TST: ndimage: add more tests in test_gaussian_truncate() #1

Open
wants to merge 2,358 commits into from
@WarrenWeckesser

This PR strengthens the existing test, and adds a few more.

endolith and others added some commits Jul 25, 2013
@endolith endolith DOC: Remove a comment redundant with nearby docstring,
change a comment to docstring,
spell check a comment, add TODO to a comment
fe7a504
@endolith endolith TST:Create tests for step and impulse based on tests for step2 and im…
…pulse2

Didn't copy tests that pass for step2 and impulse 2 but fail for step and
impulse, though these should probably be addressed at some point
266bc51
@endolith endolith TST: Add tests for step/impulse/step2/impulse2 for accepting array_li…
…ke without errors
8a005e5
@rgommers rgommers BUG: fix a couple of bugs in signal.ltisys ce3b420
@rgommers rgommers TST: de-duplicate tests for signal.ltisys d1823f9
@rgommers rgommers TST: add regression test with complex input for signal.step.
Note that step2/impulse/impulse2 don't work with this input (see gh-2654).

When unifying step/step2 and impulse/impulse2, a decision should be made on
whether complex inputs need to be supported.
0058fd0
@rgommers rgommers Merge branch 'pr/2654' into master. f4e6379
@rgommers rgommers Merge pull request #2685 from rgommers/update-release-notes
DOC: update 0.13.0 release notes.
ab5fd28
@rgommers rgommers BUG: fix test failure for rv_discrete.entropy with numpy 1.5.1. Closes 05772f7
@pv pv special/struve: deal with breakdown of the asymptotic expansion ea2bef9
@pv pv special/struve: fix convergence criterion of power series
For v < 0, |v| >> 1, the summand may start to increase again for n > v,
so push the convergence point far enough.
61a2b0e
@pv pv special/struve: update comments d98bd02
@pv pv special/struve: fix c++ style comments f3f72e8
@ClemensFMN ClemensFMN DOC: Extended tutorial/integrate 9493e6f
@ClemensFMN ClemensFMN DOC: improved integrate tutorial according to comments on PR 2708. 600019f
@ClemensFMN ClemensFMN MAINT: removed nowrap and brackets in integrate tutorial. c66ebe3
@rgommers rgommers MAINT: minor cleanup in integrate tutorial, add Clemens Novak to auth…
…or list
0ab57f1
@rgommers rgommers Merge branch 'pr/2708' into master 5776368
@rgommers rgommers TST: remove incorrect (and useless) tests for rounding numbers.
Also some general cleanup in test_stats.py.  See gh-2482.
b1ab3d4
@rgommers rgommers TST: fix equal comparison of float arrays in spatial.distance tests.
Closes gh-2482.
622f342
@pv pv TST: sparse: fix tests
Fix the slicing test, and do not run inverse test for booleans (this
fails if UMFPACK is enabled).
065f621
@cowlicks cowlicks Merge pull request #1 from pv/pr-2689
Pr 2689
d9ae272
@cowlicks cowlicks WIP: Add check for size zero slice result. 0e523ee
@WarrenWeckesser WarrenWeckesser BUG: stats: fixed tests of stats.power_divergence with empty arrays 7147e58
@rgommers rgommers Merge pull request #2716 from WarrenWeckesser/bug-powerdiv-tests
BUG: stats: fixed tests of stats.power_divergence with empty arrays
d3b827d
@michaelwimmer michaelwimmer add missing ABI wrappers for blas/lapack for MacOSX Accelerate framework 6968998
@michaelwimmer michaelwimmer fix missing blas/lapack ABI wrappers for the iterative solvers for Ma…
…cOSX Accelerate
8112345
@michaelwimmer michaelwimmer fix the blas/lapack ABI wrappers for arpack for MacOSX Accelerate fra…
…mework
5935030
@michaelwimmer michaelwimmer enable single precision in ARPACK again aad913b
@michaelwimmer michaelwimmer add tests for singular precision in iterative module 8578f17
@michaelwimmer michaelwimmer fix ABI also for scipy.lib.blas 79705ff
@michaelwimmer michaelwimmer fix readme 9486ab5
@michaelwimmer michaelwimmer fix file name of abi wrapper 26edefa
@michaelwimmer michaelwimmer remove Lapack functions introduced only from v3.2 onward 7b2776b
@michaelwimmer michaelwimmer repeat arpack test if eigenvectors are correct, but eigenvalues are n…
…ot as requested in `which`
fa7e3b9
@pv pv Merge pull request #2695 from michaelwimmer/master
BUG: Completion of g77 ABI wrappers for MacOSX

The MacOSX Accelerate framework adheres to g77 ABI conventions.  This
implies that

- every function returning a single precision floating point value
  (Fortran REAL) in fact returns a double precision value (Fortran DOUBLE PRECISION)
- every function returning a complex number is in fact internally a subroutine
  with the result as the first argument, i.e. COMPLEX FUNCTION X(A, B, C) is in
  fact SUBROUTINE X(RESULT, A, B, C)

Trying to link gfortran generated code with this different ABI results either
in crashes (typically for complex) or wrong results (especially for REAL).
Scipy already contained wrappers that solved these ABI issues for some
BLAS/LAPACK routines (in particular C,ZDOTC, C,ZDOTU), but these ABI wrappers
were rather incomplete.

As a result, the scipy tests involving single precision failed often on OSX
machines (see e.g. issue gh-2248).

In this pull request:

- the ABI wrappers for the basic lapack/blas wrappers in scipy.linalg and scipy.lib are completed
- the ABI wrappers for the Fortran routines in scipy.sparse.linalg.isolve are completed
- test cases for single precision for the iterative solvers are added
  (there were no before, so the failure of the iterative solvers on OSX was unnoticed)
- the ABI wrappers for ARPACK are completed
- single precision in ARPACK has been enabled again

With these changes, all relevant tests pass again for my MacOSX 10.8 machine
that showed many failures before these patches.
62883c9
@cowlicks cowlicks WIP: Disallow indexes which return size 0 sparse matrices.
Also add errors in this case.
c1fee97
@cowlicks cowlicks WIP, TST: Add knownfailures to tests for size == 0 cases.
Also removed tests that previously checked the size 0 workaround.
c859927
@pv pv DOC: sparse: docstring fixes 8f4a276
@pv pv Revert "WIP: Change CSR index checking to allow slices like `1:1`."
This reverts commit 32ca14c.
75eec75
@pv pv BUG: sparse/sputils: fix issues in issequence / ismatrix ad3312b
@pv pv TST: sparse/basic: enable bool index tests for LIL 7c89c32
@alexbrc alexbrc DOC: matfunc release notes 2971f25
@pv pv Merge pull request #2689 from cowlicks/fancy-indexing
ENH: sparse: Fancy indexing for CSR and CSC matrices

Add initial fancy indexing support for CSR and CSC.
76bce8b
@alexbrc alexbrc DOC: use double back-ticks to improve rendering through github f184b48
@pv pv Merge pull request #2722 from argriffing/release-notes-13
DOC: new/updated matfuncs implementation descriptions for 0.13 release notes
c5a7111
@pv pv BUG: special: add asymptotic expansions to shichi
Enable also the si+ci tests, they seem to work.
5540dbd
@woodscn woodscn ENH: add N-dimensional integration function.
New n-dimensional integration routine 'nquad' included in scipy.integrate.
a00ae73
@woodscn woodscn MAINT: some code formatting changes in integrate.nquad. b0037a0
@alexbrc alexbrc MAINT: less in-place modification 73f93c3
@alexbrc alexbrc MAINT: less complicated variable args e0cebaa
@alexbrc alexbrc MAINT: merge two functions f30e9c1
@alexbrc alexbrc BUG: recursion termination condition e40d8fe
@rgommers rgommers MAINT: fix documentation formatting and PEP8 for integrate.nquad code. 2d5a7a6
@rgommers rgommers DOC: fix a few issues in the integrate examples. 4ccb036
@rgommers rgommers BUG: fix nquad issue for depth=1 integration. Add tests against dbl/t…
…plquad.
c48c70e
@woodscn woodscn DOC: improve clarity in integrate.nquad docstring. abf9c47
@rgommers rgommers DOC: add integrate.nquad to 0.13.0 release notes. c3f79be
@rgommers rgommers Merge branch 'pr/2510' into master.
Reviewed as scipy#2510
3f43d22
@WarrenWeckesser WarrenWeckesser BUG+ENH: stats: Improve the 'fit' method of norm_gen, gamma_gen and b…
…eta_gen.

