summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
* DOC/MAINT: Python 2.6 or 3.2 requiredLars Buitinck2014-09-251-1/+1
| | | | | | | | | Installation docs still stated Python 2.4 is supported, while not mentioning 3.x at all. Also corrected the statement that distutils is in python-dev on Debian, while it's actually in the main package. Several parts of the dev docs and some comments also referenced now-irrelevant 2.5 bugs/limitations.
* TST: extend truth table test to arrays with various dtypesLars Buitinck2014-09-241-10/+32
|
* ENH: faster (branchless) logical_xorLars Buitinck2014-09-241-8/+8
|
* Merge pull request #5104 from juliantaylor/inner-fixCharles Harris2014-09-231-0/+3
|\ | | | | BUG: check success of PyArray_DescrFromType in PyArray_InnerProduct
| * BUG: check success of PyArray_DescrFromType in PyArray_InnerProductJulian Taylor2014-09-231-0/+3
| | | | | | | | closes gh-5065
* | Merge pull request #5107 from juliantaylor/na-docCharles Harris2014-09-231-9/+1
|\ \ | | | | | | DOC: remove preservena reference from docstrings
| * | DOC: remove preservena reference from docstringsJulian Taylor2014-09-231-9/+1
| |/ | | | | | | | | | | | | preservena is not not implemented. the putmask docstring is misleading, currently copyto is faster for dense or sparse masks while putmask is faster for random masks. [ci skip]
* | Merge pull request #5106 from juliantaylor/sequence-fixCharles Harris2014-09-232-1/+23
|\ \ | | | | | | BUG: check if object provides len() before trying to iterate it
| * | BUG: check if object provides len() before trying to iterate itJulian Taylor2014-09-232-1/+23
| |/ | | | | | | | | | | some libraries want object arrays from objects that are iterable but rely on not providing len() to get the right dtype from numpy. closes gh-5100
* | Merge pull request #5105 from juliantaylor/where-no-subtypeCharles Harris2014-09-231-1/+1
|\ \ | | | | | | BUG: don't propagate subtypes from np.where
| * | BUG: don't propagate subtypes from np.whereJulian Taylor2014-09-231-1/+1
| |/ | | | | | | | | | | the C implementation cannot preserve subclass attributes, this needs a python wrapper using indexing for subclasses. closes gh-5095
* | Merge pull request #5099 from perimosocordiae/patch-1Jaime2014-09-232-1/+7
|\ \ | | | | | | BUG: np.unique with chararray + inverse_index
| * | BUG: np.unique with chararray + inverse_indexCJ Carey2014-09-232-1/+7
| | | | | | | | | | | | | | | The call to `empty_like` was trying to use the `chararray` subclass, which doesn't support the `np.intp` dtype.
* | | Merge pull request #5088 from juliantaylor/not-fixCharles Harris2014-09-199-11/+14
|\ \ \ | |/ / |/| | BUG: fix wrong error checks of PyBytes_AsStringAndSize
| * | BUG: fix wrong error checks of PyBytes_AsStringAndSizeJulian Taylor2014-09-209-11/+14
| | | | | | | | | | | | | | | | | | | | | The check in ufunc_type_resolution.c is broken due to a spurious not and a few cases of TypeError being overwritten without clearing. Also change all of these error checks to a negative check for consistency.
* | | Merge pull request #5087 from juliantaylor/unicode-argminCharles Harris2014-09-192-2/+12
|\ \ \ | | | | | | | | BUG: fix out of bound access in unicode argmin/argmax
| * | | BUG: fix out of bound access in unicode argmin/argmaxJulian Taylor2014-09-192-2/+12
| | |/ | |/| | | | | | | | | | | | | elsize is in bytes and the pointer of unicode type, so it must divided by the size. Closes gh-5082
* | | Merge pull request #5034 from seberg/inplace-ravelJulian Taylor2014-09-182-17/+92
|\ \ \ | | | | | | | | ENH: Allow ravel to reshape in place when possible
| * | | ENH: Allow ravel to reshape in place when possibleSebastian Berg2014-09-142-17/+92
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This fixes a lot of corner cases in reshape 'K' when the array is not contiguous in the first place, it was previously never ravelled in place. Closes gh-5033
* | | | BUG: Readd the cblas.h header.Charles Harris2014-09-174-8/+587
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Rename cblas.h to npy_cblas.h and put it in numpy/core/src/private. Fix the various includes to account for the change. It turns out that MKL provides cblas, but not cblas.h. Rather, they provide mkl_cblas.h, which also uses typedefs for the CBLAS_XXX enums.
* | | | DOC: update numpy.extract docstringGregory R. Lee2014-09-151-1/+3
| | | |
* | | | Merge pull request #5072 from charris/remove-testcalcs.pyJaime2014-09-151-68/+0
|\ \ \ \ | | | | | | | | | | MAINT: Remove testcalcs.py file.
| * | | | MAINT: Remove testcalcs.py file.Charles Harris2014-09-151-68/+0
| |/ / / | | | | | | | | | | | | | | | | | | | | The testcalcs.py file is nowhere imported or installed, nor did anyone reply when it was proposed on the mailing list to remove it. It looks to be a leftover from the original work on datetime.
