Fix senkin_tmp predictions check

[rcannood]

In run_2026-09-28 senkin_tmp_predict still failed on every CITE dataset, now with AssertionError: test cells do not match the trained model. The cell names are identical, but the train image (pandas 3) writes the index of predictions.h5ad as a string dtype and the predict image (pandas 2.3) reads the test cells as object, so Index.equals() returns False.

Describe your changes

• senkin_tmp_predict: compare obs_names and var_names as lists.
• Regenerated the test model on S3 (models/senkin_tmp/predictions.h5ad), which only held the old model.pkl that the predict image cannot unpickle.

Checked on de.NBI against the failing pbmc_cite/normal inputs with the 0.2.0-rc6 image.

Checklist before requesting a review

• I have performed a self-review of my code

• Check the correct box. Does this PR contain:

  • Breaking changes
  • New functionality
  • Major changes
  • Minor changes
  • Bug fixes

• Proposed changes are described in the CHANGELOG.md

• CI Tests succeed and look good!