Mailman 3 January 2012 - SciPy-User

AUTO: Christopher Peters is out of the office (returning 01/30/2012)
by Christopher Peters Jan. 26, 2012

Jan. 26, 2012

I am out of the office until 01/30/2012. I am out of the office. Please email urgent requests to Mike McDonald. Note: This is an automated response to your message "[SciPy-User] Stack corruption in extension module." sent on 1/26/2012 12:46:34 AM. This is the only notification you will receive while this person is away.

1 0

Stack corruption in extension module.
by J. David Lee Jan. 26, 2012

Jan. 26, 2012

Hi, Sorry that this is a bit off-topic, but I'm having an interesting problem with a C extension module. Here is a simple test module, stack_test.c: #define PY_ARRAY_UNIQUE_SYMBOL __np_inline_stack_test #include<Python.h> #include<numpy/arrayobject.h> #include<math.h> // Forward declarations of our function. static PyObject *function(PyObject *self, PyObject *args); // Boilerplate: function list. static PyMethodDef methods[] = { { "function", function, METH_VARARGS, "Doc string."}, { NULL, NULL, 0, NULL } /* Sentinel */ }; // Boilerplate: Module initialization. PyMODINIT_FUNC initstack_test(void) { (void) Py_InitModule("stack_test", methods); import_array(); } static void calc_A(npy_int64 i, npy_int64 P, npy_int64 Q, npy_float64 *R, npy_float64 *A , npy_float64 b) { printf("Start : i: %li P: %f Q: %f R: %f A: %f b: %f\n", i, P, Q, *R, *A, b); *R = 6.0; *A = 7.0; printf("End : i: %li P: %f Q: %f R: %f A: %f b: %f\n", i, P, Q, *R, *A, b); } static PyObject *function(PyObject *self, PyObject *args) { long th; long return_val = 34; if (!PyArg_ParseTuple(args, "l",&th)) { return NULL; } npy_float64 P=0; npy_float64 Q=0; npy_float64 R=0; npy_float64 A=0; npy_float64 b=0; npy_int64 i = 0; P = 1; Q = 2; R = 3; A = 4; b = 5; printf("Before: i: %li P: %f Q: %f R: %f A: %f b: %f\n", i, P, Q, R, A, b); calc_A(i, P, Q,&R,&A, b); printf("After : i: %li P: %f Q: %f R: %f A: %f b: %f\n", i, P, Q, R, A, b); return PyLong_FromLong(return_val); } I compile with: gcc -pthread -fno-strict-aliasing -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -fPIC -UNDEBUG -I/usr/lib/pymodules/python2.7/numpy/core/include -I/usr/include/python2.7 -c stack_test.c -o stack_test.o gcc -pthread -shared -Wl,-O1 -Wl,-Bsymbolic-function stack_test.o stack_test.so Now when I run the function: stack_test.function(1) Before: i: 0 P: 1.000000 Q: 2.000000 R: 3.000000 A: 4.000000 b: 5.000000 Start : i: 0 P: 3.000000 Q: 4.000000 R: 5.000000 A: 4.000000 b: 5.000000 End : i: 0 P: 6.000000 Q: 7.000000 R: 5.000000 A: 4.000000 b: 5.000000 After : i: 0 P: 1.000000 Q: 2.000000 R: 6.000000 A: 7.000000 b: 5.000000 which is clearly wrong. I'm using python 2.7.2 built with gcc 4.6.2. If you have any suggestions, I'd love to hear them. Thanks, David PS The numpy stuff is in there because this used to be doing some array calculations. I cut it down to a minimum that still had bad behavior.

1 1

Course "Python for Scientists and Engineers" in Chicago
by Mike Müller Jan. 24, 2012

Jan. 24, 2012

Course "Python for Scientists and Engineers" in Chicago ======================================================= There will be a comprehensive Python course for scientists and engineers in Chicago end of February / beginning of March 2012. It consists of a 3-day intro and a 2-day advanced section. Both sections can be taken separately or combined. More details below and here: http://www.dabeaz.com/chicago/science.html Please let friends or colleagues who might be interested in such a course know about it. 3-Day Intro Section ------------------- - Overview of Scientific and Technical Libraries for Python. - Numerical Calculations with NumPy - Storage and Processing of Large Amounts of Data - Graphical Presentation of Scientific Data with matplotlib - Object Oriented Programming for Scientific and Technical Projects - Open Time for Problem Solving 2-Day Advanced Section ---------------------- - Extending Python with Other Languages - Unit Testing - Version Control with Mercurial The Details ----------- The course is hosted by David Beazley (http://www.dabeaz.com). Date: Feb 27 - Mar 2, 2012 Location: Chicago, IL, USA Trainer: Mike Müller Course Language: English Link: http://www.dabeaz.com/chicago/science.html

