Update app.py
Browse files
app.py
CHANGED
|
@@ -354,7 +354,6 @@
|
|
| 354 |
# gr.Markdown(f"Last updated on **{LAST_UPDATED}**")
|
| 355 |
|
| 356 |
# demo.launch()
|
| 357 |
-
|
| 358 |
import gradio as gr
|
| 359 |
import pandas as pd
|
| 360 |
import os
|
|
@@ -364,12 +363,15 @@ from datetime import datetime
|
|
| 364 |
LEADERBOARD_FILE = "leaderboard.csv" # File to store all submissions persistently
|
| 365 |
LAST_UPDATED = datetime.now().strftime("%B %d, %Y")
|
| 366 |
|
| 367 |
-
|
| 368 |
-
|
| 369 |
-
|
| 370 |
-
|
| 371 |
-
|
| 372 |
-
|
|
|
|
|
|
|
|
|
|
| 373 |
|
| 374 |
def clean_answer(answer):
|
| 375 |
"""
|
|
@@ -403,11 +405,20 @@ def load_leaderboard():
|
|
| 403 |
"""
|
| 404 |
Load all submissions from the leaderboard file.
|
| 405 |
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 406 |
return pd.read_csv(LEADERBOARD_FILE)
|
| 407 |
|
| 408 |
def evaluate_predictions_and_update_leaderboard(prediction_file):
|
| 409 |
"""
|
| 410 |
-
Evaluate predictions
|
| 411 |
"""
|
| 412 |
ground_truth_file = "ground_truth.csv"
|
| 413 |
if not os.path.exists(ground_truth_file):
|
|
@@ -449,6 +460,9 @@ def evaluate_predictions_and_update_leaderboard(prediction_file):
|
|
| 449 |
except Exception as e:
|
| 450 |
return f"Error during evaluation: {str(e)}", load_leaderboard()
|
| 451 |
|
|
|
|
|
|
|
|
|
|
| 452 |
# Gradio Interface
|
| 453 |
with gr.Blocks() as demo:
|
| 454 |
gr.Markdown("# Prediction Evaluation Tool with Leaderboard")
|
|
|
|
| 354 |
# gr.Markdown(f"Last updated on **{LAST_UPDATED}**")
|
| 355 |
|
| 356 |
# demo.launch()
|
|
|
|
| 357 |
import gradio as gr
|
| 358 |
import pandas as pd
|
| 359 |
import os
|
|
|
|
| 363 |
LEADERBOARD_FILE = "leaderboard.csv" # File to store all submissions persistently
|
| 364 |
LAST_UPDATED = datetime.now().strftime("%B %d, %Y")
|
| 365 |
|
| 366 |
+
def initialize_leaderboard_file():
|
| 367 |
+
"""
|
| 368 |
+
Check if the leaderboard file exists. If not, create it with appropriate headers.
|
| 369 |
+
"""
|
| 370 |
+
if not os.path.exists(LEADERBOARD_FILE):
|
| 371 |
+
pd.DataFrame(columns=[
|
| 372 |
+
"Model Name", "Overall Accuracy", "Valid Accuracy",
|
| 373 |
+
"Correct Predictions", "Total Questions", "Timestamp"
|
| 374 |
+
]).to_csv(LEADERBOARD_FILE, index=False)
|
| 375 |
|
| 376 |
def clean_answer(answer):
|
| 377 |
"""
|
|
|
|
| 405 |
"""
|
| 406 |
Load all submissions from the leaderboard file.
|
| 407 |
"""
|
| 408 |
+
if not os.path.exists(LEADERBOARD_FILE):
|
| 409 |
+
return pd.DataFrame({
|
| 410 |
+
"Model Name": [],
|
| 411 |
+
"Overall Accuracy": [],
|
| 412 |
+
"Valid Accuracy": [],
|
| 413 |
+
"Correct Predictions": [],
|
| 414 |
+
"Total Questions": [],
|
| 415 |
+
"Timestamp": [],
|
| 416 |
+
})
|
| 417 |
return pd.read_csv(LEADERBOARD_FILE)
|
| 418 |
|
| 419 |
def evaluate_predictions_and_update_leaderboard(prediction_file):
|
| 420 |
"""
|
| 421 |
+
Evaluate predictions, update the leaderboard, and return the updated leaderboard.
|
| 422 |
"""
|
| 423 |
ground_truth_file = "ground_truth.csv"
|
| 424 |
if not os.path.exists(ground_truth_file):
|
|
|
|
| 460 |
except Exception as e:
|
| 461 |
return f"Error during evaluation: {str(e)}", load_leaderboard()
|
| 462 |
|
| 463 |
+
# Initialize leaderboard file
|
| 464 |
+
initialize_leaderboard_file()
|
| 465 |
+
|
| 466 |
# Gradio Interface
|
| 467 |
with gr.Blocks() as demo:
|
| 468 |
gr.Markdown("# Prediction Evaluation Tool with Leaderboard")
|