Filter*ByAlignmentConfidenceJob: better alignment plots
#610
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
This was probably enforced to deal with huge outliers, but they can't happen anymore (depending on the model) because the user has a choice to deal with the alignments that haven't finished. The plot is useless when dealing with values outside of that range, for instance when aligning with NNs (negative alignment score)
Icemole
requested review from
albertz,
curufinwe,
hannah220,
kuacakuaca,
michelwi,
moothiringote and
sarahberanek
|
Do we not have a separate Job to plot alignment scores that is much more versatile and configurable? Line 429 in b19149e So instead of changing the suboptimal plotting function of the filter jobs to something slightly-less-but-still-not-so-good , I would just disregard the plots they produce and use the dedicated plotting job..? Maybe instead if this PR, add a parameter |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
The alignment plots were previously useless when dealing with values outside of the disposed range, such as when aligning with NNs (negative alignment scores). Now the X axis range entirely depends on the distribution.
Also added a custom plot for the recording alignment scores, since the Y axis label was wrong (it said "segments" instead of "recordings").
Notes:
remove_dnf_alignments=True(which is the recommended value anyway, and only set toFalsefor retrocompatibility purposes), the graph might look bad (e.g. X axis extending from 0 to 1e+36, one big bar at [0, 200] and then nothing else). However, since I use NN-based alignments, it does already look bad to me for most of my filterings (just one histogram bar at 0 and then an empty graph), so I believe this would already be an improvement. To circumvent this, the user should setremove_dnf_alignments=True.remove_dnf_alignments=Trueby default always. The alignments that did not finish aren't in the alignment caches anyway, so these being included here only pollutes the plot. What do you think?