1 0

special.kv goes from 0 to nan for large numbers
by Johann Cohen-Tanugi Jan. 24, 2012

Jan. 24, 2012

Hi there, I have scipy 0.10 vanilla from the packages distributed with Ubuntu 11.10, and I see the following : In [1]: import scipy In [2]: scipy.__version__ Out[2]: '0.10.0' In [3]: import scipy.special as sp In [4]: sp.kv(1./3.,1.08e+09) Out[4]: nan In [5]: sp.kv(1./3.,1.07e+09) Out[5]: 0.0 I guess this has to dow with a cephes and big numbers.... Is it fixed on github already, or is it a known bug? Or is it a feature and not a bug? My machine is 64bit BTW. Best, johann

3 3

Re: [SciPy-User] save/load scipy sparse csr_matrix in a portable data format
by Dinesh B Vadhia Jan. 23, 2012

Jan. 23, 2012

Ah, thank-you David. Didn't see any doc mentioning re-creating the sparse matrix. Brill! -------------------------------------------------------------------------------- Date: Mon, 23 Jan 2012 08:14:22 -0500 From: David Warde-Farley <wardefar(a)iro.umontreal.ca> Subject: Re: [SciPy-User] save/load scipy sparse csr_matrix in a portable data format To: SciPy Users List <scipy-user(a)scipy.org> Message-ID: <B5751591-4FA4-4F53-8D6C-7F96BAF9856B(a)iro.umontreal.ca> Content-Type: text/plain; charset=us-ascii On 2012-01-21, at 1:34 PM, Dinesh B Vadhia wrote: > How do you save/load a scipy sparse csr_matrix in a portable format? The sparse matrix is created in Python 3.2.2 (Windows 64-bit) to run on Python 2.7.2 (Linux 64-bit) . I've tried numpy.save and numpy.load as well as scipy.io.mmwrite() and scipy.io.mmread() and none of the methods work. A csr_matrix really has only 3 data attributes that matter: .data, .indices, and .indptr. They are all simple ndarrays, so numpy.save will work on them. Save those three arrays with numpy.save or numpy.savez, load them back with numpy.load, and then recreate the sparse matrix object with new_csr = csr_matrix((data, indices, indptr), shape=(M, N)) Could be smoother but it gets the job done. David

2 1

save/load scipy sparse csr_matrix in a portable data format
by Dinesh B Vadhia Jan. 23, 2012

Jan. 23, 2012

How do you save/load a scipy sparse csr_matrix in a portable format? The sparse matrix is created in Python 3.2.2 (Windows 64-bit) to run on Python 2.7.2 (Linux 64-bit) . I've tried numpy.save and numpy.load as well as scipy.io.mmwrite() and scipy.io.mmread() and none of the methods work.

2 1

How to calculate Yulewalk with scipy.optimize.leastsq
by klo uo Jan. 22, 2012

Jan. 22, 2012

H(z) = yulewalk(N,frq,mag) - finds the N-th order iir filter: [image: %5Cnormalsize%5C%21H%28z%29%3D%5Cfrac%7BB%28z%29%7D%7BA%28z%29%7D%3D%5Cfrac%7Bb%281%29%20%2B%20b%282%29z%5E%7B-1%7D%2B...%2Bb%28n%29z%5E%7B-%28n-1%29%7D%7D%7B1%2Ba%281%29z%5E%7B-1%7D%2B...%2Ba%28n%29z%5E%7B-%28n-1%29%7D%7D.gif] H(z) : filter B(z)/A(z) N : integer (order of desired filter) frq : real row vector (non-decreasing order), frequencies. mag : non negative real row vector (same size as frq), desired magnitudes. ---------------------------------------- yulewalk() function in "/scipy/signal/filter_design.py" is empty. It perhaps can be calculated using scipy.optimize.leastsq() but I lack skills to figure out how? All I can do right now is port it quick and dirty from Matlab to Python, but I feel it's bad idea and waste of time Can someone provide tip how can I calculate 10-th order yulewalk with scipy?

