Update app.py

app.py

@@ -11,22 +11,26 @@ def compare_csv_files():
     merged_df["CharacterErrorRate_Diff"] = merged_df["CharacterErrorRate_1.5"] - merged_df["CharacterErrorRate_1.4"]
     
     merged_df["WordErrorRate_Comparison"] = merged_df["WordErrorRate_Diff"].apply(
-        lambda x: f"1.5 is stronger than 1.4 ({x:.8f})" if x < 0 else (
-            f"1.4 is stronger than 1.5 ({-x:.8f})" if x > 0 else "1.4 is the same as 1.5 (0)"
+        lambda x: "1.4 is the same as 1.5 (Ignored due to large diff)" if abs(x) > 1 else (
+            f"1.5 is stronger than 1.4 ({x:.8f})" if x < 0 else (
+                f"1.4 is stronger than 1.5 ({-x:.8f})" if x > 0 else "1.4 is the same as 1.5 (0)"
+            )
         )
     )
     merged_df["CharacterErrorRate_Comparison"] = merged_df["CharacterErrorRate_Diff"].apply(
-        lambda x: f"1.5 is stronger than 1.4 ({x:.8f})" if x < 0 else (
-            f"1.4 is stronger than 1.5 ({-x:.8f})" if x > 0 else "1.4 is the same as 1.5 (0)"
+        lambda x: "1.4 is the same as 1.5 (Ignored due to large diff)" if abs(x) > 1 else (
+            f"1.5 is stronger than 1.4 ({x:.8f})" if x < 0 else (
+                f"1.4 is stronger than 1.5 ({-x:.8f})" if x > 0 else "1.4 is the same as 1.5 (0)"
+            )
         )
     )
     
-    avg_word_diff = merged_df["WordErrorRate_Diff"].mean()
-    avg_char_diff = merged_df["CharacterErrorRate_Diff"].mean()
+    avg_word_diff = merged_df["WordErrorRate_Diff"].loc[merged_df["WordErrorRate_Diff"].abs() <= 1].mean()
+    avg_char_diff = merged_df["CharacterErrorRate_Diff"].loc[merged_df["CharacterErrorRate_Diff"].abs() <= 1].mean()
     overall_summary = f"""
     <h3>Overall Comparison:</h3>
-    <p>Average WordErrorRate Difference: {'1.5 is stronger' if avg_word_diff < 0 else '1.4 is stronger'} ({avg_word_diff:.8f})</p>
-    <p>Average CharacterErrorRate Difference: {'1.5 is stronger' if avg_char_diff < 0 else '1.4 is stronger'} ({avg_char_diff:.8f})</p>
+    <p>Average WordErrorRate Difference (excluding large diffs): {'1.5 is stronger' if avg_word_diff < 0 else '1.4 is stronger'} ({avg_word_diff:.8f})</p>
+    <p>Average CharacterErrorRate Difference (excluding large diffs): {'1.5 is stronger' if avg_char_diff < 0 else '1.4 is stronger'} ({avg_char_diff:.8f})</p>
     """
     
     result = merged_df[[