How to use the lightfm._lightfm_fast.CSRMatrix function in lightfm

positives_lookup,
                interactions.row,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                self.k,
                self.n,
                num_threads,
                self.random_state,
            )
        else:
            fit_logistic(
                CSRMatrix(item_features),
                CSRMatrix(user_features),
                interactions.row,
                interactions.col,
                interactions.data,
                sample_weight,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                num_threads,
            )

interactions.row,
                interactions.col,
                interactions.data,
                sample_weight,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                num_threads,
                self.random_state,
            )
        elif loss == "warp-kos":
            fit_warp_kos(
                CSRMatrix(item_features),
                CSRMatrix(user_features),
                positives_lookup,
                interactions.row,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                self.k,
                self.n,
                num_threads,
                self.random_state,
            )
        else:
            fit_logistic(
                CSRMatrix(item_features),
                CSRMatrix(user_features),

)

        n_users = user_ids.max() + 1
        n_items = item_ids.max() + 1

        (user_features, item_features) = self._construct_feature_matrices(
            n_users, n_items, user_features, item_features
        )

        lightfm_data = self._get_lightfm_data()

        predictions = np.empty(len(user_ids), dtype=np.float64)

        predict_lightfm(
            CSRMatrix(item_features),
            CSRMatrix(user_features),
            user_ids,
            item_ids,
            predictions,
            lightfm_data,
            num_threads,
        )

        return predictions

CSRMatrix(user_features),
                positives_lookup,
                interactions.row,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                self.k,
                self.n,
                num_threads,
                self.random_state,
            )
        else:
            fit_logistic(
                CSRMatrix(item_features),
                CSRMatrix(user_features),
                interactions.row,
                interactions.col,
                interactions.data,
                sample_weight,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                num_threads,
            )

interactions.row,
                interactions.col,
                interactions.data,
                sample_weight,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                num_threads,
                self.random_state,
            )
        elif loss == "bpr":
            fit_bpr(
                CSRMatrix(item_features),
                CSRMatrix(user_features),
                positives_lookup,
                interactions.row,
                interactions.col,
                interactions.data,
                sample_weight,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                num_threads,
                self.random_state,
            )
        elif loss == "warp-kos":
            fit_warp_kos(
                CSRMatrix(item_features),

positives_lookup,
                interactions.row,
                interactions.col,
                interactions.data,
                sample_weight,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                num_threads,
                self.random_state,
            )
        elif loss == "warp-kos":
            fit_warp_kos(
                CSRMatrix(item_features),
                CSRMatrix(user_features),
                positives_lookup,
                interactions.row,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                self.k,
                self.n,
                num_threads,
                self.random_state,
            )
        else:
            fit_logistic(
                CSRMatrix(item_features),

ranks = sp.csr_matrix(
            (
                np.zeros_like(test_interactions.data),
                test_interactions.indices,
                test_interactions.indptr,
            ),
            shape=test_interactions.shape,
        )

        lightfm_data = self._get_lightfm_data()

        predict_ranks(
            CSRMatrix(item_features),
            CSRMatrix(user_features),
            CSRMatrix(test_interactions),
            CSRMatrix(train_interactions),
            ranks.data,
            lightfm_data,
            num_threads,
        )

        return ranks

# The CSR conversion needs to happen before shuffle indices are created.
            # Calling .tocsr may result in a change in the data arrays of the COO matrix,
            positives_lookup = CSRMatrix(
                self._get_positives_lookup_matrix(interactions)
            )

        # Create shuffle indexes.
        shuffle_indices = np.arange(len(interactions.data), dtype=np.int32)
        self.random_state.shuffle(shuffle_indices)

        lightfm_data = self._get_lightfm_data()

        # Call the estimation routines.
        if loss == "warp":
            fit_warp(
                CSRMatrix(item_features),
                CSRMatrix(user_features),
                positives_lookup,
                interactions.row,
                interactions.col,
                interactions.data,
                sample_weight,
                shuffle_indices,
                lightfm_data,
                self.learning_rate,
                self.item_alpha,
                self.user_alpha,
                num_threads,
                self.random_state,
            )
        elif loss == "bpr":
            fit_bpr(