Merge pull request #131 from daisybio/multi_features

Multi features

Author: JudithBernett

create_report.py

@@ -267,8 +267,10 @@ def draw_per_grouping_algorithm_plots(
 if __name__ == "__main__":
     parser = argparse.ArgumentParser(description="Generate reports from evaluation results")
     parser.add_argument("--run_id", required=True, help="Run ID for the current execution")
+    parser.add_argument("--dataset", required=True, help="Dataset name for which to render the result file")
     args = parser.parse_args()
     run_id = args.run_id
+    dataset = args.dataset
 
     # assert that the run_id folder exists
     if not os.path.exists(f"results/{run_id}"):
@@ -280,7 +282,7 @@ def draw_per_grouping_algorithm_plots(
         evaluation_results_per_drug,
         evaluation_results_per_cell_line,
         true_vs_pred,
-    ) = parse_results(path_to_results=f"results/{run_id}")
+    ) = parse_results(path_to_results=f"results/{run_id}", dataset=dataset)
 
     # part of pipeline: EVALUATE_FINAL, COLLECT_RESULTS
     (

drevalpy/models/utils.py

@@ -1,7 +1,6 @@
 """Utility functions for loading and processing data."""
 
 import os.path
-import warnings
 
 import numpy as np
 import pandas as pd
@@ -94,17 +93,11 @@ def iterate_features(df: pd.DataFrame, feature_type: str) -> dict[str, dict[str,
         if cl in features.keys():
             continue
         rows = df.loc[cl]
+        rows = rows.astype(float).to_numpy()
         if (len(rows.shape) > 1) and (rows.shape[0] > 1):  # multiple rows returned
-            warnings.warn(
-                f"Multiple rows returned for Cell Line {cl} (and maybe others) "
-                f"in feature {feature_type}, taking the first one.",
-                stacklevel=2,
-            )
-
-            rows = rows.iloc[0]
-        # convert to float values
-        rows = rows.astype(float)
-        features[cl] = {feature_type: rows.values}
+            # take mean
+            rows = np.mean(rows, axis=0)
+        features[cl] = {feature_type: rows}
     return features
 
 

drevalpy/visualization/utils.py

@@ -37,11 +37,12 @@ def _parse_layout(f: TextIO, path_to_layout: str) -> None:
     f.write("".join(layout))
 
 
-def parse_results(path_to_results: str) -> tuple[pd.DataFrame, pd.DataFrame, pd.DataFrame, pd.DataFrame]:
+def parse_results(path_to_results: str, dataset: str) -> tuple[pd.DataFrame, pd.DataFrame, pd.DataFrame, pd.DataFrame]:
     """
     Parse the results from the given directory.
 
     :param path_to_results: path to the results directory
+    :param dataset: dataset name, e.g., GDSC2
     :returns: evaluation results, evaluation results per drug, evaluation results per cell line, and true vs. predicted
         values
     """
@@ -54,7 +55,7 @@ def parse_results(path_to_results: str) -> tuple[pd.DataFrame, pd.DataFrame, pd.
     # Convert the path to a forward-slash version for the regex (for Windows)
     result_dir_str = str(result_dir).replace("\\", "/")
     pattern = re.compile(
-        rf"{result_dir_str}/(LPO|LCO|LDO)/[^/]+/(predictions|cross_study|randomization|robustness)/.*\.csv$"
+        rf"{result_dir_str}/{dataset}/(LPO|LCO|LDO)/[^/]+/(predictions|cross_study|randomization|robustness)/.*\.csv$"
     )
     result_files = [file for file in result_files if pattern.match(str(file).replace("\\", "/"))]
 
@@ -69,8 +70,9 @@ def parse_results(path_to_results: str) -> tuple[pd.DataFrame, pd.DataFrame, pd.
         rel_file = str(os.path.normpath(file.relative_to(result_dir))).replace("\\", "/")
         print(f'Evaluating file: "{rel_file}" ...')
         file_parts = rel_file.split("/")
-        lpo_lco_ldo = file_parts[0]
-        algorithm = file_parts[1]
+        dataset = file_parts[0]
+        lpo_lco_ldo = file_parts[1]
+        algorithm = file_parts[2]
         (
             overall_eval,
             eval_results_per_drug,

tests/test_drp_model.py

@@ -151,10 +151,9 @@ def test_iterate_features() -> None:
     """Test the iteration over features."""
     df = pd.DataFrame({"GeneA": [1, 2, 3, 2], "GeneB": [4, 5, 6, 2], "GeneC": [7, 8, 9, 2]})
     df.index = ["CellLine1", "CellLine2", "CellLine3", "CellLine1"]
-    with pytest.warns(UserWarning):
-        features = iterate_features(df, "gene_expression")
+    features = iterate_features(df, "gene_expression")
     assert len(features) == 3
-    assert np.all(features["CellLine1"]["gene_expression"] == [1, 4, 7])
+    assert np.all(features["CellLine1"]["gene_expression"] == [1.5, 3, 4.5])
 
 
 def test_load_drug_ids_from_csv() -> None:

tests/test_run_suite.py

@@ -53,7 +53,7 @@ def test_run_suite(args):
         evaluation_results_per_drug,
         evaluation_results_per_cell_line,
         true_vs_pred,
-    ) = parse_results(path_to_results=os.path.join(temp_dir.name, args.run_id, args.dataset_name))
+    ) = parse_results(path_to_results=os.path.join(temp_dir.name, args.run_id), dataset="Toy_Data")
 
     (
         evaluation_results,