pheval 0.4.5__py3-none-any.whl → 0.4.7__py3-none-any.whl

pheval/cli.py

@@ -1,4 +1,4 @@
-"""PhEval CLI Module """
+"""PhEval CLI Module"""
 
 import logging
 

pheval/post_processing/post_processing.py

@@ -1,6 +1,6 @@
 import logging
 import operator
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from enum import Enum
 from pathlib import Path
 from typing import List, Union
@@ -96,6 +96,7 @@ class PhEvalVariantResult(PhEvalResult):
     ref: str
     alt: str
     score: float
+    grouping_id: str = field(default=None)
 
 
 @dataclass
@@ -105,7 +106,7 @@ class RankedPhEvalVariantResult(PhEvalVariantResult):
         rank (int): The rank for the result entry
     """
 
-    rank: int
+    rank: int = 0
 
     @staticmethod
     def from_variant_result(pheval_variant_result: PhEvalVariantResult, rank: int):
@@ -228,26 +229,57 @@ class ResultSorter:
         )
 
 
-def _rank_pheval_result(pheval_result: [PhEvalResult], sort_order: SortOrder) -> pd.DataFrame:
-    """
-    Rank PhEval results post-processed from tool-specific output, managing tied scores (ex aequo)
+class ResultRanker:
+    def __init__(self, pheval_result: List[PhEvalResult], sort_order: SortOrder):
+        """
+        Initialise the PhEvalRanker.
+        Args:
+            pheval_result (List[PhEvalResult]): PhEval results to rank.
+            sort_order (SortOrder): Sorting order based on which ranking is performed.
+        """
+        self.pheval_result = pheval_result
+        self.sort_order = sort_order
+        self.ascending = sort_order == SortOrder.ASCENDING
 
-    Args:
-        pheval_result ([PhEvalResult]): PhEval results obtained from tool-specific output
-        sort_order (SortOrder): Sorting order based on which ranking is performed
+    def rank(self) -> pd.DataFrame:
+        """
+        Rank PhEval results, managing tied scores (ex aequo) and handling grouping_id if present.
 
-    Returns:
-        pd.DataFrame : Ranked PhEval results with tied scores managed
+        Returns:
+            pd.DataFrame : Ranked PhEval results with tied scores managed.
+        """
+        pheval_result_df = pd.DataFrame([data.__dict__ for data in self.pheval_result])
 
-    Raises:
-        ValueError: If an incompatible PhEval result type is encountered
-    """
-    pheval_result_df = pd.DataFrame([data.__dict__ for data in pheval_result])
-    if sort_order == SortOrder.ASCENDING:
-        pheval_result_df["rank"] = pheval_result_df["score"].rank(method="max", ascending=True)
-    elif sort_order == SortOrder.DESCENDING:
-        pheval_result_df["rank"] = pheval_result_df["score"].rank(method="max", ascending=False)
-    return pheval_result_df
+        if self._has_valid_grouping_id(pheval_result_df):
+            pheval_result_df = self._rank_with_grouping_id(pheval_result_df)
+        else:
+            pheval_result_df = self._rank_without_grouping_id(pheval_result_df)
+        return pheval_result_df.drop(columns=["min_rank", "grouping_id"], errors="ignore")
+
+    @staticmethod
+    def _has_valid_grouping_id(pheval_result_df: pd.DataFrame) -> bool:
+        """Check if grouping_id exists and has no None values."""
+        return (
+            "grouping_id" in pheval_result_df.columns
+            and not pheval_result_df["grouping_id"].isnull().any()
+        )
+
+    def _rank_with_grouping_id(self, pheval_result_df: pd.DataFrame) -> pd.DataFrame:
+        """Apply ranking when grouping_id is present and has no None values."""
+        pheval_result_df["min_rank"] = (
+            pheval_result_df.groupby(["score", "grouping_id"])
+            .ngroup()
+            .rank(method="dense", ascending=self.ascending)
+        ).astype(int)
+        pheval_result_df["rank"] = pheval_result_df.groupby("score")["min_rank"].transform("max")
+        return pheval_result_df
+
+    def _rank_without_grouping_id(self, pheval_result_df: pd.DataFrame) -> pd.DataFrame:
+        """Apply ranking without using grouping_id."""
+        pheval_result_df["rank"] = (
+            pheval_result_df["score"].rank(method="max", ascending=self.ascending).astype(int)
+        )
+        return pheval_result_df
 
 
 def _return_sort_order(sort_order_str: str) -> SortOrder:
@@ -282,7 +314,7 @@ def _create_pheval_result(pheval_result: [PhEvalResult], sort_order_str: str) ->
     """
     sort_order = _return_sort_order(sort_order_str)
     sorted_pheval_result = ResultSorter(pheval_result, sort_order).sort_pheval_results()
-    return _rank_pheval_result(sorted_pheval_result, sort_order)
+    return ResultRanker(sorted_pheval_result, sort_order).rank()
 
 
 def _write_pheval_gene_result(

{pheval-0.4.5.dist-info → pheval-0.4.7.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pheval
-Version: 0.4.5
+Version: 0.4.7
 Summary: 
 Author: Yasemin Bridges
 Author-email: y.bridges@qmul.ac.uk

