1433eb7ca0d16cbb0dcaf0a54de2eb06654c8749,src/pyscenic/rnkdb.py,InvertedRankingDatabase,load,#InvertedRankingDatabase#Any#,430

Before Change

    def load(self, gs: Type[GeneSignature]) -> pd.DataFrame:
        rank_unknown = np.iinfo(INVERTED_DB_DTYPE).max
        features = self.df.index; n_features = len(features)
        reference_identifiers = np.array([self.identifier2idx[identifier] for identifier in gs.genes])
        ranked_identifiers = self.df.values

        def gen_series(ranked_identifiers, reference_identifiers, name):
            idx = np.nonzero(np.isin(ranked_identifiers, reference_identifiers))[0]
            return pd.Series(index=ranked_identifiers[idx], data=idx, name=name, dtype=INVERTED_DB_DTYPE)

        //return pd.DataFrame(data=build_rankings(ranked_identifiers, reference_identifiers),
        //                    index=self.df.index,
        //                    columns=gs.genes)

        return pd.concat([gen_series(ranked_identifiers[idx, :], reference_identifiers, features[idx]) for idx in range(n_features)], axis=1).T\
                                .fillna(rank_unknown).astype(INVERTED_DB_DTYPE)\
                                .rename(columns=self.idx2identifier, inplace=False)

After Change

    def load(self, gs: Type[GeneSignature]) -> pd.DataFrame:
        rank_unknown = np.iinfo(INVERTED_DB_DTYPE).max
        reference_identifiers = np.array([self.identifier2idx[identifier] for identifier in gs.genes])
        return pd.concat([col.reindex(index=reference_identifiers, fill_value=rank_unknown) for col in self.features],
                         axis=1).T.astype(INVERTED_DB_DTYPE).rename(columns=self.idx2identifier)

        //return pd.DataFrame(data=build_rankings(ranked_identifiers, reference_identifiers),

Instances