Accepting request 1169895 from devel:languages:python:numeric

OBS-URL: https://build.opensuse.org/request/show/1169895
OBS-URL: https://build.opensuse.org/package/show/openSUSE:Factory/python-numexpr?expand=0&rev=24
This commit is contained in:
Ana Guerrero 2024-04-23 16:57:43 +00:00 committed by Git OBS Bridge
commit 1b971d956c
5 changed files with 58 additions and 7 deletions

3
numexpr-2.10.0.tar.gz Normal file
View File

@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c89e930752639df040539160326d8f99a84159bbea41943ab8e960591edaaef0
size 103563

View File

@ -1,3 +0,0 @@
version https://git-lfs.github.com/spec/v1
oid sha256:f21d12f6c432ce349089eb95342babf6629aebb3fddf187a4492d3aadaadaaf0
size 103155

View File

@ -1,3 +1,20 @@
-------------------------------------------------------------------
Mon Apr 22 09:57:01 UTC 2024 - Markéta Machová <mmachova@suse.com>
- Add patch revert-to-numpy1.patch to restore compatibility with numpy v1
* revert of upstream commit, drop when numpy v2 comes to Factory
-------------------------------------------------------------------
Fri Apr 19 11:57:33 UTC 2024 - Ben Greiner <code@bnavigator.de>
- Update to 2.10.0
* Support for NumPy 2.0.0. This is still experimental, so please
report any issues you find. Thanks to Clément Robert and Thomas
Caswell for the work.
* Avoid erroring when OMP_NUM_THREADS is empty string. Thanks to
Patrick Hoefler.
* Do not warn if OMP_NUM_THREAD set.
-------------------------------------------------------------------
Tue Feb 6 09:32:12 UTC 2024 - Dirk Müller <dmueller@suse.com>

View File

@ -18,22 +18,25 @@
%{?sle15_python_module_pythons}
Name: python-numexpr
Version: 2.9.0
Version: 2.10.0
Release: 0
Summary: Numerical expression evaluator for NumPy
License: MIT
Group: Development/Languages/Python
URL: https://github.com/pydata/numexpr/
Source: https://files.pythonhosted.org/packages/source/n/numexpr/numexpr-%{version}.tar.gz
# PATCH-FIX-UPSTREAM revert of https://github.com/pydata/numexpr/commit/beedecb3990d604d0f272c1a2f9a1d117f6120ba to enable numpy v1 again, drop when numpy is updated to v2
Patch0: revert-to-numpy1.patch
BuildRequires: %{python_module devel >= 3.7}
BuildRequires: %{python_module numpy-devel >= 1.13.3}
# Until numpy 2 is in Factory, keep allowing to build with numpy 1.x. See comments in pyproject.toml for scipy and pandas
BuildRequires: %{python_module numpy-devel >= 1.19.3}
BuildRequires: %{python_module pip}
BuildRequires: %{python_module setuptools}
BuildRequires: %{python_module wheel}
BuildRequires: fdupes
BuildRequires: gcc-c++
BuildRequires: python-rpm-macros
Requires: python-numpy >= 1.13.3
Requires: python-numpy >= 1.19.3
%python_subpackages
%description
@ -43,7 +46,7 @@ can be accelerated and use less memory than doing the same
calculation in Python.
%prep
%setup -q -n numexpr-%{version}
%autosetup -p1 -n numexpr-%{version}
# wrong-file-end-of-line-encoding
sed -i 's/\r$//' ANNOUNCE.rst AUTHORS.txt README.rst RELEASE_NOTES.rst site.cfg.example
# remove unwanted shebang

31
revert-to-numpy1.patch Normal file
View File

@ -0,0 +1,31 @@
From beedecb3990d604d0f272c1a2f9a1d117f6120ba Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Cl=C3=A9ment=20Robert?= <cr52@protonmail.com>
Date: Mon, 1 Apr 2024 14:43:52 +0200
Subject: Revert: [PATCH] BLD: fix compiling against numpy 2.0
---
numexpr/interpreter.cpp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
Index: numexpr-2.10.0/numexpr/interpreter.cpp
===================================================================
--- numexpr-2.10.0.orig/numexpr/interpreter.cpp
+++ numexpr-2.10.0/numexpr/interpreter.cpp
@@ -1203,7 +1203,7 @@ NumExpr_run(NumExprObject *self, PyObjec
Py_INCREF(dtypes[0]);
} else { // constant, like in '"foo"'
dtypes[0] = PyArray_DescrNewFromType(NPY_STRING);
- PyDataType_SET_ELSIZE(dtypes[0], (npy_intp)self->memsizes[1]);
+ dtypes[0]->elsize = (int)self->memsizes[1];
} // no string temporaries, so no third case
}
if (dtypes[0] == NULL) {
@@ -1449,7 +1449,7 @@ NumExpr_run(NumExprObject *self, PyObjec
/* Get the sizes of all the operands */
dtypes_tmp = NpyIter_GetDescrArray(iter);
for (i = 0; i < n_inputs+1; ++i) {
- self->memsizes[i] = PyDataType_ELSIZE(dtypes_tmp[i]);
+ self->memsizes[i] = dtypes_tmp[i]->elsize;
}
/* For small calculations, just use 1 thread */