Fix/supress warnings in reader tests #2689
Conversation
pnuu
added
component:readers
component:tests
cleanup
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## main #2689 +/- ##
==========================================
+ Coverage 95.32% 95.34% +0.02%
==========================================
Files 371 371
Lines 52441 52379 -62
==========================================
- Hits 49988 49943 -45
+ Misses 2453 2436 -17
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
Pull Request Test Coverage Report for Build 7212382784Warning: This coverage report may be inaccurate.We've detected an issue with your CI configuration that might affect the accuracy of this pull request's coverage report.
💛 - Coveralls |
|
Before this PR there were 472 warnings triggered by the reader tests, now there are 68:
I'll remove the one |
pnuu
requested review from
sfinkens,
mraspaud,
adybbroe and
djhoese
as code owners
| with warnings.catch_warnings(): | ||
| warnings.filterwarnings("ignore", | ||
| message=r"You will likely lose important projection information", | ||
| category=UserWarning) | ||
| proj_dict = area_def.proj_dict |
There was a problem hiding this comment.
I think the warning would disappear if the crs was passed instead of the proj dict, as proj4_radius_parameters accepts also CRSs: https://github.com/pytroll/pyresample/blob/main/pyresample/utils/proj4.py#L76-L91
There was a problem hiding this comment.
It seems I could get all the tested things from area_def.crs.to_cf() apart from the units. There is LENGTHUNIT["metre",...] in crs_wkt, but checking for that seems like a cludge. So it's either using proj_dict or doing "metre" in area_def.crs.to_cf()['crs_wkt'] 🤔
There was a problem hiding this comment.
Maybe crs.axis_info[0].unit_name?
There was a problem hiding this comment.
But how about creating an expected crs instead, and comparing it to area_def.crs?
eg:
expected_crs = CRS.from_dict(dict(a=6378137.0, b=6356752.3, h=35785863.0, lon_0=140.7, proj="geos", units="m"))
assert area_def.crs == expected_crsThere was a problem hiding this comment.
Too simple 🙈 Adjusted in 24422ad
There was a problem hiding this comment.
Ok, I didn't comment on the other ones, but there are a bunch other places where this could be used instead of suppressing the warning...
There was a problem hiding this comment.
I tried to go the "do not suppress" route, but evidently I missed this way. I'll start combing through the changes for these.
There was a problem hiding this comment.
I updated the proj_dict usages to CRS as much as possible. The remaining are used either because the original projection needs to be modified (in nwcsaf_nc) and the dict is just too simple not to be used, or because the CRS objects differ even if the proj dicts are identical (oceancolorcci_l3_nc).
This PR fixes and/or suppresses a bunch of warnings emitted by the Satpy reader tests.