* | | | Merge pull request #5070 from juliantaylor/gnubsd-fixCharles Harris2014-09-141-3/+1
|\ \ \ \ | |/ / / |/| | | TST: gnufreebsd has fixed the thread control word loss
| * | | TST: gnufreebsd has fixed the thread control word lossJulian Taylor2014-09-141-3/+1
| | | | | | | | | | | | | | | | http://lists.alioth.debian.org/pipermail/glibc-bsd-commits/2014-September/004163.html
* | | | Merge pull request #5061 from keszybz/fix-dimension-messageJulian Taylor2014-09-132-2/+2
|\ \ \ \ | | | | | | | | | | Fix dimension message
| * | | | BUG: fix off-by-one in dimension error messageZbigniew Jędrzejewski-Szmek2014-09-101-1/+1
| | | | | | | | | | | | | | | | | | | | The maximum allowed is really NP_MAX_INT, not NP_MAX_INT-1.
| * | | | BUG: fix off-by-one in dimensionality error messageZbigniew Jędrzejewski-Szmek2014-09-101-1/+1
| | | | | | | | | | | | | | | | | | | | The maximum allowed is really 32, not 31.
* | | | | Merge pull request #5024 from pp-mo/in1d_of_setJulian Taylor2014-09-131-0/+4
|\ \ \ \ \ | | | | | | | | | | | | DOC: warn about using a set with 'in1d'.
| * | | | | DOC: warn about using a set with 'in1d'.Patrick Peglar2014-08-311-0/+4
| | | | | |
* | | | | | Merge pull request #5045 from ryanblak/argmin-docJulian Taylor2014-09-131-4/+44
|\ \ \ \ \ \ | | | | | | | | | | | | | | DOC: added doc for argmin
| * | | | | | DOC: added doc for argminryanblak2014-09-041-4/+44
| | | | | | |
* | | | | | | Merge pull request #5066 from seberg/index-sequenceJulian Taylor2014-09-132-2/+36
|\ \ \ \ \ \ \ | |_|_|_|/ / / |/| | | | / / | | |_|_|/ / | |/| | | | BUG: cannot convert invalid sequence index to tuple
| * | | | | TST: Test tuple subclass index and broken sequence indexSebastian Berg2014-09-131-1/+32
| | | | | |
| * | | | | BUG: cannot convert invalid sequence index to tupleSebastian Berg2014-09-131-2/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A sequence which raises an error on item getting, should be just handled as a single index and not a tuple nd-index. Also fix some error handling. closes gh-5063
* | | | | | Merge pull request #5067 from juliantaylor/restore-dupCharles Harris2014-09-111-1/+1
|\ \ \ \ \ \ | |_|_|_|/ / |/| | | | | BUG: restore api of npy_PyFile_DupClose in python2
| * | | | | BUG: restore api of npy_PyFile_DupClose in python2Julian Taylor2014-09-111-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | breaks matplotlib
* | | | | | MAINT: make main a strict prototype in configure checksJulian Taylor2014-09-091-6/+6
| | | | | | | | | | | | | | | | | | | | | | | | removes about 400 lines of warnings from the build logs
* | | | | | Merge pull request #5056 from juliantaylor/scalar-exactCharles Harris2014-09-076-36/+107
|\ \ \ \ \ \ | | | | | | | | | | | | | | add C-API PyArray_IsAnyScalarExact
| * | | | | | ENH: skip numpy_ufunc and priority lookup for numpy scalarsJulian Taylor2014-09-072-2/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | improves rmul with scalars, gh-5030
| * | | | | | API: add PyArray_IsAnyScalarExact for a fast check for scalarsJulian Taylor2014-09-074-4/+25
| | | | | | |
| * | | | | | ENH: search typenumbers from typeobjects via a binary searchJulian Taylor2014-09-071-34/+79
| | | | | | |
* | | | | | | Merge pull request #5047 from juliantaylor/move-dotblas-to-multiarrayCharles Harris2014-09-0723-1559/+1051
|\ \ \ \ \ \ \ | | | | | | | | | | | | | | | | Move dotblas to multiarray
| * | | | | | | ENH: add better error message for misaligned matricesJulian Taylor2014-09-044-9/+67
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | based on gh-4985 by Lars Buitinck
| * | | | | | | MAINT: cleanup unnecessary codeJulian Taylor2014-09-043-41/+2
| | | | | | | |
| * | | | | | | BLD: check for CBLAS header in "unoptimized" blasJulian Taylor2014-09-041-1/+25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Allows building against system installed blas library that might be optimized.
| * | | | | | | BUG: Capitalize environmental variables in numpy/core/__init__.py.Charles Harris2014-09-041-6/+5
| | | | | | | |
| * | | | | | | TST: Add vdot tests, move tests from test_blasdot to test_multiarray.Charles Harris2014-09-042-163/+74
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Not all tests from test_blasdot were relevant, as they checked the cblas implementations against the non-cblas implementations. Those two are no longer separate. Remove test_blasdot.py as there nothing is left in it.
| * | | | | | | DOC: Update docs to reflect deprecation of alterdot and restoredot.Charles Harris2014-09-042-37/+44
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Also move docstrings into the versions in numpy/core/numeric.py as the functions are no longer in the defunct _dotblas module.
| * | | | | | | ENH: Move vdot to multiarray.Charles Harris2014-09-0415-1000/+384
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Remove vdot from _dotblas and implement it in multiarray. Remove the files in core/blasdot as they are no longer needed. Fix tests and build to reflect the changes. The vdot function is now a bit faster in the common case as ravel is used instead of flatten. There is also no need to conjugate the files for clongdouble.
View Lorry Controller Status