optimize

Data/show_plots.py

@@ -611,20 +611,31 @@ for gt_col in gt_columns:
     result_col = f'result.{base_name}'
 
     if result_col in df.columns:
-        # Convert to numeric, force invalid to NaN
-        s1 = pd.to_numeric(df[gt_col], errors='coerce')
-        s2 = pd.to_numeric(df[result_col], errors='coerce')
+        # 1. Convert to numeric, forcing errors to NaN
+        # 2. Explicitly cast to float to prevent Boolean subtraction errors
+        s1 = pd.to_numeric(df[gt_col], errors='coerce').astype(float)
+        s2 = pd.to_numeric(df[result_col], errors='coerce').astype(float)
 
         # Calculate matches (abs difference <= 0.5)
-        # Note: .sum() treats True as 1
-        matches = (np.abs(s1 - s2) <= 0.5).sum()
-        percentage = (matches / len(df)) * 100
+        # We use .count() to find how many non-NaN values we have if you want
+        # a more accurate percentage, or keep len(df) for the total row count.
+        diff = np.abs(s1 - s2)
+        matches = (diff <= 0.5).sum()
+
+        # Determine the denominator (total valid comparisons)
+        valid_count = diff.notna().sum()
+
+        if valid_count > 0:
+            percentage = (matches / valid_count) * 100
+        else:
+            percentage = 0
 
         data_list.append({
-            'GT': gt_col.replace('GT.', ''),
+            'GT': base_name,
             'Match %': round(percentage, 1)
         })
 
+
 # 2. Prepare Data for Plotting
 match_df = pd.DataFrame(data_list)
 # We pivot to get a format suitable for a heatmap