Mailman 3 June 2007 - NumPy-Discussion

numpy requires tuples for indexing where Numeric allowed them?
by Jim Kleckner 08 Jun '07

08 Jun '07

I'm fighting conversion from Numeric to numpy. One change that doesn't seem documented is that I used to be able to select items using lists and now it seems that I have to convert them to tuples. Is that correct and is there a function buried in there that will accept a list for indices? Any reason that item() can't take a list? The weird thing is that it doesn't blow up right away when a list is passed in an array ref but rather returns something I don't expect. I work with lists rather than the implicit tuples of a function call because then I can work with arbitrary dimensions. In the meantime, I guess I can just convert the list to an otherwise unnecessary tuple. Jim

3 2

SciPy Journal
by Travis Oliphant 08 Jun '07

08 Jun '07

Hi everybody, I'm sorry for the cross posting, but I wanted to reach a wide audience and I know not everybody subscribes to all the lists. I've been thinking more about the "SciPy Journal" that we discussed before and I have some thoughts. 1) I'd like to get it going so that we can push out an electronic issue after the SciPy conference (in September) 2) I think it's scope should be limited to papers that describe algorithms and code that are in NumPy / SciPy / SciKits. Perhaps we could also accept papers that describe code that depends on NumPy / SciPy that is also easily available. 3) I'd like to make a requirement for inclusion of new code in SciPy that it have an associated journal article describing the algorithms, design approach, etc. I don't see this journal article as being user-interface documentation for the code. I see this is as a place to describe why the code is organized as it is and to detail any algorithms that are used. 4) The purpose of the journal as I see it is to a) provide someplace to document what is actually done in SciPy and related software. b) provide a teaching tool of numerical methods with actual "people use-it" code that would be useful to researchers, students, and professionals. c) hopefully clever new algorithms will be developed for SciPy by people using Python that could be show-cased here d) provide a peer-review publication opportunity for people who contribute to open-source software 5) We obviously need associate editors and people willing to review submitted articles as well as people willing to submit articles. I have two articles that can be submitted within the next two months. What do other people have? As an example of the kind of thing a SciPy Journal would be useful for. I have recently over-hauled the interpolation.py file for SciPy by incorporating the B-spline stuff that is partly in fitpack. In the process I noticed two things: 1) I have (what seems to me) a different recursive algorithm for calculating derivatives of B-splines than I could find in fitpack. 2) I have developed a different way to determine the K-1 extra degrees of freedom for Kth-order spline fitting than I have seen before. The SciPy Journal would be a great place to document both of these things while describing the spline interpolation design of scipy.interpolate It is true that I could submit this stuff to other journals, but it seems like that doing that makes the information harder to find in the future and not easier. I'm also dissatisfied with how information exclusionary academic journals seem to be. They are catching up, but they are still not as accessible as other things available on the internet. Given the open nature of most scientific research, it is remarkable that getting access to the information is not as easy as it should be with modern search engines (if your internet domain does not subscribe to the e-journal). Comments and feedback is welcome. -Travis

