SciPy 0.14.0 Release Notes

SciPy 0.14.0 is the culmination of 8 months of hard work. It contains many new features, numerous bug-fixes, improved test coverage and better documentation. There have been a number of deprecations and API changes in this release, which are documented below. All users are encouraged to upgrade to this release, as there are a large number of bug-fixes and optimizations. Moreover, our development attention will now shift to bug-fix releases on the 0.14.x branch, and on adding new features on the master branch.

This release requires Python 2.6, 2.7 or 3.2-3.4 and NumPy 1.5.1 or greater.

New features

scipy.interpolate improvements

A new wrapper function scipy.interpolate.interpn for interpolation on regular grids has been added. interpn supports linear and nearest-neighbor interpolation in arbitrary dimensions and spline interpolation in two dimensions.

Faster implementations of piecewise polynomials in power and Bernstein polynomial bases have been added as scipy.interpolate.PPoly and scipy.interpolate.BPoly. New users should use these in favor of scipy.interpolate.PiecewisePolynomial.

scipy.interpolate.interp1d now accepts non-monotonic inputs and sorts them. If performance is critical, sorting can be turned off by using the new assume_sorted keyword.

Functionality for evaluation of bivariate spline derivatives in scipy.interpolate has been added.

The new class scipy.interpolate.Akima1DInterpolator implements the piecewise cubic polynomial interpolation scheme devised by H. Akima.

Functionality for fast interpolation on regular, unevenly spaced grids in arbitrary dimensions has been added as scipy.interpolate.RegularGridInterpolator .

scipy.linalg improvements

The new function scipy.linalg.dft computes the matrix of the discrete Fourier transform.

A condition number estimation function for matrix exponential, scipy.linalg.expm_cond, has been added.

scipy.optimize improvements

A set of benchmarks for optimize, which can be run with optimize.bench(), has been added.

scipy.optimize.curve_fit now has more controllable error estimation via the absolute_sigma keyword.

Support for passing custom minimization methods to optimize.minimize() and optimize.minimize_scalar() has been added, currently useful especially for combining optimize.basinhopping() with custom local optimizer routines.

scipy.stats improvements

A new class scipy.stats.multivariate_normal with functionality for multivariate normal random variables has been added.

A lot of work on the scipy.stats distribution framework has been done. Moment calculations (skew and kurtosis mainly) are fixed and verified, all examples are now runnable, and many small accuracy and performance improvements for individual distributions were merged.

The new function scipy.stats.anderson_ksamp computes the k-sample Anderson-Darling test for the null hypothesis that k samples come from the same parent population.

scipy.signal improvements

scipy.signal.iirfilter and related functions to design Butterworth, Chebyshev, elliptical and Bessel IIR filters now all use pole-zero (“zpk”) format internally instead of using transformations to numerator/denominator format. The accuracy of the produced filters, especially high-order ones, is improved significantly as a result.

The Savitzky-Golay filter was added with the new functions scipy.signal.savgol_filter and scipy.signal.savgol_coeffs.

The new function scipy.signal.vectorstrength computes the vector strength, a measure of phase synchrony, of a set of events.

scipy.special improvements

The functions scipy.special.boxcox and scipy.special.boxcox1p, which compute the Box-Cox transformation, have been added.

scipy.sparse improvements

  • Significant performance improvement in CSR, CSC, and DOK indexing speed.
  • When using Numpy >= 1.9 (to be released in MM 2014), sparse matrices function correctly when given to arguments of np.dot, np.multiply and other ufuncs. With earlier Numpy and Scipy versions, the results of such operations are undefined and usually unexpected.
  • Sparse matrices are no longer limited to 2^31 nonzero elements. They automatically switch to using 64-bit index data type for matrices containing more elements. User code written assuming the sparse matrices use int32 as the index data type will continue to work, except for such large matrices. Code dealing with larger matrices needs to accept either int32 or int64 indices.

Deprecated features

anneal

The global minimization function scipy.optimize.anneal is deprecated. All users should use the scipy.optimize.basinhopping function instead.

scipy.stats

randwcdf and randwppf functions are deprecated. All users should use distribution-specific rvs methods instead.

Probability calculation aliases zprob, fprob and ksprob are deprecated. Use instead the sf methods of the corresponding distributions or the special functions directly.

scipy.interpolate

PiecewisePolynomial class is deprecated.

