[MRG] Fix handling of cell arrays and loading of evoked mat files when channels are not in info #7304
Conversation
thht
changed the title
Codecov Report
@@ Coverage Diff @@
## master #7304 +/- ##
==========================================
- Coverage 89.89% 89.68% -0.22%
==========================================
Files 450 450
Lines 81284 81286 +2
Branches 12920 12923 +3
==========================================
- Hits 73070 72900 -170
- Misses 5391 5526 +135
- Partials 2823 2860 +37 |
|
the azure pipelines fail with some access violation. no idea whats wrong there... |
|
@agramfort should we rerun the tests? |
|
Restarted but if it fails again you can ignore |
|
ok, codecov complaining seems to come from the fact that pandas seems to be always available. is this correct? |
|
It's not available on all builds, there is a minimal one. But that build also does not have |
|
(this is fairly common though and not worth fighting here to deal with, we can live with it) |
|
ok. so, from my side this is fine now. could you guys take a look and see if there is something i missed or still need to do? |
| @testing.requires_testing_data | ||
| # Reading the sample CNT data results in a RuntimeWarning because it cannot | ||
| # parse the measurement date. We need to ignore that warning. | ||
| @pytest.mark.filterwarnings('ignore::RuntimeWarning') |
There was a problem hiding this comment.
Could you add a suitable regex in the :: so that it's more targeted? That way other warnings that we should pay attention to won't creep in easily.
larsoner
left a comment
larsoner
left a comment
There was a problem hiding this comment.
Other than a minor ignores gripe, LGTM
|
ok, i added the chnages proposed by @larsoner and already added to one test by @agramfort to the other tests needing it. CI failures seem to not come from my code.... |
|
Thanks @thht ! |
…n channels are not in info (mne-tools#7304) * added test for cell arrays * added validation of fieldtrip structure * make pydocstyle happy * added test for bug when loading evoked mat files with missing channels * fixed test * also prune avg field * pydocstring happy * updated testing data version and hash * fixed extra space * updated whatsnew * fix typo in whatsnew * address comment from @larsoner * use more precise warning filter in all tests Co-authored-by: Alexandre Gramfort <alexandre.gramfort@m4x.org>
…n channels are not in info (mne-tools#7304) * added test for cell arrays * added validation of fieldtrip structure * make pydocstyle happy * added test for bug when loading evoked mat files with missing channels * fixed test * also prune avg field * pydocstring happy * updated testing data version and hash * fixed extra space * updated whatsnew * fix typo in whatsnew * address comment from @larsoner * use more precise warning filter in all tests Co-authored-by: Alexandre Gramfort <alexandre.gramfort@m4x.org>
3 participants
Reference issue
Fixes #7293
What does this implement/fix?
This PR fixes two bugs found when investigating #7293:
infoobject but in the loaded evoked data, the function would crash.