rpm/python-sherpa
SHA256
1
0
Fork 0
forked from pool/python-sherpa
Code Pull Requests Activity

Accepting request 1109765 from devel:languages:python:numeric

Browse Source 
OBS-URL: https://build.opensuse.org/request/show/1109765
OBS-URL: https://build.opensuse.org/package/show/openSUSE:Factory/python-sherpa?expand=0&rev=17
This commit is contained in:
Ana Guerrero 2023-09-10 11:10:01 +00:00 committed by Git OBS Bridge
commit 66ebaa0d74
4 changed files with 195 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

96
numpy125-CI.patch Normal file
View File

@ -0,0 +1,96 @@
From 2aea86a2a596323e9c8e41275acbffa1e9a609a0 Mon Sep 17 00:00:00 2001
From: Doug Burke <dburke@cfa.harvard.edu>
Date: Mon, 17 Jul 2023 11:33:29 -0400
Subject: [PATCH 1/2] Silence test warnings if pytest-doctestplus is not
installed
I originally tried adding the warnings to sherpa/conftest.py but
this is happening in pytest itself, so I've taken advantage of
the general filter-warnings capability to hide thse messages.
Before this change, running pytest without pytest-doctestplus would
result in messages like
==================================================== warnings summary =====================================================
../../../lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373
/lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373: PytestConfigWarning: Unknown config option: doctest_norecursedirs
self._warn_or_fail_if_strict(f"Unknown config option: {key}\n")
../../../lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373
/lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373: PytestConfigWarning: Unknown config option: doctest_plus
self._warn_or_fail_if_strict(f"Unknown config option: {key}\n")
../../../lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373
/lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373: PytestConfigWarning: Unknown config option: doctest_plus_atol
self._warn_or_fail_if_strict(f"Unknown config option: {key}\n")
../../../lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373
/lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373: PytestConfigWarning: Unknown config option: doctest_subpackage_requires
self._warn_or_fail_if_strict(f"Unknown config option: {key}\n")
../../../lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373
/lagado2.real/local/anaconda/envs/sherpa-main/lib/python3.11/site-packages/_pytest/config/__init__.py:1373: PytestConfigWarning: Unknown config option: text_file_format
self._warn_or_fail_if_strict(f"Unknown config option: {key}\n")
-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
================================================= short test summary info =================================================
---
pytest.ini | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/pytest.ini b/pytest.ini
index fde2ab8e0c..e2c4c4847e 100644
--- a/pytest.ini
+++ b/pytest.ini
@@ -2,6 +2,10 @@
addopts = -rs --ignore=setup.py --ignore=test_requirements.txt
norecursedirs = .git build dist tmp* .eggs
text_file_format = rst
+filterwarnings =
+ # Since we can runpytest without loading pytest-doctestplus, hide
+ # the warnings we get when this is done
+ ignore::pytest.PytestConfigWarning
doctest_plus = enabled
doctest_plus_atol = 1e-4
doctest_optionflags =
From 14e43a098ecb9a5068d2886f6d4680b9cd878404 Mon Sep 17 00:00:00 2001
From: Doug Burke <dburke@cfa.harvard.edu>
Date: Mon, 17 Jul 2023 12:05:21 -0400
Subject: [PATCH 2/2] Tests: hide NumPy 1.25 ndim > 0 warnings (crates)
Although Sherpa has been updated to remove the deprecation
warnnng from NumPy 1.25:
Conversion of an array with ndim > 0 to a scalar is deprecated,
and will error in future. Ensure you extract a single element
from your array before performing this operation.
(Deprecated NumPy 1.25.)
it is still present in Crates (mid 2023), so hide the warning with
the assumption we can revert this commit by mid December 2023.
---
sherpa/conftest.py | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/sherpa/conftest.py b/sherpa/conftest.py
index 97979b4e4b..339f252a51 100644
--- a/sherpa/conftest.py
+++ b/sherpa/conftest.py
@@ -117,6 +117,11 @@ def pytest_collection_modifyitems(config, items):
r'np.asscalar\(a\) is deprecated since NumPy v1.16, use a.item\(\) instead',
r"Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated since Python 3.3,and in 3.9 it will stop working",
+ # NumPy 1.25 warnings that are raised by (mid-2023) crates code.
+ # Hopefully this can be removed by December 2023.
+ #
+ r"Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. \(Deprecated NumPy 1.25.\)",
+
],
UserWarning:
[