Closes github issue gh-2514.

* scipy.misc.doccer: Added the 'inherit_docstring_from' decorator for
  inheriting  docstrings in methods of subclasses.
* scipy.stats.beta_gen:
  * Correctly handle keyword arguments that fix parameters.
  * When both floc and fscale are given, the parameters are estimated by MLE
    (as stated in the docstring). Previously it had been using the method
    of moments.
  * When both floc and fscale are given, an error is raised if any data
    values are not in the open interval (floc, floc+fscale).
* scipy.stats.gamma_gen:
  * Correctly handle the keyword arguments that fix parameters.
  * A regularization term was added in the denominator of the formula for
    estimating the shape parameter a in gamma_gen._fitstart.
* scipy.stats.norm_gen: Implement a 'fit' method that uses the explicit
  formulas for the maximum likelihood estimation of the parameters.
00d2665
@WarrenWeckesser WarrenWeckesser Merge pull request #2711 from rgommers/vecentropy
BUG: fix test failure for rv_discrete.entropy with numpy 1.5.1.  Closes ...
7114368
@cbrueffer cbrueffer Spell Smirnov's name properly. bfa5d6b
@cbrueffer cbrueffer Remove spurious EOL and EOF whitespace. a040d53
@rgommers rgommers Merge pull request #2519 from WarrenWeckesser/bug-gamma-beta-fit
BUG+ENH: stats:  Improve the 'fit' method of norm_gen, gamma_gen and beta_gen.
91974df
@WarrenWeckesser WarrenWeckesser Merge pull request #2714 from rgommers/issue-2482-test-rounding
Fix issue 2482, test failures in spatial and stats
9e6f931
@WarrenWeckesser WarrenWeckesser BUG: stats: The formula for the skew in poisson.stats() was incorrect…
…--it did not take the sqrt.
99c4b8e
@josef-pkt josef-pkt BUG: levene didn't sort before trimming closes #2555 039f133
@cowlicks cowlicks TST: Silence some warnings. 6fc83c8
@cowlicks cowlicks TST: Small test optimizations. 5fc9e3e
@cowlicks cowlicks TST: Add warning catches for indexing and inv and exp tests. a335d56
@cowlicks cowlicks TST: Add warning catches to tests in sparse.linalg. 6a98e99
@rgommers rgommers TST: remove duplicate with-statement in sparse/tests/test_base.py 03b28bd
@rgommers rgommers Merge branch 'pr/2731' into master. 51d5f96
@pv pv BUG: spatial/qhull: fix memory leak in setlist.pxd e0c24d5
@pv pv TST: interpolate: add a benchmark for checking if interpnd leaks memory 53ce59f
@rgommers rgommers Merge pull request #2728 from WarrenWeckesser/stats-poisson-skew
BUG: stats: The formula for the skew in poisson.stats() was incorrect--it did not take the sqrt.
61f7aeb
@josef-pkt josef-pkt REF: levene test: add random seed 3f4c735
@rgommers rgommers Merge pull request #2723 from pv/special-ci
BUG: special: add asymptotic expansions to shichi
f339e7e
@rgommers rgommers TST: silence RuntimeWarning's in stats tests for empty input. b18df76
@rgommers rgommers Merge pull request #2712 from pv/special-struve
special: Reimplement Struve functions
08813e3
@gertingold gertingold more stable handling of arguments close to 1 8aa1a00
@gertingold gertingold hyp2f1 issue 1561 corrected sign of gamma functions d853fd2
@pv pv DOC: release: minor updates and reordering to the 0.13.0 release notes 3f307c7
@pv pv DOC: release: explain how to get previous LIL behavior bf605c0
@josef-pkt josef-pkt BUG: binom_test failed for some values, closes issue #2384 2938e69
@gertingold gertingold corrected signs of gamma functions df92b33
@josef-pkt josef-pkt TST: binom_test more unittests for edgecase and close c3722f8
@josef-pkt josef-pkt REF: binom_test add shortcut for special case (unit tests previous co…
…mmit)
75e9ee6
@pv pv BLD: track add_library dependencies in distutils
Distutils does not automatically track source dependencies to
numpy.distutils add_library libraries. So it needs to be manually.

