summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
* ENH: Masked array simplify copying back of unchanged valuesSebastian Berg2016-09-021-6/+1
| | | | | Also silences a spurious warning during tests (the multiplication could give a warning).
* ENH: Use new context manager for testingSebastian Berg2016-09-021-16/+27
| | | | | | Making the outer context manager a suppress warnings gives good control to print warnings only once in release mode and suppress some specific warnings which cannot be easily avoided otherwise.
* TST: Use new warnings context manager in all testsSebastian Berg2016-09-0218-257/+266
| | | | | | | | In some places, just remove aparently unnecessary filters. After this, all cases of ignore filters should be removed from the tests, making testing (even multiple runs) normally fully predictable.
* ENH: Remove warning ignoring from nanfuncsSebastian Berg2016-09-023-70/+66
| | | | | | Comment mentions a speedup, but it seems unsure why it should be there. Instead use an error state in divide_by_count. Some extra complex warnings had to be ignored (but those seemed correct)
* MAINT: Add missing teardown in deprecation test caseSebastian Berg2016-09-021-0/+2
|
* BUG: Suppress common NaT warningsSebastian Berg2016-09-024-41/+61
| | | | | | | | | Printing of datetime arrays used to cause warning due to comparison warnings in NaT. This is circumvented by using views to integer values. Part of this should be simplified when the deprecation is over. Also fixes a bug with non-native byteorder.
* ENH: Make warning testing context managers more specificSebastian Berg2016-09-022-15/+8
| | | | | | | This means that warnings of different origin then the one tested for behave normally. If the normal behaviour is to igonre them this might decrease specificity in rare cases. In most cases the specificity will be slightly higher.
* MAINT: Simplify deprecation test decoratorSebastian Berg2016-09-022-15/+12
| | | | | Also modify the corresponding test to suppress the non Deprecation warnings created to test specificity.
* Merge pull request #7985 from charris/rebase-7763Charles Harris2016-08-283-39/+490
|\ | | | | Rebase 7763, ENH: Add new warning suppression/filtering context
| * STY: Small PEP8 cleanups.Charles Harris2016-08-283-2/+5
| |
| * ENH,TST: Add new warning suppression/filtering contextSebastian Berg2016-08-283-39/+487
| | | | | | | | | | | | This context has a couple of advantages over the typical one, and can ensure that warnings are cleanly filtered without side effect for later tests.
* | Merge pull request #7823 from madphysicist/ma-bugsJulian Taylor2016-08-274-52/+133
|\ \ | | | | | | BUG: Fixed masked array behavior for scalar inputs to np.ma.atleast_*d
| * | BUG: Fixed masked array behavior for scalar inputs to np.ma.atleast_*dJoseph Fox-Rabinovitz2016-08-244-52/+133
| | |
* | | DOC: Corrected allowed keywords in add_(installed_)libraryOleksandr Pavlyk2016-08-261-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Docstring listed extra_f77_compiler_args and extra_f90_compiler_args as allowed keywords, but per command/build_clib.py, lines 193-194 the recognized keywords are in fact extra_f77_compile_args and extra_f90_compile_args. The keyword extra_compiler_args, although non-confirmant both for fortran keywords and to the keywords of add_extension, is consistent with the implementation (see line 200 of build_clib.py).
* | | DOC: Fixed documented dimension of return value Joseph Fox-Rabinovitz2016-08-261-1/+1
| |/ |/| | | The first dimension of the return value depends on the degree of the polynomial, not the number of elements being used in the fit.
* | Merge pull request #7963 from charris/fix-microsoft-compilersCharles Harris2016-08-252-25/+97
|\ \ | | | | | | BUG: MSVCCompiler grows 'lib' & 'include' env strings exponentially.
| * | BUG: MSVCCompiler grows 'lib' & 'include' env strings exponentially.Charles Harris2016-08-242-25/+97
| | | | | | | | | | | | | | | | | | | | | | | | | | | Each time an MSVCCompiler was instantiated the old values of os.environ['lib'] and os.environ['include'] were concatenated to the new values set by initializing the distutils.msvc{,9}compiler.MSVCCompiler base class. Consequently when the the old and new values of those variables were the same, they doubled in size with each instantiation, leading to quickly hitting the 32,768 character limit.
* | | Merge pull request #7943 from theultimatecrouton/doc_fixCharles Harris2016-08-251-5/+2
|\ \ \ | |_|/ |/| | DOC: #7927. Remove deprecated note for memmap relevant for Python <2.5
| * | DOC: #7927. Remove deprecated note for memmap relevant for Python <2.5 [ci skip]Tom Bird2016-08-171-0/+2
| | |
| * | DOC: #7927. Remove deprecated note for memmap relevant for Python <2.5Tom Bird2016-08-171-5/+0
| | |
* | | BUG: cannot assign to tuple after usemattip2016-08-241-8/+6
| |/ |/|
* | Merge pull request #7941 from charris/fix-numpy-reload-againCharles Harris2016-08-203-53/+81
|\ \ | | | | | | BUG: Make sure numpy globals keep identity after reload.
| * | BUG: Make sure numpy globals keep identity after reload.Charles Harris2016-08-163-53/+81
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Reloading currently causes problems because global classes defined in numpy/__init__.py change their identity (a is b) after reload. The solution taken here is to move those classes to a new non-reloadable module numpy/_globals and import them into numpy from there. Classes moved are ModuleDeprecationWarning, VisibleDeprecationWarning, and _NoValue. Closes #7844.
* | | Merge pull request #7952 from charris/fix-for-setuptools_25_4Ralf Gommers2016-08-211-30/+33
|\ \ \ | | | | | | | | BUG: Use keyword arguments to initialize Extension base class.
| * | | BUG: Use keyword arguments to initialize Extension base class.Charles Harris2016-08-201-30/+33
| |/ / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently numpy.distutils.Extension class subclasses python's distutils.extension.Extension class and initializes the base class with a call that uses positional arguments rather than keyword arguments. This causes problems with setuptools 25.4.0 where the Extension class gets a new init function that expects keyword rather than positional arguments. We should have been using keyword arguments all along and our luck has run out, so use proper keywords Closes #7951.
* | | Merge pull request #7936 from goerz/sparse-type-checkStephan Hoyer2016-08-172-2/+45
|\ \ \ | |_|/ |/| | ENH: improve duck typing inside iscomplexobj
| * | ENH: improve duck typing inside iscomplexobjMichael Goerz2016-08-152-2/+45
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Both `iscomplexobj` and `isrealobj` now try to refer to the argument's `dtype` attribute if it exists. This significantly extends the list of types for which `iscomplexobj` returns correct results (including e.g. scipy sparse matrices and pandas objects). Extended the tests of the `iscomplexobj` routine for the following cases: * simple scalars * standard lists (test internal auto-conversion to numpy arrays) * "Duck typing" for objects that define a dtype attribute (either referring to one of the existing numpy dtypes, or a custom dtype, as pandas does) This fixes #7924
* | | Merge pull request #7937 from charris/fix-quicksort-for-bogus-comparisonJulian Taylor2016-08-172-4/+19
|\ \ \ | |_|/ |/| | BUG: Guard against buggy comparisons in generic quicksort.
| * | BUG: Guard against buggy comparisons in generic quicksort.Charles Harris2016-08-152-4/+19
| |/ | | | | | | | | | | | | | | | | | | Generic types may have buggy comparison operators in which case the sentinal values in quicksort cannot be counted on to keep the pointers from running off the ends of the array. The solution here is to explicitly check that the pointers are in bounds when sorting generic types. Closes #7934.
