Codecov Report

Patch coverage: 100.00% and project coverage change: +0.03 🎉

Comparison is base (559bef4) 93.18% compared to head (cd038ef) 93.21%.

@@            Coverage Diff             @@
##             main     #603      +/-   ##
==========================================
+ Coverage   93.18%   93.21%   +0.03%     
==========================================
  Files          32       32              
  Lines        4842     4851       +9     
==========================================
+ Hits         4512     4522      +10     
+ Misses        330      329       -1     

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

@tamarl08 , @damien2012eng I have made a bunch of changes.

1. I noticed that the tables with the Shap values could not be sorted because of the index label. We remove this label in rsmtool tables so I did the same here.
2. For consistency, I upper-cased HAS_SINGLE_EXAMPLE for the variable passed as environment variable to the notebook. I also simplified the computation of has_single_example.
3. I tweaked the plots notebook to:
   • clearly show the plots that are unavailable instead of hiding them.
   • fix the weird matplotlib message being shown for waterfall plot.
4. I fixed the outdated tutorial documentation and added some useful links.

Can you both please review the changes and also test the single-example and multiple-example runs to make sure that everything looks okay in the notebook?

@damien2012eng can you also double check that when choosing the examples via sample_ids, the correct examples are being explained?

Hi @desilinguist I have run experiments with both single and multiple responses, and the reports look good to me. I also provided multiple ids, and then examined both the report and csv files, and confirmed that the correct responses had been explained.

Thanks for the improvement!