This commit causes the appropriate extension modules be rebuilt if a
change has been made in some of the sources of the libraries it depends
on.
8eb2987
@pv pv Merge pull request #2734 from gertingold/issue1561
BUG: special/hyp2f1: more stable handling of arguments close to 1
287b592
@pv pv Merge pull request #2727 from cbrueffer/misc
DOC: stats: spell Smirnov's name properly.
fff3594
@pv pv DOC: HACKING: add instructions for debugging compiled code 0bd28ff
@pv pv DOC: remove outdated TOCHANGE.txt 463d390
@pv pv DOC: remove not so useful Py3K.txt 67bc8ce
@pv pv DOC: move material from README to HACKING, remove outdated parts, and…
… move LATEST to README
0fea1f4
@andrenarchy andrenarchy [ENH] sparse.linalg.LinearOperator: add +, *, **
Enhances LinearOperator such that they can be
added, multiplied and exponentiated.
19f3aa5
@andrenarchy andrenarchy [TST] sparse.linalg.LinearOperator: test +, *, ** cd2a581
@andrenarchy andrenarchy MAINT,TST: use isintlike and add __pow__ test a644fe8
@andrenarchy andrenarchy MAINT: use subclasses for LinearOperator operations 8b2b198
@andrenarchy andrenarchy LinearOperator: make it callable 89180b1
josef-pkt and others added some commits Aug 16, 2013
@josef-pkt josef-pkt Bug: stats trim_mean for 2d array, add axis argument closes #2554 a616a14
@rgommers rgommers MAINT: docstring and style fixes for stats.trimboth/trim_mean c72b636
@rgommers rgommers Merge branch 'pr/2729' into master. 6bb29aa
@ev-br ev-br DOC: remove scale parameter from the discrete distributions docstrings 0381615
@ev-br ev-br DOC: remove xa and xb from the stats tutorial e27ed0c
@ev-br ev-br DOC: mention shape args as keywords in stats tutorial b054f79
@rgommers rgommers DOC: some minor tweaks to the stats tutorial. ce40c9d
@rgommers rgommers Merge branch 'pr/2757' into master 36d10a7
@richardtsai richardtsai BUG: fix wrong output shape in ndimage.zoom()
ndimage.zoom() may calculate the output shape incorrectly
due to round-off error in some occasions. Rounding to
nearest will be a better method for calculating the the
output shape compared to truncation.
0bf2b81
@richardtsai richardtsai issue a warning when the old and new shape differ b039e49
@rgommers rgommers TST: filter ndimage.zoom warning and clean up ndimage tests. ec4ee7a
@rgommers rgommers Merge branch 'pr/2706' into master. 61dc8f0
@rgommers rgommers REL: minor changes to release script and 0.13.0 release notes. 050ac31
@rgommers rgommers BLD: update Bento build scripts for OS X Accelerate wrapper changes. 444f7f6
@rgommers rgommers BLD: update Bento build scripts for changes to special._ufuncs
Closes gh-2752.
b7d2019
@rgommers rgommers BLD: fix MinGW Windows build issue by splitting linalg.interpolate so…
…urces.

