Mailman 3 December 2014 - SciPy-Dev

Adding routines for updating QR decompositions
by Eric Moore April 6, 2015

April 6, 2015

Hi, I've just submitted PR #4249 adding qr_update, qr_insert and qr_delete for performing rank-k updates, adding rows or columns and removing rows or colums. https://github.com/scipy/scipy/pull/4249 Comments welcome. Eric.

5 6

removing lib.blas/lapack
by Ralf Gommers Dec. 31, 2014

Dec. 31, 2014

Hi, The following have been deprecated since scipy 0.12.0: - scipy.lib.blas - scipy.lib.lapack - scipy.linalg.blas.cblas - scipy.linalg.blas.fblas - scipy.linalg.lapack.clapack - scipy.linalg.lapack.flapack See http://docs.scipy.org/doc/scipy/reference/release.0.12.0.html#deprecated-fe… It looks to me like it's time to remove those for 0.16.0. Then scipy.lib doesn't contain anything public anymore and will then be renamed scipy._lib. Any objections? Ralf

3 3

ANN: Scipy 0.14.1 release
by Pauli Virtanen Dec. 30, 2014

Dec. 30, 2014

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Dear all, We are pleased to announce the Scipy 0.14.1 release. The 0.14.1 release is a bugfix-only release, addressing the following issues: - - gh-3630 NetCDF reading results in a segfault - - gh-3631 SuperLU object not working as expected for complex matrices - - gh-3733 Segfault from map_coordinates - - gh-3780 Segfault when using CSR/CSC matrix and uint32/uint64 - - gh-3781 Fix omitted types in sparsetools typemaps - - gh-3802 0.14.0 API breakage: _gen generators are missing from scipy.stats.distributions API - - gh-3805 Ndimge test failures with numpy 1.10 - - gh-3812 == sometimes wrong on csr_matrix - - gh-3853 Many scipy.sparse test errors/failures with numpy 1.9.0b2 - - gh-4084 Fix exception declarations for Cython 0.21.1 compatibility - - gh-4093 Avoid a memory error in splev(x, tck, der=k) - - gh-4104 Workaround SGEMV segfault in Accelerate (maintenance 0.14.x) - - gh-4143 Fix ndimage functions for large data - - gh-4149 Bug in expm for integer arrays - - gh-4154 Ensure that the 'size' argument of PIL's 'resize' method is a tuple - - gh-4163 ZeroDivisionError in scipy.sparse.linalg.lsqr - - gh-4164 Remove use of deprecated numpy API in lib/lapack/ f2py wrapper - - gh-4180 PIL resize support tuple fix - - gh-4168 Address arpack test failures on windows 32 bits with numpy 1.9.1 - - gh-4203 Sparse matrix multiplication in 0.14.x slower compared to 0.13.x - - gh-4218 Make ndimage interpolation compatible with numpy relaxed strides - - gh-4225 Off-by-one error in PPoly shape checks - - gh-4248 Fix issue with incorrect use of closure for slsqp Source tarballs and binaries are available at https://sourceforge.net/projects/scipy/files/SciPy/0.14.1/ Best regards, Pauli Virtanen -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iEYEARECAAYFAlSjCcYACgkQ6BQxb7O0pWBxcwCfcnd4uva5hzMHQlHmWxlfbja3 T0AAn2QQmhcotDRB2c2p41Xzjb4MJ13f =yBxH -----END PGP SIGNATURE-----

2 1

Possibly a bug in scipy.spatial.distance.mahalanobis
by Simen Langseth Dec. 29, 2014

Dec. 29, 2014

Dear Scipy Developers: The documentation below mentions the following for the formula of mahalanobis distance: [image: \sqrt{ (u-v) V^{-1} (u-v)^T }] http://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.spatial.di… However, in the following source code, the transpose of the delta seems to be missing. u = _validate_vector(u) v = _validate_vector(v)VI = np.atleast_2d(VI)delta = u - vm = np.dot(np.dot(delta, VI), delta)return np.sqrt(m) https://github.com/scipy/scipy/blob/6a7327e8bb8248b2ea165180bc602edf1ab33dd… Please forgive me if I could not understand the source code perfectly. Thanks. Simen

3 3

welcome Lars Buitinck and Matthew Brett to the Scipy team
by Ralf Gommers Dec. 28, 2014

Dec. 28, 2014

Hi all, On behalf of the Scipy developers I'd like to welcome Lars Buitinck and Matthew Brett as members of the core dev team. Both Lars and Matthew have been important contributors for a long time, this just makes it "official" (i.e. they now have commit rights). Matthew's first patch that I can find is from 2006; currently he is the maintainer of scipy.io.matlab and builds OS X wheels for our releases. Lars has been making improvements and fixes all over the code base for quite a while - 55 PRs and counting. I'm looking forward to their continued contributions! Cheers, Ralf

2 1

Wallenius' noncentral hypergeometric distribution
by Adam Stone Dec. 28, 2014

Dec. 28, 2014

Hello, I recently wrote some python code that involved calculating Wallenius' noncentral hypergeometric distribution ( http://en.wikipedia.org/wiki/Wallenius%27_noncentral_hypergeometric_distrib… ). This is similar to the scipy.stats.hypergeom distribution, which is applied to situations like drawing marbles from an urn without replacement, but the noncentral version introduces unequally-weighted probabilities for drawing the different types of marbles. What I have is a multivariate implementation of this (i.e. drawing from 3+ types of marbles). I thought that it might be a good addition to the scipy.stats module. My initial implementation is here: https://github.com/AdamStone/hypergeometric Obviously the above isn't yet in a form consistent with the other scipy.stats distributions, but I wanted to check with the mailing list before committing the time to restructure it. If the list approves, I will go ahead with implementing it in the manner of the other scipy.stats distributions. Thanks, Adam Stone

2 1

Fast sizes for FFT
by Robert McGibbon Dec. 24, 2014

Dec. 24, 2014

Hey, The performance of fftpack depends very strongly on the array size -- sizes that are powers of two are good, but also powers of three, five and seven, or numbers whose only prime factors are from (2,3,5,7). For problems that can use padding, rounding up the size (using np.fft.fft(x, n=size_with_padding)) to one of these multiples makes a big difference. Some other packages expose a function for calculating the next fast size, e.g: http://ltfat.sourceforge.net/notes/ltfatnote017.pdf. Is there anything like this in numpy/scipy? If not, would this be a reasonable feature to add? -Robert

3 2

Efficient operations on sparse matrices
by Alex Susu Dec. 23, 2014

Dec. 23, 2014

Hello. I am very interested in implementing efficient operations on sparse matrices from scipy, such as: interp2, imresize, imfilter, mean2, gradient - all these functions are actually the ones from Matlab, used in computer vision (you can google to see exactly what they do). It seems it's not very difficult - I used also some information from an earlier email on this list (see below). Did anybody else try something similar with the sparse matrices in scipy? Thank you, Alex > On Fri, Oct 10, 2014 at 12:00 PM, Saullo Castro <saullogiovani <at> gmail.com> wrote: > I developed a function (shown below) to check if a sparse matrix is > symmetric and would like to know if the community is interested to include > in scipy.sparse. > > Regards, > Saullo > > def is_symmetric(m): > """Check if a sparse matrix is symmetric > """ > if m.shape[0] != m.shape[1]: > raise ValueError('m must be a square matrix') > > if not isinstance(m, coo_matrix): > m = coo_matrix(m) > > r, c, v = m.row, m.col, m.data > tril_no_diag = r > c > triu_no_diag = c > r > > if triu_no_diag.sum() != tril_no_diag.sum(): > return False > > rl = r[tril_no_diag] > cl = c[tril_no_diag] > vl = v[tril_no_diag] > ru = r[triu_no_diag] > cu = c[triu_no_diag] > vu = v[triu_no_diag] > > sortl = np.lexsort((cl, rl)) > sortu = np.lexsort((ru, cu)) > vl = vl[sortl] > vu = vu[sortu] > > check = np.allclose(vl, vu) > > return check

2 1

(no subject)
by Clint Valentine Dec. 23, 2014

Dec. 23, 2014

I ported an algorithm originally written in MatLab to Python for generating a random contingency table (when one exists) for a statistical test I am writing. I know something like this doesn't exist in scipy.stats.contingency and was wondering if it would be a good addition? I haven't developed any tests for the function yet but can if needed. I have just started developing in Python as an undergraduate student studying computational biology. This would be my first time contributing to open source and I am looking forward to the the experience and all the advice you are willing to offer. Thanks for your time. -Clint def rcont(ncol, nrow, ncolt, nrowt): """ Return a random contingency table based on the two-dimensional shape and marginal totals of a given contingency table There must be 1 or many tables with nonnegative, integral entries that have the given shape and marginal sums as input. The function will generate, at random, one of the tables and return it. Repeated calls to the function will return new random selections. Parameters: ----------- ncol : integer The vector length of the columns in the contingency table nrow : integer The vector length of the rows in the contingency table ncolt : array_like The vector of the marginal column totals in the contingency table nrowt : array_like The vector of the marginal row totals in the contingency table Returns ------- imatrix : ndarray of integers The randomly generated two-dimensional ndarray preserving the same shape and marginal totals provided if exists Reference --------- Algorithm AS144: Appl. Statist. (1979) v.28, pp.329-332 Licensing: GNU LGPL license Author: Original FORTRAN77 version by James Boyett. MATLAB version by John Burkardt. Python version by Clint Valentine Reference: James Boyett, Algorithm AS 144: Random R x C Tables with Given Row and Column Totals, Applied Statistics, Volume 28, Number 3, pages 329-332, 1979. """ # Initialize and permute vector nvect = np.array(xrange(sum(ncolt)), dtype=np.int_) np.random.shuffle(nvect) # Construct random matrix and partial column sums imatrix = np.zeros([nrow, ncol], dtype=np.int_) nsubt = np.cumsum(ncolt) # Generate random RxC contingency tables preserving shape and # marginal column and row sums and total sum. count = 0 for i in xrange(nrow): limit = nrowt[i] for _ in xrange(limit): for j in xrange(ncol): if nvect[count] <= nsubt[j]: count += 1 imatrix[i][j] += 1 break return imatrix

3 2

ANN: Scipy 0.14.1 release candidate 1
by Pauli Virtanen Dec. 19, 2014

Dec. 19, 2014

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Dear all, We have finished preparing the Scipy 0.14.1 release candidate 1. If no regressions turn up, the final release is planned within the following weeks. The 0.14.1 release will be a bugfix-only release, addressing the following issues: - - gh-3630 NetCDF reading results in a segfault - - gh-3631 SuperLU object not working as expected for complex matrices - - gh-3733 Segfault from map_coordinates - - gh-3780 Segfault when using CSR/CSC matrix and uint32/uint64 - - gh-3781 Fix omitted types in sparsetools typemaps - - gh-3802 0.14.0 API breakage: _gen generators are missing from scipy.stats.distributions API - - gh-3805 ndimge test failures with numpy 1.10 - - gh-3812 == sometimes wrong on csr_matrix - - gh-3853 Many scipy.sparse test errors/failures with numpy 1.9.0b2 - - gh-4084 Fix exception declarations for Cython 0.21.1 compatibility - - gh-4093 Avoid a memory error in splev(x, tck, der=k) - - gh-4104 Workaround SGEMV segfault in Accelerate (maintenance 0.14.x) - - gh-4143 Fix ndimage functions for large data - - gh-4149 Bug in expm for integer arrays - - gh-4154 Ensure that the 'size' argument of PIL's 'resize' method is a tuple - - gh-4163 ZeroDivisionError in scipy.sparse.linalg.lsqr - - gh-4164 Remove use of deprecated numpy API in lib/lapack/ f2py wrapper - - gh-4180 pil resize support tuple fix - - gh-4168 Address arpack test failures on windows 32 bits with numpy 1.9.1 - - gh-4218 make ndimage interpolation compatible with numpy relaxed strides - - gh-4225 off-by-one error in PPoly shape checks - - gh-4248 fix issue with incorrect use of closure for slsqp Source tarballs and binaries are available at https://sourceforge.net/projects/scipy/files/scipy/0.14.1rc1/ Best regards, Pauli Virtanen -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iEYEARECAAYFAlSOD0YACgkQ6BQxb7O0pWDO5ACfccLqMvZWfkHqSzDCkMSoRKAU n7cAni6XhWJRy7oJ757rlGeIi0e34HTn =9bB/ -----END PGP SIGNATURE-----

2 1