Backwards incompatible changes

scipy.special.lpmn

lpmn no longer accepts complex-valued arguments. A new function clpmn with uniform complex analytic behavior has been added, and it should be used instead.

scipy.sparse.linalg

Eigenvectors in the case of generalized eigenvalue problem are normalized to unit vectors in 2-norm, rather than following the LAPACK normalization convention.

The deprecated UMFPACK wrapper in scipy.sparse.linalg has been removed due to license and install issues. If available, scikits.umfpack is still used transparently in the spsolve and factorized functions. Otherwise, SuperLU is used instead in these functions.

scipy.stats

The deprecated functions glm, oneway and cmedian have been removed from scipy.stats.

stats.scoreatpercentile now returns an array instead of a list of percentiles.

scipy.interpolate

The API for computing derivatives of a monotone piecewise interpolation has changed: if p is a PchipInterpolator object, p.derivative(der) returns a callable object representing the derivative of p. For in-place derivatives use the second argument of the __call__ method: p(0.1, der=2) evaluates the second derivative of p at x=0.1.

The method p.derivatives has been removed.

Authors

  • Marc Abramowitz +
  • Anders Bech Borchersen +
  • Vincent Arel-Bundock +
  • Petr Baudis +
  • Max Bolingbroke
  • François Boulogne
  • Matthew Brett
  • Lars Buitinck
  • Evgeni Burovski
  • CJ Carey +
  • Thomas A Caswell +
  • Pawel Chojnacki +
  • Phillip Cloud +
  • Stefano Costa +
  • David Cournapeau
  • David Menendez Hurtado +
  • Matthieu Dartiailh +
  • Christoph Deil +
  • Jörg Dietrich +
  • endolith
  • Francisco de la Peña +
  • Ben FrantzDale +
  • Jim Garrison +
  • André Gaul
  • Christoph Gohlke
  • Ralf Gommers
  • Robert David Grant
  • Alex Griffing
  • Blake Griffith
  • Yaroslav Halchenko
  • Andreas Hilboll
  • Kat Huang
  • Gert-Ludwig Ingold
  • James T. Webber +
  • Dorota Jarecka +
  • Todd Jennings +
  • Thouis (Ray) Jones
  • Juan Luis Cano Rodríguez
  • ktritz +
  • Jacques Kvam +
  • Eric Larson +
  • Justin Lavoie +
  • Denis Laxalde
  • Jussi Leinonen +
  • lemonlaug +
  • Tim Leslie
  • Alain Leufroy +
  • George Lewis +
  • Max Linke +
  • Brandon Liu +
  • Benny Malengier +
  • Matthias Kümmerer +
  • Cimarron Mittelsteadt +
  • Eric Moore
  • Andrew Nelson +
  • Niklas Hambüchen +
  • Joel Nothman +
  • Clemens Novak
  • Emanuele Olivetti +
  • Stefan Otte +
  • peb +
  • Josef Perktold
  • pjwerneck
  • poolio
  • Jérôme Roy +
  • Carl Sandrock +
  • Andrew Sczesnak +
  • Shauna +
  • Fabrice Silva
  • Daniel B. Smith
  • Patrick Snape +
  • Thomas Spura +
  • Jacob Stevenson
  • Julian Taylor
  • Tomas Tomecek
  • Richard Tsai
  • Jacob Vanderplas
  • Joris Vankerschaver +
  • Pauli Virtanen
  • Warren Weckesser

A total of 80 people contributed to this release. People with a “+” by their names contributed a patch for the first time. This list of names is automatically generated, and may not be fully complete.

