 0.] yt : [INFO ] 2011-04-21 15:19:13,131 Parameters: domain_right_edge = [ 1. 1. 1.] yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: cosmological_simulation = 1 yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: current_redshift = 7.03540239852 yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: omega_lambda = 0.726 yt : [INFO ] 2011-04-21 15:19:13,133 Parameters: omega_matter = 0.274 yt : [INFO ] 2011-04-21 15:19:13,133 Parameters: hubble_constant = 0.705 On Thu, 2011-04-21 at 16:02 -0400, Matthew Turk wrote:
Hi Christine,
On Thu, Apr 21, 2011 at 3:46 PM, Christine Simpson <csimpson@astro.columbia.edu> wrote:
Hi all,
I'm not sure if this is the right list to send this to, but I think a bunch of knowledgeable people read this one, and hopefully one of you can help me. I'm trying to use Reason and I'm kind of confused about the ssh tunneling part. I start up yt serve on a remote machine (ranger) and I'm told to create a ssh tunnel by executing these commands:
~C -L55185:localhost:55185
My local machine doesn't recognize "~C" and I'm not sure what it is either. I tried creating an ssh tunnel with this:
Ah, so ~C should dump you at an ssh console. Sometimes it takes a couple tries, but you should just have to type "~C" right in the window and an ssh prompt should come up, that looks like this:
ssh>
Then you would type -L ...
ssh tg803911@ranger.tacc.utexas.edu -L 55185:localhost:55185
As a quick note, the ports change randomly for each instance of reason. If someone were to know the hostname, port, AND the UUID of your session, they could run as you. (Not cool. :)
I was then able to successfully open the gui in a browser with the url provided, but I am unable to load data (I can view the help though). In order to do the ssh tunnel properly do I need to install something on my local machine?
It sounds like if you were able to open the GUI, you are already there. Have you been able to "load" any data in the window? For instance,
load("/path/to/your/data")
and then hitting shift-enter, to send the command? When this is done, it should pop up with a new item in the left hand side.
We're still working on the interface, so these comments and issues ARE very useful, particularly if they present stumbling blocks.
Best,
Matt
Christine
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Hi Christine, On Thu, Apr 21, 2011 at 4:28 PM, Christine Simpson <csimpson@astro.columbia.edu> wrote:
Okay,
I tried ~C a few times and my system keeps telling me "command not found."
Ah, you should be hitting it while the yt serve command is going -- i.e., run yt serve, and in that same terminal window while yt serve is running, type "~C".
In the gui I was able to open, I tried loading the data as you suggest with the full path and hitting shift+enter. A red box pops up around the yt input panel and nothing pops up on the left. The yt input panel is then unresponsive (I can't type in it anymore). However, in the terminal where I've set up yt serve, all the text outputs that yt spits out when it loads data are output:
yt : [INFO ] 2011-04-21 15:19:13,126 Parameters: current_time = 35.8164705155 yt : [INFO ] 2011-04-21 15:19:13,126 Parameters: domain_dimensions = [128, 128, 128] yt : [INFO ] 2011-04-21 15:19:13,127 Parameters: domain_left_edge = [ 0. 0. 0.] yt : [INFO ] 2011-04-21 15:19:13,131 Parameters: domain_right_edge = [ 1. 1. 1.] yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: cosmological_simulation = 1 yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: current_redshift = 7.03540239852 yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: omega_lambda = 0.726 yt : [INFO ] 2011-04-21 15:19:13,133 Parameters: omega_matter = 0.274 yt : [INFO ] 2011-04-21 15:19:13,133 Parameters: hubble_constant = 0.705
This looks like it could be a bug. (That's why reason is still in beta! :) If you hop into IRC or stop by my office we can take a look at this. Thanks for testing this all out while it's still new! -Matt
On Thu, 2011-04-21 at 16:02 -0400, Matthew Turk wrote:
Hi Christine,
On Thu, Apr 21, 2011 at 3:46 PM, Christine Simpson <csimpson@astro.columbia.edu> wrote:
Hi all,
I'm not sure if this is the right list to send this to, but I think a bunch of knowledgeable people read this one, and hopefully one of you can help me. I'm trying to use Reason and I'm kind of confused about the ssh tunneling part. I start up yt serve on a remote machine (ranger) and I'm told to create a ssh tunnel by executing these commands:
~C -L55185:localhost:55185
My local machine doesn't recognize "~C" and I'm not sure what it is either. I tried creating an ssh tunnel with this:
Ah, so ~C should dump you at an ssh console. Sometimes it takes a couple tries, but you should just have to type "~C" right in the window and an ssh prompt should come up, that looks like this:
ssh>
Then you would type -L ...
ssh tg803911@ranger.tacc.utexas.edu -L 55185:localhost:55185
As a quick note, the ports change randomly for each instance of reason. If someone were to know the hostname, port, AND the UUID of your session, they could run as you. (Not cool. :)
I was then able to successfully open the gui in a browser with the url provided, but I am unable to load data (I can view the help though). In order to do the ssh tunnel properly do I need to install something on my local machine?
It sounds like if you were able to open the GUI, you are already there. Have you been able to "load" any data in the window? For instance,
load("/path/to/your/data")
and then hitting shift-enter, to send the command? When this is done, it should pop up with a new item in the left hand side.
We're still working on the interface, so these comments and issues ARE very useful, particularly if they present stumbling blocks.
Best,
Matt
Christine
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Hi, I have it working. For whatever reason, it didn't occur to me to open the tunnel in the window I was running yt serve. My bad. Opening the tunnel separately didn't allow me to load the data properly. Thanks for your help Matt. Christine On Thu, 2011-04-21 at 16:31 -0400, Matthew Turk wrote:
Hi Christine,
On Thu, Apr 21, 2011 at 4:28 PM, Christine Simpson <csimpson@astro.columbia.edu> wrote:
Okay,
I tried ~C a few times and my system keeps telling me "command not found."
Ah, you should be hitting it while the yt serve command is going -- i.e., run yt serve, and in that same terminal window while yt serve is running, type "~C".
In the gui I was able to open, I tried loading the data as you suggest with the full path and hitting shift+enter. A red box pops up around the yt input panel and nothing pops up on the left. The yt input panel is then unresponsive (I can't type in it anymore). However, in the terminal where I've set up yt serve, all the text outputs that yt spits out when it loads data are output:
yt : [INFO ] 2011-04-21 15:19:13,126 Parameters: current_time = 35.8164705155 yt : [INFO ] 2011-04-21 15:19:13,126 Parameters: domain_dimensions = [128, 128, 128] yt : [INFO ] 2011-04-21 15:19:13,127 Parameters: domain_left_edge = [ 0. 0. 0.] yt : [INFO ] 2011-04-21 15:19:13,131 Parameters: domain_right_edge = [ 1. 1. 1.] yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: cosmological_simulation = 1 yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: current_redshift = 7.03540239852 yt : [INFO ] 2011-04-21 15:19:13,132 Parameters: omega_lambda = 0.726 yt : [INFO ] 2011-04-21 15:19:13,133 Parameters: omega_matter = 0.274 yt : [INFO ] 2011-04-21 15:19:13,133 Parameters: hubble_constant = 0.705
This looks like it could be a bug. (That's why reason is still in beta! :) If you hop into IRC or stop by my office we can take a look at this. Thanks for testing this all out while it's still new!
-Matt
On Thu, 2011-04-21 at 16:02 -0400, Matthew Turk wrote:
Hi Christine,
On Thu, Apr 21, 2011 at 3:46 PM, Christine Simpson <csimpson@astro.columbia.edu> wrote:
Hi all,
I'm not sure if this is the right list to send this to, but I think a bunch of knowledgeable people read this one, and hopefully one of you can help me. I'm trying to use Reason and I'm kind of confused about the ssh tunneling part. I start up yt serve on a remote machine (ranger) and I'm told to create a ssh tunnel by executing these commands:
~C -L55185:localhost:55185
My local machine doesn't recognize "~C" and I'm not sure what it is either. I tried creating an ssh tunnel with this:
Ah, so ~C should dump you at an ssh console. Sometimes it takes a couple tries, but you should just have to type "~C" right in the window and an ssh prompt should come up, that looks like this:
ssh>
Then you would type -L ...
ssh tg803911@ranger.tacc.utexas.edu -L 55185:localhost:55185
As a quick note, the ports change randomly for each instance of reason. If someone were to know the hostname, port, AND the UUID of your session, they could run as you. (Not cool. :)
I was then able to successfully open the gui in a browser with the url provided, but I am unable to load data (I can view the help though). In order to do the ssh tunnel properly do I need to install something on my local machine?
