[Bug]: beam_PreCommit_Python_ML fails with AttributeError

[yuwtennis]

What happened?

Problem

I have been working on PR #36932 and experienced error in github actions.
So I have created as an separated issue.

Github Action for beam_PreCommit_Python_ML is failing due to AttributeError.
Below is the snippet from python version 3.11 .

AttributeError: The following error was raised: 'IForest' object has no attribute '__sklearn_tags__'. It seems that there are no classes that implement `__sklearn_tags__` in the MRO and/or all classes in the MRO call `super().__sklearn_tags__()`. Make sure to inherit from `BaseEstimator` which implements `__sklearn_tags__` (or alternatively define `__sklearn_tags__` but we don't recommend this approach). Note that `BaseEstimator` needs to be on the right side of other Mixins in the inheritance order. [while running 'AnomalyDetection/RunOfflineDetector/Call RunInference (OfflineDetector:808c63)/BeamML_RunInference']

https://github.com/apache/beam/actions/runs/19757409550/job/56611473889?pr=36932#step:9:3263
https://github.com/apache/beam/actions/runs/19757409550/job/56611473885?pr=36932#step:9:3084
https://github.com/apache/beam/actions/runs/19757409550/job/56611473905?pr=36932#step:9:3091

Issue Priority

Priority: 2 (default / most bugs should be filed as P2)

Issue Components

• Component: Python SDK
• Component: Java SDK
• Component: Go SDK
• Component: Typescript SDK
• Component: IO connector
• Component: Beam YAML
• Component: Beam examples
• Component: Beam playground
• Component: Beam katas
• Component: Website
• Component: Infrastructure
• Component: Spark Runner
• Component: Flink Runner
• Component: Samza Runner
• Component: Twister2 Runner
• Component: Hazelcast Jet Runner
• Component: Google Cloud Dataflow Runner

[yuwtennis]

Quick diagnosis

scikit_learn had made breaking change in getting estimator tags in v1.6.0 .
The model that validates with scikit_learn is required to implement the _sklearn_tags_ but IForest used in the test_scoring_with_matched_features does not . This eventually causes the run_interface of the PyODModelHandler to fail.

At moment of this error , versions of scikit_learn and pyod is below.

https://github.com/apache/beam/actions/runs/19757409550/job/56611473889?pr=36932#step:9:1800

scikit-learn==1.8.0rc1
pyod==2.0.5

BTW The last scheduled unit test in pyod was using scikit-learn 1.7.2 . The test did not fail because the scikit-learn package was raising Warning instead .

[shunping]

Thanks for reporting this. See also #31285 (comment).