Type of errors that this fixes::

    Argument #5 (named `ifac') of `zfftb1' is one type at (2) but is some
    other type at (1) [info -f g77 M GLOBALS]

Warning shows up with gfortran as::

    Warning: Type mismatch in argument 'ifac' at (1); passed REAL(8) to INTEGER(4)

Closes gh-2709.
72dc1d6
@rgommers rgommers BLD: ignore split Fortran source files created by the build process.
Also add some other generated files to .gitignore
b99c669
@rgommers rgommers BLD: use relative paths in linalg/setup.py 10dfce5
@rgommers rgommers MAINT: address review comments on splitting Fortran files at build time. 9c89dad
@rgommers rgommers BLD: exclude split Fortran files from sdist generated source bundles. 86ba46c
@rgommers rgommers TST: silence test warnings about deprecation of expm2/expm3. 165bbb6
@cowlicks cowlicks BUG: Raise error when attempting to create dtype=object sparse matrix. f23a87d
@cowlicks cowlicks DOC: Add note about SciPy being for numeric data. d301a57
@cowlicks cowlicks ENH: Add getrow and getcol to DOK. 5570f8d
@cowlicks cowlicks BUG: Fix casting in _mul_scalar for DOK and LIL.
The way I did the casting is pretty hackish. There should be a better
way.
6dd6e01
@cowlicks cowlicks TST: Turn on scalar multiplication tests for LIL and DOK. e5c9e67
@cowlicks cowlicks BUG: Remove broken take and split methods from DOK. cf3215c
@pv pv ENH: sparse: move scalar upcast dtype computation to a helper function 3cc9213
@pv pv MAINT: sparse: convert object dtype error to a warning e941098
@pv pv Merge pull request #2719 from cowlicks/dok-lil-improvements
BUG: sparse: DOK and LIL improvements

* Raise a warning when you try to create a sparse matrix with
  dtype=object, and add a note to the docs stating that SciPy Sparse is for
  numeric data.
* Remove the broken and undocumented take and split methods from DOK, add getrow and getcol.
* Fix the casting for scalar multiplication with LIL and DOK and turn on tests for them.

Fixes gh-2528, gh-2645, and gh-2542.
6619f15
@pv pv BLD: recognize one more Fortran comment character 557b6f0
@pv pv Merge pull request #2761 from rgommers/fix-mingw-compilation
BLD: Fix MinGW compilation by splitting source files

Fixes gh-2709
ba8eec1
@alexbrc alexbrc MAINT: attempt debugging some code that behaves differently on differ…
…ent LAPACKs
26a076e
@alexbrc alexbrc DOC: remove an obsolete file f544095
@alexbrc alexbrc MAINT: disable a test that involves singular matrices and which seems…
… to not be stable across different LAPACKs
b1d4749
@rgommers rgommers Merge pull request #2768 from argriffing/lax-fractional-matrix-power-…
…singularity

TST: mark a fractional_matrix_power test as knownfail
7c36b55
@rgommers rgommers REL: set version of master to 0.14.0-dev, add 0.14.0 release notes. 78eb93c
@cowlicks cowlicks TST sparse: Tests for sparse matirx min and max axis args. 257f278
@cowlicks cowlicks ENH sparse: Add support for sparse matrix .min and .max axis arguments. 9494b22
@rgommers rgommers BLD: fix Bento build also on OS X. ab6b69e
@dlax dlax DOC: fix list of minimization methods supporting Jacobian 18e5c49
@dlax dlax BUG: count function evaluations in approximate Jacobian for L-BFGS-B
Closes #2305.
1cc015e
@WarrenWeckesser WarrenWeckesser TST: stats: Added a regression test for stats.wilcoxon. Closes gh-2391. c7988d1
@endolith endolith DOC: additional notes about FFTs, typo fixes, etc. e925071
@endolith endolith MAINT: Use boolean defaults for boolean parameter overwrite_x fabc4dd
@endolith endolith MAINT: Use boolean defaults for boolean overwrite_x parameter ed5e94f
@endolith endolith DOC: Add simple ifft Returns section d5ed83c
@endolith endolith DOC: Add examples for DCT, minor changes to DCT docstrings b4ab782
@endolith endolith DOC: elaborate on rfft and dct efficiencies cc9d3f7
@pv pv Merge pull request #2773 from dlax/issue-2305
ENH: optimize: count function evaluations in approximate jacobian for L-BFGS-B