7 12

Sparse jacobian in scipy ODE solver
by Per Nielsen Jan. 22, 2012

Jan. 22, 2012

Hi all I am working with large linear system 1st order ODE systems of the form dydt = M * y where M is very sparse complex valued matrix. I currently use the ode class from scipy.integrator like: from scipy.integrate import ode r = ode(f) r.set_integrator('zvode', method='adams') and calculate f = dydt as M.dot(y). I would like to make use of the jacobian of dydt, which for my linear problem is simply M, to speed up my simulations. However, when I supply a sparse matrix as the jacobian I get the following error (see attached script): rv_cb_arr is NULL Call-back cb_jac_in_zvode__user__routines failed. --------------------------------------------------------------------------- TypeError Traceback (most recent call last) /Users/per/Dropbox/DTU/python/tryouts/ode_jac_test.py in <module>() 61 62 while r.successful() and r.t < t1: ---> 63 r.integrate(r.t + dt) 64 ts.append(r.t) 65 ys.append(r.y) /Library/Frameworks/EPD64.framework/Versions/7.1/lib/python2.7/site-packages/scipy/integrate/ode.py in integrate(self, t, step, relax) 324 self.y,self.t = mth(self.f,self.jac or (lambda :None), 325 self.y,self.t,t, --> 326 self.f_params,self.jac_params) 327 return self.y 328 /Library/Frameworks/EPD64.framework/Versions/7.1/lib/python2.7/site-packages/scipy/integrate/ode.py in run(self, *args) 684 685 def run(self,*args): --> 686 y1,t,istate = self.runner(*(args[:5]+tuple(self.call_args)+args[5:])) 687 if istate < 0: 688 warnings.warn('zvode: ' + TypeError: a float is required It seems that the jacobian can not be sparse. But is this a fundamental problem or could one somehow work around it and still use ode from scipy? Any help or comments would be appreciated. Best, Per

2 3

ANN: SOLVCON 0.1.1
by Yung-Yu Chen Jan. 21, 2012

Jan. 21, 2012

Hello, I am pleased to announce version 0.1.1 of SOLVCON. SOLVCON is a Python-based, multi-physics software framework for solving first-order hyperbolic PDEs. The source tarball can be downloaded at http://bitbucket.org/yungyuc/solvcon/downloads . More information can be found at http://solvcon.net/ . This release adds a loader of Gmsh mesh format and fixes several bugs. New features: - Add a loader for Gmsh ASCII mesh format. The loader locates in solvcon.io.gmsh and is implemented as pure Python code. ``scg mesh`` command line tool can recognize the format. Issue #52. - Revamp the dependency building system to support older OSes and proxies that need authentication. Issue #53. - Extract the SCons commands for building the Epydoc and Sphinx document from SConstruct into standalone SCons tools. Two new tools are added in the directory ``site_scons/site_tools/``: ``sphinx.py`` and ``scons_epydoc.py``. Note that the SCons tool for Epydoc cannot be named as ``epydoc.py`` or the name collides with the real ``epydoc`` package. - Add Gmsh and Sphinx into ground/. Bug-fix: - Issue #49: "No Vtk for final time step". Output timing of CollectHook and MarchSave. - Issue #54: "Shared objects are not found under Mac OS X". Thank you Nathan. - Issue #38: "soln/dsoln shouldn't be hard-coded". -- Yung-Yu Chen http://solvcon.net/yyc/ +886 (99) 129 4763

1 0

matrix multiplication
by Chris Lirakis Jan. 21, 2012

Jan. 21, 2012

I have tried the following and they all yield the same result. A = zeros(3,1) B = A * A.T B = [[ 0 0 0] [0 0 0] [0 0 0]] A = matrix(zeros(3,1) B = A * A.T B = [[ 0 0 0] [0 0 0] [0 0 0]] I would have thought that the result should have been [[0]] if I do the following: A = matrix([[0 0 0]]) Then A * A.T yields [[0]] Can someone tell me why and how I might fix this? Chris -- Chris Lirakis

3 2