Issues closed

  • #1325: add custom axis keyword to dendrogram function in scipy.cluster.hierarchy...
  • #1437: Wrong pochhammer symbol for negative integers (Trac #910)
  • #1555: scipy.io.netcdf leaks file descriptors (Trac #1028)
  • #1569: sparse matrix failed with element-wise multiplication using numpy.multiply()...
  • #1833: Sparse matrices are limited to 2^32 non-zero elements (Trac #1307)
  • #1834: scipy.linalg.eig does not normalize eigenvector if B is given...
  • #1866: stats for invgamma (Trac #1340)
  • #1886: stats.zipf floating point warnings (Trac #1361)
  • #1887: Stats continuous distributions - floating point warnings (Trac...
  • #1897: scoreatpercentile() does not handle empty list inputs (Trac #1372)
  • #1918: splint returns incorrect results (Trac #1393)
  • #1949: kurtosistest fails in mstats with type error (Trac #1424)
  • #2092: scipy.test leaves darwin27compiled_catalog, cpp and so files...
  • #2106: stats ENH: shape parameters in distribution docstrings (Trac...
  • #2123: Bad behavior of sparse matrices in a binary ufunc (Trac #1598)
  • #2152: Fix mmio/fromfile on gzip on Python 3 (Trac #1627)
  • #2164: stats.rice.pdf(x, 0) returns nan (Trac #1639)
  • #2169: scipy.optimize.fmin_bfgs not handling functions with boundaries...
  • #2177: scipy.cluster.hierarchy.ClusterNode.pre_order returns IndexError...
  • #2179: coo.todense() segfaults (Trac #1654)
  • #2185: Precision of scipy.ndimage.gaussian_filter*() limited (Trac #1660)
  • #2186: scipy.stats.mstats.kurtosistest crashes on 1d input (Trac #1661)
  • #2238: Negative p-value on hypergeom.cdf (Trac #1719)
  • #2283: ascending order in interpolation routines (Trac #1764)
  • #2288: mstats.kurtosistest is incorrectly converting to float, and fails...
  • #2396: lpmn wrong results for |z| > 1 (Trac #1877)
  • #2398: ss2tf returns num as 2D array instead of 1D (Trac #1879)
  • #2406: linkage does not take Unicode strings as method names (Trac #1887)
  • #2443: IIR filter design should not transform to tf representation internally
  • #2572: class method solve of splu return object corrupted or falsely...
  • #2667: stats endless loop ?
  • #2671: .stats.hypergeom documentation error in the note about pmf
  • #2691: BUG scipy.linalg.lapack: potrf/ptroi interpret their ‘lower’...
  • #2721: Allow use of ellipsis in scipy.sparse slicing
  • #2741: stats: deprecate and remove alias for special functions
  • #2742: stats add rvs to rice distribution
  • #2765: bugs stats entropy
  • #2832: argrelextrema returns tuple of 2 empty arrays when no peaks found...
  • #2861: scipy.stats.scoreatpercentile broken for vector per
  • #2891: COBYLA successful termination when constraints violated
  • #2919: test failure with the current master
  • #2922: ndimage.percentile_filter ignores origin argument for multidimensional...
  • #2938: Sparse/dense matrix inplace operations fail due to __numpy_ufunc__
  • #2944: MacPorts builds yield 40Mb worth of build warnings
  • #2945: FAIL: test_random_complex (test_basic.TestDet)
  • #2947: FAIL: Test some trivial edge cases for savgol_filter()
  • #2953: Scipy Delaunay triangulation is not oriented
  • #2971: scipy.stats.mstats.winsorize documentation error
  • #2980: Problems running what seems a perfectly valid example
  • #2996: entropy for rv_discrete is incorrect?!
  • #2998: Fix numpy version comparisons
  • #3002: python setup.py install fails
  • #3014: Bug in stats.fisher_exact
  • #3030: relative entropy using scipy.stats.distribution.entropy when...
  • #3037: scipy.optimize.curve_fit leads to unexpected behavior when input...
  • #3047: mstats.ttest_rel axis=None, requires masked array
  • #3059: BUG: Slices of sparse matrices return incorrect dtype
  • #3063: range keyword in binned_statistics incorrect
  • #3067: cumtrapz not working as expected
  • #3069: sinc
  • #3086: standard error calculation inconsistent between ‘stats’ and ‘mstats’
  • #3094: Add a perm function into scipy.misc and an enhancement of...
  • #3111: scipy.sparse.[hv]stack don’t respect anymore the dtype parameter
  • #3172: optimize.curve_fit uses different nomenclature from optimize.leastsq
  • #3196: scipy.stats.mstats.gmean does not actually take dtype
  • #3212: Dot product of csr_matrix causes segmentation fault
  • #3227: ZeroDivisionError in broyden1 when initial guess is the right...
  • #3238: lbfgsb output not suppressed by disp=0
  • #3249: Sparse matrix min/max/etc don’t support axis=-1
  • #3251: cdist performance issue with ‘sqeuclidean’ metric
  • #3279: logm fails for singular matrix
  • #3285: signal.chirp(method=’hyp’) disallows hyperbolic upsweep
  • #3299: MEMORY LEAK: fmin_tnc
  • #3330: test failures with the current master
  • #3345: scipy and/or numpy change is causing tests to fail in another...
  • #3363: splu does not work for non-vector inputs
  • #3385: expit does not handle large arguments well
  • #3395: specfun.f doesn’t compile with MinGW
  • #3399: Error message bug in scipy.cluster.hierarchy.linkage
  • #3404: interpolate._ppoly doesn’t build with MinGW
  • #3412: Test failures in signal
  • #3466: `scipy.sparse.csgraph.shortest_path` does not work on `scipy.sparse.csr_matrix` or `lil_matrix`

Pull requests

  • #442: ENH: sparse: enable 64-bit index arrays & nnz > 2**31
  • #2766: DOC: remove doc/seps/technology-preview.rst
  • #2772: TST: stats: Added a regression test for stats.wilcoxon. Closes...
  • #2778: Clean up stats._support, close statistics review issues
  • #2792: BUG io: fix file descriptor closing for netcdf variables
  • #2847: Rice distribution: extend to b=0, add an explicit rvs method.
  • #2878: [stats] fix formulas for higher moments of dweibull distribution
  • #2904: ENH: moments for the zipf distribution
  • #2907: ENH: add coverage info with coveralls.io for Travis runs.
  • #2932: BUG+TST: setdiag implementation for dia_matrix (Close #2931)...
  • #2942: Misc fixes pointed out by Eclipse PyDev static code analysis
  • #2946: ENH: allow non-monotonic input in interp1d
  • #2986: BUG: runtests: chdir away from root when running tests
  • #2987: DOC: linalg: don’t recommend np.linalg.norm
  • #2992: ENH: Add “limit” parameter to dijkstra calculation
  • #2995: ENH: Use int shape
  • #3006: DOC: stats: add a log base note to the docstring
  • #3007: DEP: stats: Deprecate randwppf and randwcdf
  • #3008: Fix mstats.kurtosistest, and test coverage for skewtest/normaltest
  • #3009: Minor reST typo
  • #3010: Add scipy.optimize.Result to API docs
  • #3012: Corrects documentation error
  • #3052: PEP-8 conformance improvements
  • #3064: Binned statistic
  • #3068: Fix Issue #3067 fix cumptrapz that was raising an exception when...
  • #3073: Arff reader with nominal value of 1 character
  • #3074: Some maintenance work
  • #3080: Review and clean up all Box-Cox functions
  • #3083: Bug: should return 0 if no regions found
  • #3085: BUG: Use zpk in IIR filter design to improve accuracy
  • #3101: refactor stats tests a bit
  • #3112: ENH: implement Akima interpolation in 1D
  • #3123: MAINT: an easier way to make ranges from slices
  • #3124: File object support for imread and imsave
  • #3126: pep8ify stats/distributions.py
  • #3134: MAINT: split distributions.py into three files
  • #3138: clean up tests for discrete distributions
  • #3155: special: handle the edge case lambda=0 in pdtr, pdtrc and pdtrik
  • #3156: Rename optimize.Result to OptimizeResult
  • #3166: BUG: make curve_fit() work with array_like input. Closes gh-3037.
  • #3170: Fix numpy version checks
  • #3175: use numpy sinc
  • #3177: Update numpy version warning, remove oldnumeric import
  • #3178: DEP: remove deprecated umfpack wrapper. Closes gh-3002.
  • #3179: DOC: add BPoly to the docs
  • #3180: Suppress warnings when running stats.test()
  • #3181: altered sem func in mstats to match stats
  • #3182: Make weave tests behave
  • #3183: ENH: Add k-sample Anderson-Darling test to stats module
  • #3186: Fix stats.scoreatpercentile
  • #3187: DOC: make curve_fit nomenclature same as leastsq
  • #3201: Added axis keyword to dendrogram function
  • #3207: Make docstring examples in stats.distributions docstrings runnable
  • #3218: BUG: integrate: Fix banded jacobian handling in the “vode” and...
  • #3222: BUG: limit input ranges in special.nctdtr
  • #3223: Fix test errors with numpy master
  • #3224: Fix int32 overflows in sparsetools
  • #3228: DOC: tf2ss zpk2ss note controller canonical form
  • #3234: Add See Also links and Example graphs to filter design *ord functions
  • #3235: Updated the buttord function to be consistent with the other...
  • #3239: correct doc for pchip interpolation
  • #3240: DOC: fix ReST errors in the BPoly docstring
  • #3241: RF: check write attr of fileobject without writing
  • #3243: a bit of maintanence work in stats
  • #3245: BUG/ENH: stats: make frozen distributions hold separate instances
  • #3247: ENH function to return nnz per row/column in some sparse matrices
  • #3248: ENH much more efficient sparse min/max with axis
  • #3252: Fast sqeuclidean
  • #3253: FIX support axis=-1 and -2 for sparse reduce methods
  • #3254: TST tests for non-canonical input to sparse matrix operations
  • #3272: BUG: sparse: fix bugs in dia_matrix.setdiag
  • #3278: Also generate a tar.xz when running paver sdist
  • #3286: DOC: update 0.14.0 release notes.
  • #3289: TST: remove insecure mktemp use in tests
  • #3292: MAINT: fix a backwards incompatible change to stats.distributions.__all__
  • #3293: ENH: signal: Allow upsweeps of frequency in the ‘hyperbolic’...
  • #3302: ENH: add dtype arg to stats.mstats.gmean and stats.mstats.hmean
  • #3307: DOC: add note about different ba forms in tf2zpk
  • #3309: doc enhancements to scipy.stats.mstats.winsorize
  • #3310: DOC: clarify matrix vs array in mmio docstrings
  • #3314: BUG: fix scipy.io.mmread() of gzipped files under Python3
  • #3323: ENH: Efficient interpolation on regular grids in arbitrary dimensions
  • #3332: DOC: clean up scipy.special docs
  • #3335: ENH: improve nanmedian performance
  • #3347: BUG: fix use of np.max in stats.fisher_exact
  • #3356: ENH: sparse: speed up LIL indexing + assignment via Cython
  • #3357: Fix “imresize does not work with size = int”
  • #3358: MAINT: rename AkimaInterpolator to Akima1DInterpolator
  • #3366: WHT: sparse: reindent dsolve/*.c *.h
  • #3367: BUG: sparse/dsolve: fix dense matrix fortran order bugs in superlu...
  • #3369: ENH minimize, minimize_scalar: Add support for user-provided...
  • #3371: scipy.stats.sigmaclip doesn’t appear in the html docs.
  • #3373: BUG: sparse/dsolve: detect invalid LAPACK parameters in superlu...
  • #3375: ENH: sparse/dsolve: make the L and U factors of splu and spilu...
  • #3377: MAINT: make travis build one target against Numpy 1.5
  • #3378: MAINT: fftpack: Remove the use of 'import *' in a couple test...
  • #3381: MAINT: replace np.isinf(x) & (x>0) -> np.isposinf(x) to avoid...
  • #3383: MAINT: skip float96 tests on platforms without float96
  • #3384: MAINT: add pyflakes to Travis-CI
  • #3386: BUG: stable evaluation of expit
  • #3388: BUG: SuperLU: fix missing declaration of dlamch
  • #3389: BUG: sparse: downcast 64-bit indices safely to intp when required
  • #3390: BUG: nonlinear solvers are not confused by lucky guess
  • #3391: TST: fix sparse test errors due to axis=-1,-2 usage in np.matrix.sum().
  • #3392: BUG: sparse/lil: fix up Cython bugs in fused type lookup
  • #3393: BUG: sparse/compressed: work around bug in np.unique in earlier...
  • #3394: BUG: allow ClusterNode.pre_order() for non-root nodes
  • #3400: BUG: cluster.linkage ValueError typo bug
  • #3402: BUG: special: In specfun.f, replace the use of CMPLX with DCMPLX,...
  • #3408: MAINT: sparse: Numpy 1.5 compatibility fixes
  • #3410: MAINT: interpolate: fix blas defs in _ppoly
  • #3411: MAINT: Numpy 1.5 fixes in interpolate
  • #3413: Fix more test issues with older numpy versions
  • #3414: TST: signal: loosen some error tolerances in the filter tests....
  • #3415: MAINT: tools: automated close issue + pr listings for release...
  • #3440: MAINT: wrap sparsetools manually instead via SWIG
  • #3460: TST: open image file in binary mode
  • #3467: BUG: fix validation in csgraph.shortest_path