10 12

Weird numpy.arange behavoir
by Luke 07 Jun '07

07 Jun '07

I am integrating some equations and need to generate rank 1 time arrays to pass to my integrator. I need them to have the same interval between each entry and have the same number of elements. In matlab this is trivial, and it is in numpy as well, except I'm getting some weird behavoir: import numpy as N T = 0.1 dt = 0.01 k=0 t = N.arange(k*T,(k+1)*T+h,h) Output: array([ 0. , 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1 ]) So far so good. Here is where the problem arises: k=1 t = N.arange(k*T,(k+1)*T+h,h) array([ 0.1 , 0.11, 0.12, 0.13, 0.14, 0.15, 0.16, 0.17, 0.18, 0.19, 0.2 , 0.21]) Note that this time array has more entries, and in fact, the last entry is greater than (k+1)*T = (1+1)*0.01 = 2*0.01 = 0.2 Now if it was consistent for all k>0, then it would be fine. However, this is not the case: k=3 t = N.arange(k*T,(k+1)*T+h,h) Output: array([ 0.3 , 0.31, 0.32, 0.33, 0.34, 0.35, 0.36, 0.37, 0.38, 0.39, 0.4 ]) Now, this one has the same number of entries as the case where k=0. Can anybody: 1) Offer a solution to this? 2) Explain this behavoir would occur and ever be desirable? I read the numpy.arange docstring and it says that this may occur, but I don't understand why you would ever want this to occur. Apparently, the length of the returned array is: ceil((stop-start)/step) The weird thing is that in this simple example, (stop-start)/step is always exactly 11, since ((k+1)*T + h - k*T)/(h) = (T+h)/h = (0.1+0.01)/(0.01) = 11.0. In this, there shouldn't be any roundoff error. So in this simple example that was harmlessly constructed (i.e., my period time was an exact integer multiple of my step time), arange behaves undesirably (at least I think it does). After a few tests, I found that if instead of ceil, round was used, then it eliminated my problem, but I don't know if this would have other undesirable effects in other situations. I guess I could use range, but it is just a bit more tedious to code. Thanks, ~Luke

2 1

Re: [Numpy-discussion] Vista installer?
by Ryan Krauss 05 Jun '07

05 Jun '07

Success. I think this is officially a viable alternative for running Scipy/Numpy/Matplotlib/IPython in Vista. I turned off my firewall on my host (XP) and I was able to browse the internet in Ubuntu. Andrew was right that I still couldn't ping, but I seem to have everything else. And with shared folders enabled I seem to have some default SMB stuff set up. I couldn't successfully browse the shares of my partitions (C$, E$, ...), but I did specifically share a folder (from windows) and editted a Python file in that folder from within VMware. The available shared folders showed up under Places > Network as I should have expected for Ubuntu SMB shares. Another of my students supposedly has everything working in Vista with no problems. I will have to check that out and report back. Ryan On 5/29/07, Andrew Straw <strawman(a)astraw.com> wrote: > Ryan Krauss wrote: > > Thanks for your help in getting me up and going with VMWare. Andrew, > > your virtual appiance seems to work quite well. > > > > I have two hurdles left (I hope there are only 2). The first is still > > networking. I have NAT chosen in the VMWare player and that seems to > > give me limited connectivity. I seem to be able to use apt but can't > > ping or browse the internet. Any thoughts on how to trouble shoot > > this? > I'm guessing that the firewall on the host computer is preventing a lot > of this. I'm not sure I'd expect ping to work ever over NAT, as that's a > UDP thing, which is stateless and thus my guess is that NAT doesn't know > where to direct the incoming packet. However, I think you should be able > to browse, and the fact that apt works but browsing doesn't is certainly > odd, since the standard apt repositories are simply web servers. > > You can try bridged mode. That's usually what I do, but I figured it was > less likely to work than NAT (because it would require 2 external IP > addresses, and I'm not sure how your network is configured), so although > it allows full connectivity, including UDP, I chose NAT for that image. > > > > My second hurdle is how to move files between the host and guest OS. > > I see a shared folders option in the player, but don't seem to be able > > to add to the list of folders. My students need to be able to import > > python modules I give them and then submit files they would be > > developing within the guest OS. I googled for vmware shared folders > > and it seems like a non-free option. > Hmm, I haven't done any sharing through VMware server. On my mac, I have > using VMware fusion, but generally found that using SSH was more > reliable. So, maybe you can have them SSH (well, scp, perhaps using > WinSCP on the Windows side) between their two virtual computers? > > > > I haven't found a good "VMWare for dummies" or getting started guide, > > so feel free to redirect me. > VMware's docs are pretty good but necessarily get indexed by Google -- > make sure you browse their website. > > Good luck, > Andrew > > > > Thanks, > > > > Ryan > > > > On 5/29/07, David Cournapeau <david(a)ar.media.kyoto-u.ac.jp> wrote: > >> Ryan Krauss wrote: > >> > Hey David, > >> > > >> > Thanks for all your help. I run Ubuntu stand alone on my dual boot > >> > machines and have no network issues (actually, the computer running > >> > VMWare is dual boot). So, I think my network issues are VMWare > >> > specific. I don't know how it handles the hardware. > >> In this case, it should be even easier. In ethernet options of vmware, > >> you should choose NAT, and then vmware will have a dhcp server on the > >> HOST which will provide an address by dhcp to your guest. > >> > >> David > >> > >

