Refactor Correctors and remove statsmodels requirement - #679
Conversation
Codecov Report
@@ Coverage Diff @@
## main #679 +/- ##
==========================================
- Coverage 85.32% 85.21% -0.12%
==========================================
Files 41 41
Lines 4503 4483 -20
==========================================
- Hits 3842 3820 -22
- Misses 661 663 +2
Continue to review full report at Codecov.
|
jdkent
left a comment
There was a problem hiding this comment.
Looks good!
Main comment is I'm curious if the method parameter documentation could be defined once and referenced in other places, Don't consider it a blocking review comment, but it's nice not to have to repeat yourself, maybe rst substitutions would do what you want? https://stackoverflow.com/questions/8320909/repeated-parameters-with-python-sphinx
I think that's what nilearn does, and I certainly like it. Would you mind if I tackled that in a separate PR? I might open an issue about it. |
adelavega
left a comment
There was a problem hiding this comment.
Looks good to me, although I wonder if these non-imaging basic functions should perhaps live in pyMARE?
That seems reasonable to me, but PyMARE doesn't use these methods at the moment, so we would probably only want to add them to that package if we also create a Corrector class within PyMARE as well. |
…to drop-statsmodels
|
@adelavega I've moved the new functions to PyMARE. I'll make a release once the associated PR (neurostuff/PyMARE#87) is merged. EDIT: I've released PyMARE 0.0.4rc1 and pinned it here. |
|
I've opened an issue about the parameter definitions (#687), so I'm going to resolve the conflicts and merge this once CI passes. |
Closes #667.
Changes proposed in this pull request:
nimare.stats.fdr.MKDAChi2.correct_fwe_bhtoMKDAChi2.correct_fwe_indep(breaking change).To do: