Mailman 3 February 2009 - SciPy-Dev

some new ode solvers
by John Travers 24 Feb '09

24 Feb '09

Hi All, Attached is a patch which adds two new ODE solvers to the scipy.integrate.ode module. The solvers are dopri5 and dop853, which are explicit Runge-Kutta pairs originally developed by Dormand and Prince. The fortran code was downloaded from: http://www.unige.ch/~hairer/software.html The license is clearly BSD and SciPy compatible. These are excellent solvers, described in detail in the authors book: Solving Ordinary Differential Equations. Nonstiff Problems. 2nd edition. Springer Series in Comput. Math., vol. 8. The dopri5 code is what Matlab's ode45 is based on. I think they would be a good addition to SciPy and I have used them often (in Fortran). The attached patch tries to follow the (somewhat strange to me) coding practices of the current ode module. I have added them to the test suite, but note that there are not many tests! (I might add some if I get time). I have also tested them with my own code which uses the ode module. If I get the go ahead from some regular SciPy contributers then I'll go ahead and commit this patch. I think I still have SVN access, but I wanted this pacth to be reviewed first as it has been a long time since I did anything with SciPy. Cheers, John -- Telephone: (+44) (0) 7739 105209

8 10

Subclassed ndarray fails with ValueError when assigning to a sliced array
by Christoph T. Weidemann 23 Feb '09

23 Feb '09

Per, who created ticket 1023 tried to send the following message to the mailing list, but it bounced, so I am resending it on his behalf. ================================================================= Hi Folks: Someone mentioned that I should send an email to the mailing list when reporting a bug / issue, so here's the repeat of a new ticket I just opened: http://scipy.org/scipy/numpy/ticket/1023 I've been working on a subclass of ndarray that has named dimensions that you can use to slice within a custom __getitem__. For example, you can do something like this: x['time>0'] to slice the array where the time dimension is greater than zero. This is all working nicely, right up until I try to assign to a slice like this: x['time>0'] += 1 or this x['time>0'] = x['time>0'] + 1 giving me the following error: ValueError: field named time>0 not found. I'm not sure what is raising this error (the interpreter gives no module or line number that's raising it), so I'm scared it's actually in the interpreter or somewhere deep in the numpy code, but I figured I'd try here first. Any ideas for what's going on here? All assignments with standard slices work just fine. Is there some code in numpy that is checking the contents of __getitem__ and causing it to fail here? Thanks for any thoughts, Per PS-> I'm running this on NumPy? 1.1.0 in Debian Lenny.

2 3

scipy wiki throwing 500 errors
by Peter Skomoroch 23 Feb '09

23 Feb '09

It seems like the wiki still has some server issues, I've had several 500 errors while attempting to edit the wiki in the last few minutes. -- Peter N. Skomoroch 617.285.8348 http://www.datawrangling.com http://delicious.com/pskomoroch http://twitter.com/peteskomoroch

2 1

scipy.org is down
by Nils Wagner 23 Feb '09

23 Feb '09

scipy.org is down again...

3 2

new member.
by Kevin Daley 22 Feb '09

22 Feb '09

Hi, all. Just want to let everyone know I'm here. Name's Kevin Daley, from Atlanta. What I know: math: everything through linear algebra Lie calculus/Functional Calculus/Geometric Calculus (usually) chaos theory differential geometry (still needs a bit of polishing) Functional Integrals Stochastic calculus Fourier Theory/Harmonic Analysis/Wavelets science: Everything through Classical Mechanics (of course) Statistical Thermodynamics General Relativity/ESCK gravity (still needs a little work) Quantum Electrodynamics (am learning quantum optics) Fluid Dynamics/Plasmas/Turbulence statistical mechanics/information theory (some) physical neuroscience/EEG analysis computers: Python (8 years' experience. also fluent in C/C++) GPU parallelism (approx. 2-year CUDA experience, learning OpenCL) a variety of grid-based simulation techniques (the list is too long) wavelet compression (a little) computer graphics (a lot) basic FEM/level-set stuff bunch of api-level stuff: OpenGL, cg/glsl/hlsl, most widely-used python modules (and a few very arcane ones), miscellaneous: excellent verbal skills. when I'm available: most of the time. what I am most excited to help with: suggesting and helping out with new features/enhancements, extensions to new platforms, optimizations of numeric methods, user interface, integration with other modules. Everyone at scipy has been doing a great job...it's great software. But I don't see why we can't compete with industry technology's feature set----while providing extra modularity, flexibility, portability, and extensibility. You'll hear back from me soon. Cheers!