2 2

signal processing chapter for book
by Giorgio Luciano 04 Jun '07

04 Jun '07

first of all sorry for cross posting As I wrote some time ago we are trying to write a book proposal about the use of python/scipy/numpy in chemometrics and analytical chemistry. Since now I've received positive answer from eight authors and the only "missing" chapter is one about the use of python in digital signal processing (I've contacted some possible authors but since now they are busy). The schedule will not too tight and the chapter doesnt' need to be too long. Hope to hear from you soon Giorgio

1 0

corrcoef of masked array
by Jesper Larsen 04 Jun '07

04 Jun '07

Hi numpy users, I have a masked array of dimension (nvariables, nobservations) that contain missing values at arbitrary points. Is it safe to rely on numpy.corrcoeff to calculate the correlation coefficients of a masked array (it seems to give reasonable results)? Cheers, Jesper

3 5

Re: [Numpy-discussion] GPU implementation?
by Bob Lewis 04 Jun '07

04 Jun '07

James Turner wrote: > Hi Martin, > > > I was wondering if anyone has thought about accelerating NumPy with a > > GPU. For example nVidia's CUDA SDK provides a feasible way to offload > > vector math onto the very fast SIMD processors available on the GPU. > > Currently GPUs primarily support single precision floats and are not > > IEEE compliant, but still could be useful for some applications. > > I wasn't actually there, but I noticed that last year's SciPy > conference page includes a talk entitled "GpuPy: Using GPUs to > Accelerate NumPy", by Benjamin Eitzen (I think I also found his Web > page via Google): > > http://www.scipy.org/SciPy2006/Schedule > > I also wondered whether Benjamin or anyone else who is interested had > come across the Open Graphics Project (hadn't got around to asking)? Thanks for your interest. Ben and I (mostly Ben, it's his MS thesis) are working on "gpupy" and expect to have a version ready for testing by people other than ourselves some time this summer. (Very) preliminary results are promising. - Bob Lewis School of EECS Washington State University

2 1

Installing from egg ?
by Gary Pajer 04 Jun '07

04 Jun '07

Because I'm a glutton for punishment, I'm starting to install things using easy_install. I'm very new at this, and it's not going well. python 2.5 / WinXP (and Kubuntu, but we'll start with Windows) if I type easy_install numpy stuff downloads, but stops with a prompt to select a compiler by passing -c xxxxx to setup.py. I have Mingw installed, and atlas binaries, and have no trouble building numpy from svn the old fashioned way. Is the behavior up to now expected? How do I pass -c to setup.py via easy_install (misnamed, imho. maybe mho will chage with time) regards, gary

2 1

are there any numpy equivalents to MATLAB nnz(), sparse(), sparsity()?
by dmitrey 04 Jun '07

04 Jun '07

are there any numpy equivalents to MATLAB nnz(), sparse(), sparsity()? I didn't see the ones in numpy for MATLAB users page. Thx, D.

