Fixed an issue in dominance computation (#28)

The problem was in the case where all values are the same. Co-authored-by: Raoul Schram <[email protected]>
sodascience · Mar 27, 2024 · 6a9fd3a · 6a9fd3a
1 parent f09222b
commit 6a9fd3a
Showing 1 changed file with 3 additions and 0 deletions.
diff --git a/metasyncontrib/disclosure/utils.py b/metasyncontrib/disclosure/utils.py
@@ -20,6 +20,9 @@ def _compute_dominance(block_values, reverse=False):
         same_vals = np.all(block_values == max_values, axis=1)
     diff_sum = diff_values.sum(axis=1)
     dominance = diff_values[~same_vals].max(axis=1) / diff_sum[~same_vals]
+    # If all values are the same, then dominance is 0.
+    if len(dominance) == 0:
+        return 0
     return np.max(dominance)