-
Notifications
You must be signed in to change notification settings - Fork 3.3k
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
Enh: deprecate legacy plots and code #3209
base: master
Are you sure you want to change the base?
Conversation
511960b
to
a04d9a5
Compare
11a793e
to
5c994fd
Compare
Maybe update the message to say something like "will be removed in 0.44"? |
Codecov ReportPatch coverage:
Additional details and impacted files@@ Coverage Diff @@
## master #3209 /- ##
==========================================
Coverage 57.58% 57.70% 0.11%
==========================================
Files 88 88
Lines 12511 12521 10
==========================================
Hits 7205 7225 20
Misses 5306 5296 -10
☔ View full report in Codecov by Sentry. |
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.
Few suggestions (up for discussion, except the last one):
- Use the phrase "marked for deprecation" instead of "deprecated". I've always found it confusing myself, but maybe it's just me.
- Maybe
shap v0.43.0
instead ofVersion 0.43.0
. I'm thinking to specify "shap" because there might be wrapper libraries around shap, and I'm afraid end users will confuse 0.43 for the wrapper library's version rather than shap's. - Mention what the alternative should be. "Please use
shap.plots.bar
instead" or words to that effect. For summary_legacy, maybe just nudge the user towards using the dedicated shap.plots functions. - The docstrings of the legacy functions should be updated as well with the deprecation message & reference to the suggested alternative, in a deprecation admonition / warning box.
Overview
Part of #3156
Description of the changes proposed in this pull request:
Potentially could create own decorator using the python warnings module.
Did not include unit tests for bar_legacy as function is breaking outside of decorator.
Checklist