Two tests are currently skipped because of a problem with the random
forest training -- setting the random state does not eliminate
differences.
@DocSavage, I've shared a gist showing how I generated the test data, including the two tests that are currently being skipped. I can't figure out why the training datasets come out different between example-run.py (the gist) and test_gala.py (created in this PR), since I set the random state... Perhaps some non-numpy random number generator is used... Anyway, for now those tests are skipped.
Two tests are currently skipped because of a problem with the random forest training -- setting the random state does not eliminate differences.
@DocSavage, I've shared a gist showing how I generated the test data, including the two tests that are currently being skipped. I can't figure out why the training datasets come out different between example-run.py (the gist) and
test_gala.py
(created in this PR), since I set the random state... Perhaps some non-numpy random number generator is used... Anyway, for now those tests are skipped.