2 1

numpy r3857 build problem
by Christopher Hanley 04 Jun '07

04 Jun '07

Hi, I cannot build the latest version of numpy in svn (r3857) on my Intel MacBook running OSX 10.4.9. I'm guessing that the problem is that a fortran compiler isn't found. Since NUMPY doesn't require FORTRAN I found this surprising. Has there been a change in policy? I'm attaching the build log to this message. Cheers, Chris Running from numpy source directory. non-existing path in 'numpy/distutils': 'site.cfg' F2PY Version 2_3857 blas_opt_info: FOUND: extra_link_args = ['-Wl,-framework', '-Wl,Accelerate'] define_macros = [('NO_ATLAS_INFO', 3)] extra_compile_args = ['-msse3', '-I/System/Library/Frameworks/vecLib.framework/Headers'] lapack_opt_info: FOUND: extra_link_args = ['-Wl,-framework', '-Wl,Accelerate'] define_macros = [('NO_ATLAS_INFO', 3)] extra_compile_args = ['-msse3'] running install running build running config_cc unifing config_cc, config, build_clib, build_ext, build commands --compiler options running config_fc unifing config_fc, config, build_clib, build_ext, build commands --fcompiler options running build_src building py_modules sources creating build creating build/src.macosx-10.3-i386-2.5 creating build/src.macosx-10.3-i386-2.5/numpy creating build/src.macosx-10.3-i386-2.5/numpy/distutils building extension "numpy.core.multiarray" sources creating build/src.macosx-10.3-i386-2.5/numpy/core Generating build/src.macosx-10.3-i386-2.5/numpy/core/config.h customize NAGFCompiler Could not locate executable f95 customize AbsoftFCompiler customize IBMFCompiler customize IntelFCompiler customize GnuFCompiler Traceback (most recent call last): File "setup.py", line 89, in <module> setup_package() File "setup.py", line 82, in setup_package configuration=configuration ) File "/Users/chanley/dev/numpy/numpy/distutils/core.py", line 176, in setup return old_setup(**new_attr) File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/core.py", line 151, in setup dist.run_commands() File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/dist.py", line 974, in run_commands self.run_command(cmd) File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/dist.py", line 994, in run_command cmd_obj.run() File "/Users/chanley/dev/numpy/numpy/distutils/command/install.py", line 16, in run r = old_install.run(self) File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/command/install.py", line 506, in run self.run_command('build') File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/cmd.py", line 333, in run_command self.distribution.run_command(command) File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/dist.py", line 994, in run_command cmd_obj.run() File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/command/build.py", line 112, in run self.run_command(cmd_name) File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/cmd.py", line 333, in run_command self.distribution.run_command(command) File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/dist.py", line 994, in run_command cmd_obj.run() File "/Users/chanley/dev/numpy/numpy/distutils/command/build_src.py", line 130, in run self.build_sources() File "/Users/chanley/dev/numpy/numpy/distutils/command/build_src.py", line 147, in build_sources self.build_extension_sources(ext) File "/Users/chanley/dev/numpy/numpy/distutils/command/build_src.py", line 250, in build_extension_sources sources = self.generate_sources(sources, ext) File "/Users/chanley/dev/numpy/numpy/distutils/command/build_src.py", line 307, in generate_sources source = func(extension, build_dir) File "numpy/core/setup.py", line 51, in generate_config_h library_dirs = default_lib_dirs) File "/usr/stsci/pyssgdev/Python-2.5/lib/python2.5/distutils/command/config.py", line 278, in try_run self._check_compiler() File "/Users/chanley/dev/numpy/numpy/distutils/command/config.py", line 30, in _check_compiler dry_run=self.dry_run, force=1) File "/Users/chanley/dev/numpy/numpy/distutils/fcompiler/__init__.py", line 785, in new_fcompiler compiler = get_default_fcompiler(plat, requiref90=requiref90) File "/Users/chanley/dev/numpy/numpy/distutils/fcompiler/__init__.py", line 769, in get_default_fcompiler requiref90=requiref90) File "/Users/chanley/dev/numpy/numpy/distutils/fcompiler/__init__.py", line 721, in _find_existing_fcompiler c.customize(dist) File "/Users/chanley/dev/numpy/numpy/distutils/fcompiler/__init__.py", line 506, in customize self.set_library_dirs(self.get_library_dirs()) File "/Users/chanley/dev/numpy/numpy/distutils/fcompiler/gnu.py", line 126, in get_library_dirs d = self.get_libgcc_dir() File "/Users/chanley/dev/numpy/numpy/distutils/fcompiler/gnu.py", line 118, in get_libgcc_dir use_tee=0) File "/Users/chanley/dev/numpy/numpy/distutils/exec_command.py", line 244, in exec_command **env) File "/Users/chanley/dev/numpy/numpy/distutils/exec_command.py", line 263, in _exec_command_posix command_str = ' '.join(list(command)) TypeError: sequence item 0: expected string, NoneType found

2 2