It sounds like if you were able to open the GUI, you are already there. Have you been able to "load" any data in the window? For instance,
load("/path/to/your/data")
and then hitting shift-enter, to send the command? When this is done, it should pop up with a new item in the left hand side.
We're still working on the interface, so these comments and issues ARE very useful, particularly if they present stumbling blocks.
Best,
Matt
Christine
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
I am trying to make a multi-plot of phase plots, but running into difficulty. Phase plots seem to be different from the other plot objects I am used to dealing with. If I copy the cookbook version of multi_plot.py, but replace the lines like p = pc.add_slice("Density", 0, figure = fig, axes = axes[0][0], use_colorbar=False) with the global_phase_plots.py cookbook lines dd = pf.h.all_data() p = pc.add_phase_object(dd, ["Density","Temperature","x-velocity"], figure=fig, axes=axes[0][0], use_colorbar=False) The code immediately crashes because phase plots do not take the use_colorbar argument. When I take out 'use_colorbar=False', the script runs, but the resulting plot is all purple (the lowest value on the default colorbar). This is especially confusing because phase plots normally have lots of white no matter what you do with the colorbar. Does anyone know how to fix this? For what it is worth, I am still on the old svn trunk (I still use Raven). Thanks, Charles
Hi Charles, On Fri, Apr 22, 2011 at 8:27 PM, Charles Hansen <chansen@astro.berkeley.edu> wrote:
I am trying to make a multi-plot of phase plots, but running into difficulty. Phase plots seem to be different from the other plot objects I am used to dealing with.
If I copy the cookbook version of multi_plot.py, but replace the lines like
p = pc.add_slice("Density", 0, figure = fig, axes = axes[0][0], use_colorbar=False)
with the global_phase_plots.py cookbook lines
dd = pf.h.all_data() p = pc.add_phase_object(dd, ["Density","Temperature","x-velocity"], figure=fig, axes=axes[0][0], use_colorbar=False)
The code immediately crashes because phase plots do not take the use_colorbar argument. When I take out 'use_colorbar=False', the script runs, but the resulting plot is all purple (the lowest value on the default colorbar). This is especially confusing because phase plots normally have lots of white no matter what you do with the colorbar.
Does anyone know how to fix this? For what it is worth, I am still on the old svn trunk (I still use Raven).
I have replicated the error using the multi_plot.py script. I was able to get rid of it by calling _redraw_image() on each of the plot objects... but in general it's a bit ugly and I'm not sure that I'm entirely happy with how it works for phase plots. You might be better suited to setting up subplot() calls directly and feeding those axes; multi_plot is mostly useful when you have multiple synchronized colorbars, which is tricky with phase plots. Unfortunately I don't think I can backport something to make it easier to the old SVN version at this time. Best, Matt
Thanks, Charles _______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Hi Matt, Do you mind if I resurrect this discussion? I am trying to do something similar, create a multiplot phase plot. Could you say a bit more about how to feed the yt phase plots to subplot()? Thanks Christine On Fri, 2011-04-22 at 21:44 -0400, Matthew Turk wrote:
I have replicated the error using the multi_plot.py script. I was able to get rid of it by calling _redraw_image() on each of the plot objects... but in general it's a bit ugly and I'm not sure that I'm entirely happy with how it works for phase plots. You might be better suited to setting up subplot() calls directly and feeding those axes; multi_plot is mostly useful when you have multiple synchronized colorbars, which is tricky with phase plots. Unfortunately I don't think I can backport something to make it easier to the old SVN version at this time.
