pheval 0.5.3__tar.gz → 0.5.5__tar.gz

{pheval-0.5.3 → pheval-0.5.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pheval
-Version: 0.5.3
+Version: 0.5.5
 Summary: 
 Author: Yasemin Bridges
 Author-email: y.bridges@qmul.ac.uk

{pheval-0.5.3 → pheval-0.5.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pheval"
-version = "0.5.3"
+version = "0.5.5"
 description = ""
 authors = ["Yasemin Bridges <y.bridges@qmul.ac.uk>",
   "Julius Jacobsen <j.jacobsen@qmul.ac.uk>",

{pheval-0.5.3 → pheval-0.5.5}/src/pheval/analyse/rank_stats.py

@@ -124,24 +124,11 @@ class Ranks:
         return np.mean(precision_at_k)
 
     @classmethod
-    def mean_average_precision_at_k(cls, df: pl.LazyFrame, k: int) -> pl.LazyFrame:
-        """
-        Compute Mean Average Precision at K (MAP@K) by averaging AP@K scores.
-        Args:
-            df (pl.LazyFrame): The dataframe calculate MAP@K for each query.
-            k (int): The upper rank limit.
-        Returns:
-            pl.LazyFrame: The dataframe with MAP@K for each query.
-        """
+    def mean_average_precision_at_k(cls, df: pl.LazyFrame, k: int) -> float:
         ap_at_k_df = cls._average_precision_at_k(df, k)
-        return (
-            ap_at_k_df.select(
-                pl.col(f"ap@{k}").sum() / df.select(Ranks.NUMBER_OF_SAMPLES).collect()
-            )
-            .fill_null(0.0)
-            .collect()
-            .item()
-        )
+        ap_sum = ap_at_k_df.select(pl.col(f"ap@{k}").sum()).collect().item()
+        num_samples = df.select(Ranks.NUMBER_OF_SAMPLES).collect().item()
+        return ap_sum / num_samples
 
     @classmethod
     def _calculate_ndcg_at_k(cls, ranks: List[int], k: int) -> float:
@@ -165,29 +152,18 @@ class Ranks:
         )
 
     @classmethod
-    def mean_normalised_discounted_cumulative_gain(cls, df: pl.LazyFrame, k: int) -> pl.Float64:
-        """
-        Compute mean normalised discounted cumulative gain.
-        Args:
-            df (pl.LazyFrame): The dataframe to calculate mean normalised cumulative gain.
-            k (int): The upper rank limit.
-        Returns:
-            pl.LazyFrame: The dataframe with mean normalised cumulative gain.
-        """
+    def mean_normalised_discounted_cumulative_gain(cls, df: pl.LazyFrame, k: int) -> float:
         filtered_df = cls._filter_results(df, k)
-        return (
-            filtered_df.with_columns(
-                pl.struct("ranks")
-                .map_elements(
-                    lambda row: cls._calculate_ndcg_at_k(row["ranks"], k), return_dtype=pl.Float64
-                )
-                .alias(f"NDCG@{k}")
+        ndcg_df = filtered_df.with_columns(
+            pl.struct("ranks")
+            .map_elements(
+                lambda row: cls._calculate_ndcg_at_k(row["ranks"], k), return_dtype=pl.Float64
             )
-            .select(pl.col(f"NDCG@{k}").sum() / df.select(Ranks.NUMBER_OF_SAMPLES).collect())
-            .fill_null(0.0)
-            .collect()
-            .item()
+            .alias(f"NDCG@{k}")
         )
+        ndcg_sum = ndcg_df.select(pl.col(f"NDCG@{k}").sum()).collect().item()
+        num_samples = df.select(Ranks.NUMBER_OF_SAMPLES).collect().item()
+        return ndcg_sum / num_samples
 
 
 def compute_rank_stats(run_identifier: str, result_scan: pl.LazyFrame) -> pl.LazyFrame:

{pheval-0.5.3 → pheval-0.5.5}/src/pheval/post_processing/post_processing.py

@@ -151,6 +151,10 @@ def create_empty_pheval_result(
     """
     if result_type in executed_results:
         return
+    logger.info(
+        f"Writing classified results for {len(all_files(phenopacket_dir))} "
+        f"phenopackets to {output_dir}"
+    )
     executed_results.add(result_type)
     phenopacket_truth_set = PhenopacketTruthSet(phenopacket_dir)
     classify_method, write_method = _get_result_type(result_type, phenopacket_truth_set)
@@ -180,10 +184,6 @@ def generate_gene_result(
         phenopacket_dir (Path): Path to the Phenopacket directory
     """
     output_file = output_dir.joinpath(f"pheval_gene_results/{result_path.stem}-gene_result.parquet")
-    logger.info(
-        f"Writing classified results for {len(all_files(phenopacket_dir))} "
-        f"phenopackets to {output_dir.joinpath('pheval_gene_results')}"
-    )
     create_empty_pheval_result(
         phenopacket_dir, output_dir.joinpath("pheval_gene_results"), ResultType.GENE
     )
@@ -214,10 +214,6 @@ def generate_variant_result(
     output_file = output_dir.joinpath(
         f"pheval_variant_results/{result_path.stem}-variant_result.parquet"
     )
-    logger.info(
-        f"Writing classified results for {len(all_files(phenopacket_dir))} "
-        f"phenopackets to {output_dir.joinpath('pheval_variant_results')}"
-    )
     create_empty_pheval_result(
         phenopacket_dir, output_dir.joinpath("pheval_variant_results"), ResultType.VARIANT
     )
@@ -250,10 +246,6 @@ def generate_disease_result(
     output_file = output_dir.joinpath(
         f"pheval_disease_results/{result_path.stem}-disease_result.parquet"
     )
-    logger.info(
-        f"Writing classified results for {len(all_files(phenopacket_dir))} "
-        f"phenopackets to {output_dir.joinpath('pheval_disease_results')}"
-    )
     create_empty_pheval_result(
         phenopacket_dir, output_dir.joinpath("pheval_disease_results"), ResultType.DISEASE
     )