2 1

Re: [SciPy-dev] new member.
by Joe Harrington 22 Feb '09

22 Feb '09

Hi Kevin, welcome to the list. > But I don't see why we can't compete with industry technology's > feature set----while providing extra modularity, flexibility, > portability, and extensibility. I think the answer is simply one of available labor and coordination, so if you're willing to work, you are part of the solution. Others will have differing opinions, but I feel we are weakest, still, in documentation and packaging. Both areas are improving, but both have a long way to go. So, if you want to get this package adopted fastest, helping in those areas would probably bring the biggest bang per hour spent. My position is that if we make this package easy for new people to install and learn (as in, click in synaptic or yum or whatever package manager you use, read an obviously-placed getting started doc, and start playing), it will spread rapidly. Teachers and professors will use it in their classes because they know anyone can run it on any machine, easily and for free. The resulting, enlarged community will then make it much easier to build out and maintain, since that will solve the labor problem. So, head on over to docs.scipy.org, or contact the packaging team, and dive in. (And if it isn't obvious, the same invitation goes to all our lurkers. Pitching in is easy, and very rewarding!) Welcome again, --jh--

2 1

Response to ticket #872: overwrite_a has no effect for cholesky
by Stéfan van der Walt 22 Feb '09

22 Feb '09

Hi all, I think ticket #872 may have uncovered a larger problem, in that none of the scipy.linalg functions support overwrite_a. Maybe I'm mistaken; could anyone shed some light? Thanks Stéfan

2 1

Another request to edit documentation...
by Luis Saavedra 22 Feb '09

22 Feb '09

Hi list, I'm learning to use the C-API for numpy and I had some problems... but I have already autosolved and I'd like to share experience of working on the documentation, regards, Luis

2 3

New test failures scipy 0.8.0.dev5585
by Nils Wagner 22 Feb '09

22 Feb '09

Hi all, I see some new test failures beside the special function issues (python2.4, scipy 0.8.0.dev5585, numpy 1.3.0.dev6450) ====================================================================== FAIL: Test generator for parametric tests ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.4/site-packages/nose-0.10.3-py2.4.egg/nose/case.py", line 182, in runTest self.test(*self.arg) File "/usr/lib/python2.4/site-packages/scipy/misc/tests/test_pilutil.py", line 35, in tst_fromimage assert img.min() >= imin AssertionError ====================================================================== FAIL: Test generator for parametric tests ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.4/site-packages/nose-0.10.3-py2.4.egg/nose/case.py", line 182, in runTest self.test(*self.arg) File "/usr/lib/python2.4/site-packages/scipy/misc/tests/test_pilutil.py", line 35, in tst_fromimage assert img.min() >= imin AssertionError ====================================================================== FAIL: test_morestats.test_fligner ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.4/site-packages/nose-0.10.3-py2.4.egg/nose/case.py", line 182, in runTest self.test(*self.arg) File "/usr/lib/python2.4/site-packages/scipy/stats/tests/test_morestats.py", line 117, in test_fligner (3.2282229927203558, 0.072379187848207877)) File "/usr/lib/python2.4/site-packages/numpy/testing/utils.py", line 303, in assert_array_equal verbose=verbose, header='Arrays are not equal') File "/usr/lib/python2.4/site-packages/numpy/testing/utils.py", line 295, in assert_array_compare raise AssertionError(msg) AssertionError: Arrays are not equal (mismatch 100.0%) x: array([ 3.228, 0.072]) y: array([ 3.228, 0.072]) Nils

2 1

Request to help edit documentation
by Daniel Jensen 22 Feb '09

22 Feb '09

I'm just following the instructions on the page: http://docs.scipy.org/numpy/Front%20Page/ for requesting rights to help edit documentation. I've found a few simple spelling errors in the numpy documentation and thought that I could help contribute a little. Just an example, the following page: http://docs.scipy.org/doc/numpy/reference/generated/numpy.array_str.html#nu… is missing the 'f' in 'floating point'. -Daniel

2 1