create_run_from_scores

lightning_ir.base.validation_utils.create_run_from_scores(query_ids: Sequence[str], doc_ids: Sequence[Sequence[str]], scores: Tensor) → DataFrame[source]

Convenience function to create a run DataFrame from query and document ids and scores.

Parameters:

  • query_ids (Sequence[str]) – List of query IDs.

  • doc_ids (Sequence[Sequence[str]]) – List of lists containing document IDs for each query.

  • scores (torch.Tensor) – Tensor containing scores for each query-document pair.

Returns:

DataFrame containing the run information with columns:

query_id, q0, doc_id, score, rank, and system.

Return type:

pd.DataFrame