Best,
Matt
It is kinda tricky. With everything except for phaseplot, you can use something like... (using matplotlib) pc = PlotCollection(....) subplot(211) fig1 = gcf() axis1 = gca() p = pc.add_slice(...., figure=fig1, axes=axis1, use_colorbar=False) suplot(212) fig2 = gcf() axis2 = gca() p = pc.add_slice(...., figure=fig2, axes=axis2, use_colorbar=False) With phaseplot, it just doesn't work. I do not know why. I eventually just saved each figure individual and pasted them together using imagemagick convert. For the sake of argument suppose you want a 'quad chart' arrangement of four figures named as 1.png through 4.png. Use convert to stick them together: convert +append 1.png 2.png row1.png convert +append 3.png 4.png row2.png convert -append row1.png row2.png final_image.png I'd be happy to know if there is a better way. Charles On 6/7/2011 3:52 PM, Christine Simpson wrote:
Hi Matt,
Do you mind if I resurrect this discussion? I am trying to do something similar, create a multiplot phase plot. Could you say a bit more about how to feed the yt phase plots to subplot()?
Thanks Christine
On Fri, 2011-04-22 at 21:44 -0400, Matthew Turk wrote:
I have replicated the error using the multi_plot.py script. I was able to get rid of it by calling _redraw_image() on each of the plot objects... but in general it's a bit ugly and I'm not sure that I'm entirely happy with how it works for phase plots. You might be better suited to setting up subplot() calls directly and feeding those axes; multi_plot is mostly useful when you have multiple synchronized colorbars, which is tricky with phase plots. Unfortunately I don't think I can backport something to make it easier to the old SVN version at this time.
Best,
Matt
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
Hi Christine and Charles, This is something I've been looking at recently. The PlotCollection as a whole has not aged ... particularly well. It was designed for a mode of operation which is 4 years obsolete at this point. To that end, I've been working on isolated phase and profile plots, and Jeff Oishi has been working on the Plot Window. These are designed to be independent of Plot Collections, and are supposed to be more agile; more control is left to the user for setting up contexts, and fewer knobs and twiddles will be provided, on the basis that ... it's hard to get that right reliably. The phase plots and profile plots currently work, and I have been using them for a paper I am currently writing. I've also figured out how to correlated colorbars across image plots of projections&slices, if you end up needing any code for that, inside the multiplot context. The sample script I came up with for generating phase plots is here: http://paste.enzotools.org/show/1689/ The idea here is that it creates a figure, and then for each plot it creates a new subplot, into which it plots our plot. You can then use the standard Matplotlib colors and limits and so on and so forth. Hope that helps, Matt On Tue, Jun 7, 2011 at 4:08 PM, Charles Hansen <chansen@astro.berkeley.edu> wrote:
It is kinda tricky. With everything except for phaseplot, you can use something like...
(using matplotlib)
pc = PlotCollection(....) subplot(211) fig1 = gcf() axis1 = gca() p = pc.add_slice(...., figure=fig1, axes=axis1, use_colorbar=False)
suplot(212) fig2 = gcf() axis2 = gca() p = pc.add_slice(...., figure=fig2, axes=axis2, use_colorbar=False)
With phaseplot, it just doesn't work. I do not know why.
I eventually just saved each figure individual and pasted them together using imagemagick convert. For the sake of argument suppose you want a 'quad chart' arrangement of four figures named as 1.png through 4.png. Use convert to stick them together:
convert +append 1.png 2.png row1.png convert +append 3.png 4.png row2.png convert -append row1.png row2.png final_image.png
I'd be happy to know if there is a better way.
Charles
On 6/7/2011 3:52 PM, Christine Simpson wrote:
Hi Matt,
Do you mind if I resurrect this discussion? I am trying to do something similar, create a multiplot phase plot. Could you say a bit more about how to feed the yt phase plots to subplot()?
Thanks Christine
On Fri, 2011-04-22 at 21:44 -0400, Matthew Turk wrote:
I have replicated the error using the multi_plot.py script. I was able to get rid of it by calling _redraw_image() on each of the plot objects... but in general it's a bit ugly and I'm not sure that I'm entirely happy with how it works for phase plots. You might be better suited to setting up subplot() calls directly and feeding those axes; multi_plot is mostly useful when you have multiple synchronized colorbars, which is tricky with phase plots. Unfortunately I don't think I can backport something to make it easier to the old SVN version at this time.
Best,
Matt
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
participants (3)
-
Charles Hansen -
Christine Simpson -
Matthew Turk