kz930 opened a new pull request, #8765: URL: https://github.com/apache/texera/pull/8765
### What changes were proposed in this PR? The Machine Learning Scorer's schema check let any two columns of the same type through as a classification pair. For BINARY and LARGE_BINARY pairs, every metric then failed inside scikit-learn. For TIMESTAMP pairs, Precision, Recall and F1 failed the same way. The error did not name either column. The check now rejects these pairs and names the two columns and the metrics that cannot score them. A TIMESTAMP pair scored by Accuracy alone is still accepted, since scikit-learn scores it (it only counts equal rows). ### Any related issues, documentation, discussions? Closes #8756. ### How was this PR tested? Three new cases in `MachineLearningScorerOpDescSpec`: - BINARY and LARGE_BINARY pairs are rejected. - A TIMESTAMP pair is rejected for F1 and not for Accuracy. - A TIMESTAMP pair with Accuracy alone passes. The first two fail without the change. The runtime failures were reproduced first by running the operator's generated metric code under scikit-learn 1.7.2 on each column type. ### Was this PR authored or co-authored using generative AI tooling? Generated-by: Claude Code (Claude Opus 5.5) 🤖 Generated with [Claude Code](https://claude.com/claude-code) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
