Avoiding Pool in get_cross_nmi as it leads to hanging processes#259
Open
gbrunin wants to merge 5 commits intoppdebreuck:masterfrom
Open
Avoiding Pool in get_cross_nmi as it leads to hanging processes#259gbrunin wants to merge 5 commits intoppdebreuck:masterfrom
gbrunin wants to merge 5 commits intoppdebreuck:masterfrom
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Hi,
For quite some time now I have faced the issue that feature selection gets stuck around the end of the process because some MI computations fail silently and are not handled properly by
pool.imap_unordered. This could be somehow mitigated by putting everything into amain()method but in some cases it remained unavoidable, making life more difficult than it should be.In this PR, I suggest to use
concurrent.futuresto handle the parallelization inget_cross_nmi. This allows to handle edge cases more cleanly. Here, when the MI computation fails, I set the mutual_information to 0.I have not modified anything else than the cross-MI computation. Maybe this switch from multiprocessing to concurrent should be propagated to other parts of the code? Let me know what you think.
Tests are passing without modification.
Cheers,
Guillaume