89
numpy125.patch Normal file
View File

@ -0,0 +1,89 @@
From 740fcd0da87d3aefa41aa3962cc8f1ddb7ab3cab Mon Sep 17 00:00:00 2001
From: Doug Burke <dburke@cfa.harvard.edu>
Date: Tue, 20 Jun 2023 15:31:55 -0400
Subject: [PATCH] NumPy 1.25 support
NumPy 1.25 deprecates converting a ndarray that is not a 0d value,
so for our cases this is
int(value)
info("message %g", value)
where value is a ndarray of length 1. The simple fix is to explictly
access the first element - so
answer = int(value[0])
info("message %g", value[0])
although for the int case (in the PSF convolution instrument code) I
have not looked through to determine if value is always guaranteed to
be a ndarray, so it is more like
if np.isscalar(value):
newval = value
else:
newval = value[0]
answer = int(newval)
---
sherpa/fit.py | 13 +++++++++----
sherpa/instrument.py | 12 ++++++++++--
2 files changed, 19 insertions(+), 6 deletions(-)
Index: sherpa-4.15.1/sherpa/fit.py
===================================================================
--- sherpa-4.15.1.orig/sherpa/fit.py
+++ sherpa-4.15.1/sherpa/fit.py
@@ -1258,8 +1258,13 @@ class Fit(NoNewAttributesAfterInit):
def get_par_name(ii):
return self.model.pars[self.thaw_indices[ii]].fullname
- # Call from a parameter estimation method, to report
- # that limits for a given parameter have been found
+ # Call from a parameter estimation method, to report that
+ # limits for a given parameter have been found At present (mid
+ # 2023) it looks like lower/upper are both single-element
+ # ndarrays, hence the need to convert to a scalar by accessing
+ # the first element (otherwise there's a deprecation warning
+ # from NumPy 1.25).
+ #
def report_progress(i, lower, upper):
if i < 0:
pass
@@ -1268,11 +1273,11 @@ class Fit(NoNewAttributesAfterInit):
if isnan(lower) or isinf(lower):
info("%s \tlower bound: -----" % name)
else:
- info("%s \tlower bound: %g" % (name, lower))
+ info("%s \tlower bound: %g" % (name, lower[0]))
if isnan(upper) or isinf(upper):
info("%s \tupper bound: -----" % name)
else:
- info("%s \tupper bound: %g" % (name, upper))
+ info("%s \tupper bound: %g" % (name, upper[0]))
# If starting fit statistic is chi-squared or C-stat,
# can calculate reduced fit statistic -- if it is
Index: sherpa-4.15.1/sherpa/instrument.py
===================================================================
--- sherpa-4.15.1.orig/sherpa/instrument.py
+++ sherpa-4.15.1/sherpa/instrument.py
@@ -357,9 +357,17 @@ class PSFKernel(Kernel):
# and Python 3.8 - causes a TypeError with the message
# "only integer scalar arrays can be converted to a scalar index"
# to be thrown here if sent directly to set_origin. So
- # we convert to a Python integer type.
+ # we convert to a Python integer type. In NumPy 1.25 it became
+ # a deprecation error to call int on an array with ndim > 0.
#
- origin = set_origin(kshape, int(brightPixel))
+ # assume there is only one element in brightPixel if not
+ # a scalar
+ #
+ if not numpy.isscalar(brightPixel):
+ loc = brightPixel[0]
+ else:
+ loc = brightPixel
+ origin = set_origin(kshape, int(loc))
if self.origin is None:
self.origin = origin

5
python-sherpa.changes
View File

@ -1,3 +1,8 @@
-------------------------------------------------------------------
Thu Sep 7 12:24:59 UTC 2023 - Markéta Machová <mmachova@suse.com>
- Add upstream patches numpy125.patch and numpy125-CI.patch
-------------------------------------------------------------------
Mon May 29 13:52:34 UTC 2023 - Dirk Müller <dmueller@suse.com>

6
python-sherpa.spec
View File

@ -24,6 +24,10 @@ License: GPL-3.0-only
URL: https://github.com/sherpa/sherpa/
Source0: https://github.com/sherpa/sherpa/archive/%{version}.tar.gz#/sherpa-%{version}.tar.gz
Source1: https://github.com/sherpa/sherpa-test-data/archive/refs/tags/%{version}.tar.gz#/sherpa-test-data-%{version}.tar.gz
# PATCH-FIX-UPSTREAM https://github.com/sherpa/sherpa/pull/1807 NumPy 1.25 support
Patch: numpy125.patch
# PATCH-FIX-UPSTREAM https://github.com/sherpa/sherpa/pull/1819 CI: hide NumPy 1.25 array ndim>0 deprecation warnings
Patch: numpy125-CI.patch
BuildRequires: %{python_module devel >= 3.8}
BuildRequires: %{python_module numpy-devel >= 1.19}
BuildRequires: %{python_module pip}
@ -54,7 +58,7 @@ user to construct models from definitions and fit those models to
data, using a variety of statistics and optimization methods.
%prep
%setup -q -n sherpa-%{version} -a1
%autosetup -p1 -n sherpa-%{version} -a1
# uncomment system libs https://sherpa.readthedocs.io/en/latest/install.html#fftw
sed -i "s|#fftw=local|fftw=local|" setup.cfg
sed -i "s|#fftw-include[-_]dirs.*$|fftw-include-dirs=%{_includedir}|" setup.cfg