Fixes gh-2305
30c08f7
@rgommers rgommers Merge pull request #2772 from WarrenWeckesser/stats-wilcoxon-tie-test
TST: stats: Added a regression test for stats.wilcoxon.  Closes gh-2391.
fea2a30
Alain Leufroy BUG: special/specfun: fix argument value check in KLVNZO
documentation of KLVNZO(NT,KD,ZO) in scipy/special/specfun.f
says 1<=KD<=8. Previous code resulted in always True.
303077a
@pawelchojnacki pawelchojnacki scipy.signal.wavelets.cwt doc easyfix from #2748 ticket 9294f5e
@pv pv Merge pull request #2780 from aleufroy/fix-2718
BUG: special/specfun: fix argument value check in KLVNZO
08e33a0
@rgommers rgommers MAINT: remove leftover check for Python <= 2.5. Some cleanups of comm…
…ents.
03b4239
@pv pv Merge pull request #2775 from endolith/patch-3
DOC: fftpack: add more notes about FFTs, real FFTs, fix typos, change boolean defaults to False
ba4fbd5
@pv pv Merge pull request #2784 from duskglow/doc-cwt
DOC: scipy.signal.wavelets.cwt argument order fix 

Fixes gh-2748
88aa5c6
@andreas-h andreas-h DOC: enhance interp2d docstring, see #2275 178c0a3
@pv pv Merge pull request #2783 from rgommers/cleanup
MAINT: remove leftover check for Python <= 2.5.  Some cleanups of comments
637170e
@andreas-h andreas-h DOC: enhance splrep docstring, see #2223 d6e9d40
@pv pv Merge pull request #2785 from andreas-h/interpolate_doc2
DOC: interpolate: enhance interp2d docstring, see #2275
498f7b7
@loujine loujine ENH interpolate: remove warnings in pchip
Division by zero occurs only at points excluded in the
following calculations, so warning messages can be removed
safely

Closes gh-2357
fd44b7f
@pv pv Merge pull request #2769 from cowlicks/min-max-axis-args
ENH: sparse: matrix min & max axis arguments.
2f1560c
@pv pv Merge pull request #2786 from loujine/fix-2357
ENH: interpolate: remove warnings in pchip
448e522
@andbo andbo Fix issues/2331: Bode plot returns continuous phase 04f4f41
@andbo andbo BUG: issues/2331, change to assert_almost_equal and style improvement ca0c159
@gertingold gertingold ensure analyticity of associate Legendre function of first kind 38ceae1
@andreas-h andreas-h DOC: add cython dependency to INSTALL 51b7145
@rgommers rgommers Merge pull request #2799 from andreas-h/INSTALL_cython
DOC: add cython dependency to INSTALL
a9d14a1
@rgommers rgommers DOC: make more clear when Cython is required. fb15a61
@WarrenWeckesser WarrenWeckesser BUG: stats: When handling the arguments to beta_gen.fit, don't rely o…
…n np.log

