def _get_ids_list()

in kilt/eval_retrieval.py [0:0]

• 25 lines of code
• 8 McCabe index (conditional complexity)

def _get_ids_list(datapoint, rank_keys, verbose=False):
    # collect all gold ids
    ids_list = []
    for output in datapoint["output"]:
        current_ids_list = []
        if "provenance" in output:
            for provenance in output["provenance"]:
                if any(rank_key not in provenance for rank_key in rank_keys):
                    missing = set(rank_keys) - set(
                        list(provenance.keys())
                    ).intersection(set(rank_keys))
                    if verbose:
                        print(
                            f"WARNING: missing key(s) {missing} in provenance, unable to compute retrieval for those."
                        )
                else:
                    current_ids_list.append(
                        "+".join(
                            [
                                str(provenance[rank_key]).strip()
                                for rank_key in rank_keys
                            ]
                        )
                    )
        ids_list.append(_remove_duplicates(current_ids_list))  # remove duplicates

    # consider only unique ids
    return ids_list