Closed aewallwi closed 3 years ago
Merging #112 (ee8bf61) into main (01d0e32) will increase coverage by
0.23%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## main #112 +/- ##
==========================================
+ Coverage 86.84% 87.08% +0.23%
==========================================
Files 7 7
Lines 1460 1487 +27
==========================================
+ Hits 1268 1295 +27
Misses 192 192
Flag | Coverage Δ | |
---|---|---|
unittests | 87.08% <100.00%> (+0.23%) |
:arrow_up: |
Flags with carried forward coverage won't be shown. Click here to find out more.
Impacted Files | Coverage Δ | |
---|---|---|
uvtools/dspec.py | 96.53% <100.00%> (+0.11%) |
:arrow_up: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 01d0e32...ee8bf61. Read the comment docs.
Thanks for the review @r-pascua, especially the helpful pytest tips. I've restricted this PR to only address the 1d gridding stuff. I've moved the remainder of pytest migration to #116 (which is translating the existing tests verbatum into pytest compatible lines and tweaking the imports and test dir structure to accomodate codecov). I've moved the suggestions on reorganizing the existing tests (which is beyond the verbatum pytest translation) into their own issues which can be addressed in #117 . I think the gridding specific stuff is ready for another look when you have time! After that, I think its worth taking a look at #116 focusing on issues with the verbatum migration. We can take care of the test reorg in #117.
Technically depends on #114 and #113. Reviewer should focuse on changes to
dspec.py
andtest_dspec.py
. This PR allows _fit_basis_1d to handle missing integrations by replacing them with flagged zeros on a uniform grid. Missing data causes the x-axis to not be uniformly spaced but it is spaced by integer multiples of a fundamental dx so we can fill in these gaps with flagged weight zero data to make the basis fitting behave and then throw away these points when we return the model fits.