Update app.py
Browse files
app.py
CHANGED
|
@@ -194,20 +194,6 @@ def clean_answer(answer):
|
|
| 194 |
return clean[0].upper()
|
| 195 |
return None
|
| 196 |
|
| 197 |
-
def update_leaderboard(results):
|
| 198 |
-
new_entry = {
|
| 199 |
-
"Model Name": results['model_name'],
|
| 200 |
-
"Overall Accuracy": round(results['overall_accuracy'] * 100, 2),
|
| 201 |
-
"Valid Accuracy": round(results['valid_accuracy'] * 100, 2),
|
| 202 |
-
"Correct Predictions": results['correct_predictions'],
|
| 203 |
-
"Total Questions": results['total_questions'],
|
| 204 |
-
"Timestamp": datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
|
| 205 |
-
}
|
| 206 |
-
leaderboard_df = pd.DataFrame([new_entry])
|
| 207 |
-
if os.path.exists(LEADERBOARD_FILE):
|
| 208 |
-
existing_df = pd.read_csv(LEADERBOARD_FILE)
|
| 209 |
-
leaderboard_df = pd.concat([existing_df, leaderboard_df], ignore_index=True)
|
| 210 |
-
leaderboard_df.to_csv(LEADERBOARD_FILE, index=False)
|
| 211 |
|
| 212 |
def evaluate_predictions(prediction_file):
|
| 213 |
ground_truth_file = "ground_truth.csv"
|
|
@@ -245,18 +231,50 @@ def evaluate_predictions(prediction_file):
|
|
| 245 |
except Exception as e:
|
| 246 |
return f"Error during evaluation: {str(e)}"
|
| 247 |
|
| 248 |
-
def load_leaderboard():
|
| 249 |
-
if not os.path.exists(LEADERBOARD_FILE):
|
| 250 |
-
return pd.DataFrame({"Message": ["Leaderboard is empty."]})
|
| 251 |
-
return pd.read_csv(LEADERBOARD_FILE)
|
| 252 |
|
| 253 |
# Build Gradio App
|
| 254 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 255 |
def load_leaderboard():
|
|
|
|
|
|
|
|
|
|
| 256 |
if not os.path.exists(LEADERBOARD_FILE):
|
| 257 |
-
return pd.DataFrame({
|
| 258 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 259 |
return pd.read_csv(LEADERBOARD_FILE)
|
|
|
|
|
|
|
| 260 |
def evaluate_predictions_and_update_leaderboard(prediction_file):
|
| 261 |
"""
|
| 262 |
Evaluate predictions and update the leaderboard.
|
|
|
|
| 194 |
return clean[0].upper()
|
| 195 |
return None
|
| 196 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 197 |
|
| 198 |
def evaluate_predictions(prediction_file):
|
| 199 |
ground_truth_file = "ground_truth.csv"
|
|
|
|
| 231 |
except Exception as e:
|
| 232 |
return f"Error during evaluation: {str(e)}"
|
| 233 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 234 |
|
| 235 |
# Build Gradio App
|
| 236 |
|
| 237 |
+
def update_leaderboard(results):
|
| 238 |
+
"""
|
| 239 |
+
Update the leaderboard file with new results.
|
| 240 |
+
"""
|
| 241 |
+
new_entry = {
|
| 242 |
+
"Model Name": results['model_name'],
|
| 243 |
+
"Overall Accuracy": round(results['overall_accuracy'] * 100, 2),
|
| 244 |
+
"Valid Accuracy": round(results['valid_accuracy'] * 100, 2),
|
| 245 |
+
"Correct Predictions": results['correct_predictions'],
|
| 246 |
+
"Total Questions": results['total_questions'],
|
| 247 |
+
"Timestamp": datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
|
| 248 |
+
}
|
| 249 |
+
|
| 250 |
+
# Convert new entry to DataFrame
|
| 251 |
+
new_entry_df = pd.DataFrame([new_entry])
|
| 252 |
+
|
| 253 |
+
# Append to leaderboard file
|
| 254 |
+
if not os.path.exists(LEADERBOARD_FILE):
|
| 255 |
+
# If file does not exist, create it with headers
|
| 256 |
+
new_entry_df.to_csv(LEADERBOARD_FILE, index=False)
|
| 257 |
+
else:
|
| 258 |
+
# Append without headers
|
| 259 |
+
new_entry_df.to_csv(LEADERBOARD_FILE, mode='a', index=False, header=False)
|
| 260 |
+
|
| 261 |
+
|
| 262 |
def load_leaderboard():
|
| 263 |
+
"""
|
| 264 |
+
Load the leaderboard from the leaderboard file.
|
| 265 |
+
"""
|
| 266 |
if not os.path.exists(LEADERBOARD_FILE):
|
| 267 |
+
return pd.DataFrame({
|
| 268 |
+
"Model Name": [],
|
| 269 |
+
"Overall Accuracy": [],
|
| 270 |
+
"Valid Accuracy": [],
|
| 271 |
+
"Correct Predictions": [],
|
| 272 |
+
"Total Questions": [],
|
| 273 |
+
"Timestamp": [],
|
| 274 |
+
})
|
| 275 |
return pd.read_csv(LEADERBOARD_FILE)
|
| 276 |
+
|
| 277 |
+
|
| 278 |
def evaluate_predictions_and_update_leaderboard(prediction_file):
|
| 279 |
"""
|
| 280 |
Evaluate predictions and update the leaderboard.
|