* | Merge pull request #7912 from mattip/pypy-fixes2Charles Harris2016-08-1512-92/+140
|\ \ | | | | | | ENH: skip or avoid gc/objectmodel differences btwn pypy and cpython
| * | ENH: skip or avoid gc/objectmodel differences btwn pypy and cpythonmattip2016-08-1512-92/+140
| | |
* | | Merge pull request #7931 from charris/fix-npymath-for-ppcCharles Harris2016-08-151-1/+8
|\ \ \ | | | | | | | | BUG: Check for HAVE_LDOUBLE_DOUBLE_DOUBLE_LE in npy_math_complex.
| * | | BUG: Check for HAVE_LDOUBLE_DOUBLE_DOUBLE_LE in npy_math_complex.Charles Harris2016-08-111-1/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The `_real_part_reciprocal` function is coded in two ways, one depending on functions specific to IEEE floating point and the other using generic code that should always work. Because PPC long double is not IEEE the generic version should always be chosen for that architecture, but that is currently only done when the PPC is configured as big endian. This PR makes sure that the generic version is also chosen when the PPC is configured as little endian. Closes #7836.
* | | | DOC: Update subclassing doc regarding downstream compatgfyoung2016-08-111-0/+47
| |_|/ |/| | | | | | | | [ci skip]
* | | Monkey-patch _msvccompile.gen_lib_option like any other compilatorsValentin Valls2016-08-101-1/+1
|/ / | | | | | | | | | | | | `setuptools` is now using `_msvccompile` instead of `msvccompile9`. numpy is monkey-patching symmetrically `gen_lib_options` and `spawn` function for all compilators. But `_msvccompile.gen_lib_options` was not monkey-patched while `_msvccompile.spawn` is already monkey-patched throug the super class `ccompiler.spawn`. This patch only symmetrically patch `_msvccompile` to prevent param file quoting.
* | Merge pull request #7911 from skwbc/fix_issue#7835ahaldane2016-08-082-8/+46
|\ \ | | | | | | BUG: fix for issue#7835 (ma.median of 1d)
| * | BUG: modified ndim condition to use masked_arrayskwbc2016-08-071-1/+1
| | |
| * | BUG: fix for issue#7835 (ma.median of 1d)Shota Kawabuchi2016-08-072-8/+46
| |/
* | Fixed MKL detection for recent versions of this library.Xavier Abellan Ecija2016-08-051-19/+3
|/
* ENH: added axis param for np.count_nonzerogfyoung2016-08-044-39/+198
| | | | Closes gh-391.
* BUG: fix float16 type not being called due to wrong orderingJulian Taylor2016-08-053-4/+6
| | | | closes gh-7897
* Spelling and grammar fix.Daniel Ching2016-08-041-2/+2
| | | Histogram documentation was missing an 'a' in the phrase 'all around performance'.
* DOC: Docstring capitalization/sentence fixes in numpy/ma/core.py.Charles Harris2016-08-031-7/+7
|
* BUG: construct ma.array from np.array which contains paddingAllan Haldane2016-08-032-15/+56
| | | | | | | Structured dtypes with padding bytes fail in MaskedArrays because of padding bytes in `dtype.descr`. Fix is to avoid using dtype.descr.
* Handle NetBSD specific <sys/endian.h>Kamil Rytarowski2016-07-313-1/+11
|
* Merge pull request #7871 from juliantaylor/introsort2Charles Harris2016-07-291-23/+72
|\ | | | | BUG: handle introsort depth limit properly
| * BUG: handle introsort depth limit properlyJulian Taylor2016-07-291-23/+72
| | | | | | | | | | | | | | the larger partition needs its own depth limit else it grows faster than logarithmic. Also increase the stack size to hold up to 64 instead of 50 pointer pairs else it might overflow on arrays larger than than 2^50 elements.
* | Merge pull request #7879 from dunneff/patch-1Charles Harris2016-07-281-1/+1
|\ \ | | | | | | DOC: fix typo in documentation of loadtxt (closes #7878)
| * | DOC: fix typo in documentation of loadtxt (closes #7878)dunneff2016-07-281-1/+1
| |/
* | Merge pull request #7699 from nolta/preserve-lib-orderCharles Harris2016-07-281-45/+35
|\ \ | |/ |/| BLD: preserve library order
View Lorry Controller Status