{pheval-0.4.5.dist-info → pheval-0.4.7.dist-info}/RECORD

@@ -16,7 +16,7 @@ pheval/analyse/prioritisation_result_types.py,sha256=qJoB6O-lFYmzAMcTQeDJZQNLJ6h
 pheval/analyse/rank_stats.py,sha256=vNLVuG_NzhKDXxKmklYNPz44MczlyKUqcuHqbiuOXwI,17993
 pheval/analyse/run_data_parser.py,sha256=VQBUoOIRYRWc5uqURUvaWdaW3E3C7Su0JvLavQLHQaY,4105
 pheval/analyse/variant_prioritisation_analysis.py,sha256=HhDeczF7wmJjXt0ejAtF0qdczyMe25glqiS6uX_TFl8,6408
-pheval/cli.py,sha256=EBGh6TIxAiWs0eDdQiefq6YuD0mb93siGsNmsVO1j7c,1527
+pheval/cli.py,sha256=SPB8-BCIRt1fUaAalhZ5Y6JUlnJX6Cj2S52QXCovJR8,1526
 pheval/cli_pheval.py,sha256=fWbKUcPTZZSa1EJEtH_lNn1XE6qRApRHihqUZS5owrA,2424
 pheval/cli_pheval_utils.py,sha256=O6tWnE85QQHGNcP08OwJGANMfXJPsZtFEu-D6ATld00,16700
 pheval/config_parser.py,sha256=lh-Dy_FflXJUnRC3HYaEdSvPAsNZWQZlEr1hHQigrTM,1227
@@ -24,7 +24,7 @@ pheval/implementations/__init__.py,sha256=BMUTotjTdgy5j5xubWCIQgRXrSQ1ZIcjooer7r
 pheval/infra/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 pheval/infra/exomiserdb.py,sha256=pM9-TfjrgurtH4OtM1Enk5oVhIxGQN3rKRlrxHuObTM,5080
 pheval/post_processing/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-pheval/post_processing/post_processing.py,sha256=tqeVRWF6PMHpOe681ONeGaqxdviLgVJgze3o6qSpXEg,13438
+pheval/post_processing/post_processing.py,sha256=MdacHoVjmwvmWBnHCSSKBboCgMW4MRGP-d_7-t1zZew,14808
 pheval/prepare/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 pheval/prepare/create_noisy_phenopackets.py,sha256=ydhA4mpqKTDc4hBu8YfvNW2nMubHK3dbO-cv0lA4JFQ,11504
 pheval/prepare/create_spiked_vcf.py,sha256=90A-Mi8QKhvN036vtFEVWAHgzHO37itiLYrqYlG4LiA,23953
@@ -50,8 +50,8 @@ pheval/utils/file_utils.py,sha256=m21cz-qjDYqnI8ClUv3J9fKizex98a-9bSEerQ75i_c,35
 pheval/utils/phenopacket_utils.py,sha256=6xQ8WCLdR1VhiU3nCDzaqEVKjGvDWrzvPA50_6ZAHXM,27310
 pheval/utils/semsim_utils.py,sha256=s7ZCR2VfPYnOh7ApX6rv66eGoVSm9QJaVYOWBEhlXpo,6151
 pheval/utils/utils.py,sha256=9V6vCT8l1g4O2-ZATYqsVyd7AYZdWGd-Ksy7_oIC3eE,2343
-pheval-0.4.5.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-pheval-0.4.5.dist-info/METADATA,sha256=fUGgR-undfldkiVtLHXx1AjoN4gaweABMbbk79Qp0cA,6469
-pheval-0.4.5.dist-info/WHEEL,sha256=IYZQI976HJqqOpQU6PHkJ8fb3tMNBFjg-Cn-pwAbaFM,88
-pheval-0.4.5.dist-info/entry_points.txt,sha256=o9gSwDkvT4-lqKy4mlsftd1nzP9WUOXQCfnbqycURd0,81
-pheval-0.4.5.dist-info/RECORD,,
+pheval-0.4.7.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+pheval-0.4.7.dist-info/METADATA,sha256=JfraeowwRp8eQjKiFBBrIVUoA0fchchznGj4t8sXgFE,6469
+pheval-0.4.7.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
+pheval-0.4.7.dist-info/entry_points.txt,sha256=o9gSwDkvT4-lqKy4mlsftd1nzP9WUOXQCfnbqycURd0,81
+pheval-0.4.7.dist-info/RECORD,,

{pheval-0.4.5.dist-info → pheval-0.4.7.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: poetry-core 2.0.1
+Generator: poetry-core 2.1.1
 Root-Is-Purelib: true
 Tag: py3-none-any