Source: spectral-cube
Version: 0.6.5-6
Severity: serious
Tags: patch
Justification: ftbfs
Forwarded: https://github.com/radio-astro-tools/spectral-cube/issues/935
spectral-cube FTBFS in unstable due to a change somewhere between
python3-dask 2024.5.2+dfsg-1 and 2024.12.1+dfsg-1. (This dask upload
and the associated dask.distributed upload fix four other RC bugs
between them, so I'm keen to get them into testing.) Here's the test
log:
I: pybuild base:311: cd
/build/reproducible-path/spectral-cube-0.6.5/.pybuild/cpython3_3.13_spectral-cube/build;
python3.13 -m pytest
============================= test session starts
==============================
platform linux -- Python 3.13.1, pytest-8.3.4, pluggy-1.5.0
rootdir: /build/reproducible-path/spectral-cube-0.6.5
configfile: setup.cfg
plugins: doctestplus-1.3.0, astropy-header-0.2.2, arraydiff-0.6.1,
astropy-0.11.0, remotedata-0.4.1, mock-3.14.0, filter-subpackage-0.2.0,
typeguard-4.4.1, hypothesis-6.122.1, cov-5.0.0
collected 1719 items
spectral_cube/spectral_axis.py . [
0%]
spectral_cube/tests/test_analysis_functions.py ................... [
1%]
spectral_cube/tests/test_casafuncs.py ......sssssssssssssss [
2%]
spectral_cube/tests/test_cube_utils.py ....... [
2%]
spectral_cube/tests/test_dask.py .s..s.......s [
3%]
spectral_cube/tests/test_io.py ....................... [
4%]
spectral_cube/tests/test_masks.py ...................................... [
7%]
...............................................XX.... [
10%]
spectral_cube/tests/test_moments.py .................................... [
12%]
........................................................................ [
16%]
........................................................................ [
20%]
........ [
21%]
spectral_cube/tests/test_performance.py ...s.. [
21%]
spectral_cube/tests/test_projection.py .......................x......... [
23%]
.....xxx.......x........................................................ [
27%]
.................. [
28%]
spectral_cube/tests/test_regrid.py ..........ssss....s.................s [
30%]
ss....sssssssssssss [
31%]
spectral_cube/tests/test_spectral_axis.py .............................. [
33%]
...................... [
34%]
spectral_cube/tests/test_spectral_cube.py .............................. [
36%]
........................................................................ [
40%]
........................................................................ [
45%]
........................................................................ [
49%]
..............................s.....s...........sssss................... [
53%]
..............................ssssss......................ss............ [
57%]
........................................................................ [
61%]
......................................ssssss......ssssssssssss......ssss [
65%]
ssssssss......ssssssssssss......ssssssssssss......ssssss................ [
70%]
........................................................................ [
74%]
..ssssss......ssssssssssss......ssssssssssss......ssssssssssss......ssss [
78%]
ssssssss......ssssss.................................s.s.F............xx [
82%]
xx....................................................................ss [
86%]
ss............................xxxxxxxx..............sss................. [
91%]
....... [
91%]
spectral_cube/tests/test_stokes_spectral_cube.py ....................... [
92%]
........................................................................ [
97%]
..... [
97%]
spectral_cube/tests/test_subcubes.py ..ssssssssssssssssssss [
98%]
spectral_cube/tests/test_visualization.py ....ssssssss [
99%]
spectral_cube/tests/test_wcs_utils.py ............
[100%]
=================================== FAILURES
===================================
________________________ test_oned_collapse_beams[True]
________________________
data_sdav_beams =
PosixPath('/tmp/pytest-of-cjwatson/pytest-0/test_oned_collapse_beams_True_0/sdav_beams.fits')
use_dask = True
def test_oned_collapse_beams(data_sdav_beams, use_dask):
# Check that an operation along the spatial dims returns an
appropriate
# spectrum
cube, data = cube_and_raw(data_sdav_beams, use_dask=use_dask)
cube._meta['BUNIT'] = 'K'
cube._unit = u.K
spec = cube.mean(axis=(1,2))
assert isinstance(spec, VaryingResolutionOneDSpectrum)
# data has a redundant 1st axis
> np.testing.assert_equal(spec.value, data.mean(axis=(1,2,3)))
spectral_cube/tests/test_spectral_cube.py:1531:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_
args = (<built-in function eq>, array([0.49465334, 0.53732362, 0.37316832,
0.35550457]), array([0.49465334, 0.53732362, 0.37316832, 0.35550457]))
kwds = {'err_msg': '', 'header': 'Arrays are not equal', 'strict': False,
'verbose': True}
@wraps(func)
def inner(*args, **kwds):
with self._recreate_cm():
> return func(*args, **kwds)
E AssertionError:
E Arrays are not equal
E
E Mismatched elements: 1 / 4 (25%)
E Max absolute difference: 1.11022302e-16
E Max relative difference: 2.06620925e-16
E x: array([0.494653, 0.537324, 0.373168, 0.355505])
E y: array([0.494653, 0.537324, 0.373168, 0.355505])
/usr/lib/python3.13/contextlib.py:85: AssertionError
----------------------------- Captured stderr call
-----------------------------
WARNING: StokesWarning: Cube is a Stokes cube, returning spectral cube for I
component [spectral_cube.io.core]
------------------------------ Captured log call
-------------------------------
WARNING astropy:logger.py:244 StokesWarning: Cube is a Stokes cube,
returning spectral cube for I component
=========================== short test summary info
============================
FAILED
spectral_cube/tests/test_spectral_cube.py::test_oned_collapse_beams[True]
= 1 failed, 1487 passed, 212 skipped, 17 xfailed, 2 xpassed in 257.65s
(0:04:17) =
E: pybuild pybuild:389: test: plugin distutils failed with: exit code=1: cd
/build/reproducible-path/spectral-cube-0.6.5/.pybuild/cpython3_3.13_spectral-cube/build;
python3.13 -m pytest
(3.12 behaves the same way; I've omitted it for brevity.)
This looks like the first issue in
https://github.com/radio-astro-tools/spectral-cube/issues/935. I tried
cherry-picking a couple of patches from there, and that fixed it. Would
you consider doing the same in unstable? I've attached a possible
debdiff.
Thanks,
--
Colin Watson (he/him) [[email protected]]
diff -Nru spectral-cube-0.6.5/debian/changelog
spectral-cube-0.6.5/debian/changelog
--- spectral-cube-0.6.5/debian/changelog 2024-12-17 07:12:11.000000000
+0000
+++ spectral-cube-0.6.5/debian/changelog 2025-01-07 14:19:23.000000000
+0000
@@ -1,3 +1,10 @@
+spectral-cube (0.6.5-6.1) UNRELEASED; urgency=medium
+
+ * Non-maintainer upload.
+ * mean -> nanmean in failing test (closes: #-1).
+
+ -- Colin Watson <[email protected]> Tue, 07 Jan 2025 14:19:23 +0000
+
spectral-cube (0.6.5-6) unstable; urgency=medium
* Remove six dependency (Closes: #1090171)
diff -Nru spectral-cube-0.6.5/debian/patches/series
spectral-cube-0.6.5/debian/patches/series
--- spectral-cube-0.6.5/debian/patches/series 2024-12-17 07:11:57.000000000
+0000
+++ spectral-cube-0.6.5/debian/patches/series 2025-01-07 14:19:23.000000000
+0000
@@ -6,3 +6,4 @@
Don-t-run-LMV-test-on-big-endian-archs.patch
Fix-test-for-Astropy-7.patch
Remove-six-dependency.patch
+test-nanmean.patch
diff -Nru spectral-cube-0.6.5/debian/patches/test-nanmean.patch
spectral-cube-0.6.5/debian/patches/test-nanmean.patch
--- spectral-cube-0.6.5/debian/patches/test-nanmean.patch 1970-01-01
01:00:00.000000000 +0100
+++ spectral-cube-0.6.5/debian/patches/test-nanmean.patch 2025-01-07
14:19:23.000000000 +0000
@@ -0,0 +1,25 @@
+From: "Adam Ginsburg (keflavich)" <[email protected]>
+Date: Mon, 6 Jan 2025 11:13:06 -0500
+Subject: mean -> nanmean in failing test
+
+Origin: other,
https://github.com/radio-astro-tools/spectral-cube/pull/929/commits/63d92273e7f423afd1f9080db93a95300a9a5064
+Origin: other,
https://github.com/radio-astro-tools/spectral-cube/pull/929/commits/beb1233fb51cbbf8c3199a72512a37971f25b1d2
+Last-Update: 2025-01-07
+---
+ spectral_cube/tests/test_spectral_cube.py | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/spectral_cube/tests/test_spectral_cube.py
b/spectral_cube/tests/test_spectral_cube.py
+index bbc1aaa..4ccd135 100644
+--- a/spectral_cube/tests/test_spectral_cube.py
++++ b/spectral_cube/tests/test_spectral_cube.py
+@@ -1528,7 +1528,8 @@ def test_oned_collapse_beams(data_sdav_beams, use_dask):
+ spec = cube.mean(axis=(1,2))
+ assert isinstance(spec, VaryingResolutionOneDSpectrum)
+ # data has a redundant 1st axis
+- np.testing.assert_equal(spec.value, data.mean(axis=(1,2,3)))
++ # we changed to assert_almost_equal in 2025 because, for no known reason,
epsilon-level differences crept in
++ np.testing.assert_almost_equal(spec.value, np.nanmean(data, axis=(1,2,3)))
+ assert cube.unit == spec.unit
+ assert spec.header['BUNIT'] == cube.header['BUNIT']
+
