Merge pull request #380 from monarch-initiative/379-infinity-values-c…

…reate-an-error-with-roc-curves 379 infinity values create an error with roc curves
monarch-initiative · Jan 7, 2025 · d28e702 · d28e702
2 parents eb7ba11 + 80386bc
commit d28e702
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 3 deletions.
diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "pheval"
-version = "0.4.2"
+version = "0.4.3"
 description = ""
 authors = ["Yasemin Bridges <[email protected]>",
   "Julius Jacobsen <[email protected]>",

diff --git a/src/pheval/analyse/generate_plots.py b/src/pheval/analyse/generate_plots.py
@@ -2,6 +2,7 @@
 from typing import List
 
 import matplotlib
+import numpy as np
 import pandas as pd
 import seaborn as sns
 from matplotlib import pyplot as plt
@@ -368,9 +369,16 @@ def generate_roc_curve(
         """
         plt.clf()
         for i, benchmark_result in enumerate(benchmarking_results):
+            y_score = np.array(benchmark_result.binary_classification_stats.scores)
+            y_score = np.nan_to_num(
+                y_score,
+                nan=0.0,
+                posinf=max(y_score[np.isfinite(y_score)]),
+                neginf=min(y_score[np.isfinite(y_score)]),
+            )
             fpr, tpr, thresh = roc_curve(
                 benchmark_result.binary_classification_stats.labels,
-                benchmark_result.binary_classification_stats.scores,
+                y_score,
                 pos_label=1,
             )
             roc_auc = auc(fpr, tpr)
@@ -411,9 +419,16 @@ def generate_precision_recall(
         plt.clf()
         plt.figure()
         for i, benchmark_result in enumerate(benchmarking_results):
+            y_score = np.array(benchmark_result.binary_classification_stats.scores)
+            y_score = np.nan_to_num(
+                y_score,
+                nan=0.0,
+                posinf=max(y_score[np.isfinite(y_score)]),
+                neginf=min(y_score[np.isfinite(y_score)]),
+            )
             precision, recall, thresh = precision_recall_curve(
                 benchmark_result.binary_classification_stats.labels,
-                benchmark_result.binary_classification_stats.scores,
+                y_score,
             )
             precision_recall_auc = auc(recall, precision)
             plt.plot(