Remove obsolete NanoBEIR-en results

#20
HAKARI-Bench org

Remove obsolete NanoBEIR-en result files from hakari-bench/results.

Current hakari-bench main no longer includes NanoBEIR-en as a builtin dataset, and MNanoBEIR now excludes the English NanoBEIR slice. Keeping these files in the results dataset can cause rebuilds that scan raw result JSON to include obsolete NanoBEIR-en rows.

This PR deletes only compressed result JSON files matching:

hakari-results//hakari-bench__NanoBEIR-en/.json.xz

Deletion summary:

  • Files deleted: 1001
  • Result directories affected: 77
  • Files per affected directory: 13
  • Non-json result files deleted: 0
  • DuckDB/cache/report files deleted: 0

Validation performed locally before opening this PR:

  • DatasetRegistry.load_builtin().dataset_names() does not contain NanoBEIR-en.
  • resolve_eval_tasks(... dataset_names ...) returns 538 builtin tasks.
  • resolve_eval_tasks(... collection MNanoBEIR ...) returns 169 tasks across 13 non-English NanoBEIR language datasets.
  • Remote delete target list was generated from the current Hub file list and restricted to the path pattern above.
hotchpotch changed pull request status to merged

Sign up or log in to comment