New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
API: Multi-output summary plot legend #3062
API: Multi-output summary plot legend #3062
Conversation
Hi @101AlexMartin , thanks for the PR! A couple of things before we kick off the PR review:
|
Hello! I implemented your proposed changes :) |
Hi @101AlexMartin , next steps would be to make sure the tests pass :) In particular for plotting tests, we are using pytest-matplotlib. You will need to generate the baseline images and commit that into the repository. This way we can ensure that the function always generates the same-looking plot. |
I added the tests and the baseline figures :) |
Codecov Report
@@ Coverage Diff @@
## master #3062 +/- ##
==========================================
+ Coverage 55.33% 55.35% +0.02%
==========================================
Files 90 90
Lines 12875 12883 +8
==========================================
+ Hits 7124 7132 +8
Misses 5751 5751
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
tests/plots/baseline/test_random_multi_class_summary_legend.png
Outdated
Show resolved
Hide resolved
What is the status of this then? |
@101AlexMartin Sorry, let me get back to you over the weekend. Bit busy at work this week. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi @101AlexMartin Sorry this review took so long! This should be the last round, I just have a couple more suggestions for you to make.
Done :) |
Just a heads up, the 3.11 tests are failing, not anything to do with your PR. Probably due to the 4.0 lightgbm release. #3092 This PR looks good to me, I'll come back to this once we get our builds running again. Hang tight! |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Overview
Added flag to allow the printing of the mean SHAP values in the legend of a multi-output summary bar plot. The motivation behind this change is to easily compare rankings when different chunks of data are used to get the SHAP values. Comparing visually the overall ranking in a multi-output problem was quite challenging, so this pull request solves the problem.
Checklist
CHANGELOG.md
(if changes will affect users)