to generate a warning when its argument is less than or equal to 0.
Fixes gh-2771.
f5bbbc4
@WarrenWeckesser WarrenWeckesser BUG: stats: Raise a ValueError in beta_gen.fit if all the parameters …
…are fixed.
4adecca
@rgommers rgommers Merge pull request #2766 from argriffing/remove-obsolete-docs
DOC: remove doc/seps/technology-preview.rst
cec98d7
@andreas-h andreas-h ENH: allow optional rescaling of input coordinates in ND interpolation 0f43721
@rgommers rgommers Merge pull request #2802 from rgommers/issue-2062-circfuncs
BUG: handle array_like input in stats.circ* functions.  Closes gh-2062.
d2600d3
@cgohlke cgohlke BLD: Fix msvc error C2036: 'void *' : unknown size
(cherry-picked from 827083a)
5ec576b
@rgommers rgommers BUG: handle array_like input in stats.circ* functions. Closes gh-2062.
Also add test for empty input, return nans.
04403f9
@rgommers rgommers MAINT: remove lazy import stuff from lib.six. Closes gh-2764. 38dd1e4
@gertingold gertingold removed superfluous comments 27ff52d
@gertingold gertingold adapted documentation e863b62
@gertingold gertingold removed known failure for clpmn test 6539210
@rgommers rgommers Merge pull request #2801 from WarrenWeckesser/bug-beta-fit
BUG: stats: Fix a couple issues in the recently updated beta_gen.fit method.
57f649c
@gertingold gertingold improved documentation of lpmn and clpmn de82d0d
@pv pv BUG: special: fix square root conventions in clpmn and lpmn
Having the branch cut in (-1, 1) requires some care in the square roots.
815393d
@pv pv TST: special: fix mpmath tests for legenp 705c1ee
@pv pv DOC: special: adjust lpmn/clpmn documentation 039809f
@pv pv TST: interpolate: fix for Python 3 6a280b9
@pv pv BUG: interpolate/interpnd: rescale after translation 2f10b28
@pv pv Merge pull request #2802 from andreas-h/ndinterpolator
ENH: allow optional rescaling of input coordinates in ND interpolation

Fixes gh-2191
d625c40
@pv pv Merge pull request #2803 from rgommers/issue-2764-sixmoves
MAINT: remove lazy import stuff from lib.six
b061536
@pv pv BLD: special: fix build when system doesn't have isfinite macro 091e6fc
@gertingold gertingold do not allow evaluation of clpmn on the cut [-1,1] a022042
@gertingold gertingold remove exception if argument for clpmn is on the cut 468f702
@pv pv BUG: special: return inf rather than 1e300 in clpmn/lpmn to signal in…
…finity
472b8b2
@pv pv Merge pull request #2800 from gertingold/gh-2396
BUG: special: ensure analyticity of associated Legendre function of first kind

gh-2396 discusses the problem of phase factors in the Legendre function
for complex arguments. In fact, the Legendre function of the first kind
as it is implemented, exhibits phase jumps on the unit circle and thus
is not analytic.

These commits introduce a new function clpmn which provides an analytic
associated Legendre function of the first kind. The existing function
lpmn in the new version raises an exception if used for complex
arguments. The present implementation follows the conventions stated in
http://dlmf.nist.gov/14.21 with a cut on the interval (-1,1). Some tests
have been added to test_basic.py focusing on the behavior close to the
real axis and close to the unit circle. The latter tests the absence of
phase jumps which were present before. Also, one test in test_mpmath.py
has been adapted to the new function.
44e7c61
@pv pv DOC: release: document the change in lpmn 2b3b9c2
@pv pv Merge pull request #2791 from andbo/master
BUG: signal: Bode plot should return continuous phase
1dff009
@thouis ENH: expose control truncation of ndimage.gaussian_filter().
This change gives the user control over the size at which the Gaussian filter is truncated,
with default behavior to truncate at 4 standard deviations (the previous fixed value).
19eff1c
@thouis ENH: Pass extra keyword arguments to gaussian_laplace() and gaussian_…
…gradient_magnitude() to gaussian_filter()

This change allows control over the truncation radius of the gaussian filter used in
gaussian_laplace() and gaussian_gradient_magnitude().
f6c8d9d
@WarrenWeckesser WarrenWeckesser TST: ndimage: add more tests in test_gaussian_truncate() cf3aa17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment