Update src/leaderboard.py

src/leaderboard.py

@@ -118,88 +118,6 @@ def save_leaderboard(df: pd.DataFrame) -> bool:
         print(f"Error saving leaderboard: {e}")
         return False
 
-# def add_model_to_leaderboard(
-#     model_name: str,
-#     author: str,
-#     evaluation_results: Dict,
-#     validation_info: Dict,
-#     model_type: str = "",
-#     description: str = ""
-# ) -> pd.DataFrame:
-#     """Add new model results to leaderboard."""
-    
-#     # Load current leaderboard
-#     df = load_leaderboard()
-    
-#     # Check if model already exists
-#     existing_mask = df['model_name'] == model_name
-#     if existing_mask.any():
-#         print(f"Model '{model_name}' already exists. Updating...")
-#         df = df[~existing_mask]  # Remove existing entry
-    
-#     # Extract metrics
-#     averages = evaluation_results.get('averages', {})
-#     google_averages = evaluation_results.get('google_comparable_averages', {})
-#     summary = evaluation_results.get('summary', {})
-    
-#     # Create new entry
-#     new_entry = {
-#         'submission_id': create_submission_id(),
-#         'model_name': sanitize_model_name(model_name),
-#         'author': author[:100] if author else 'Anonymous',
-#         'submission_date': datetime.datetime.now().isoformat(),
-#         'model_type': model_type[:50] if model_type else 'unknown',
-#         'description': description[:500] if description else '',
-        
-#         # Primary metrics
-#         'quality_score': float(averages.get('quality_score', 0.0)),
-#         'bleu': float(averages.get('bleu', 0.0)),
-#         'chrf': float(averages.get('chrf', 0.0)),
-        
-#         # Secondary metrics
-#         'rouge1': float(averages.get('rouge1', 0.0)),
-#         'rouge2': float(averages.get('rouge2', 0.0)),
-#         'rougeL': float(averages.get('rougeL', 0.0)),
-#         'cer': float(averages.get('cer', 0.0)),
-#         'wer': float(averages.get('wer', 0.0)),
-#         'len_ratio': float(averages.get('len_ratio', 0.0)),
-        
-#         # Google comparable metrics
-#         'google_quality_score': float(google_averages.get('quality_score', 0.0)),
-#         'google_bleu': float(google_averages.get('bleu', 0.0)),
-#         'google_chrf': float(google_averages.get('chrf', 0.0)),
-        
-#         # Coverage info
-#         'total_samples': int(summary.get('total_samples', 0)),
-#         'language_pairs_covered': int(summary.get('language_pairs_covered', 0)),
-#         'google_pairs_covered': int(summary.get('google_comparable_pairs', 0)),
-#         'coverage_rate': float(validation_info.get('coverage', 0.0)),
-        
-#         # Detailed results
-#         'detailed_metrics': json.dumps(evaluation_results),
-#         'validation_report': validation_info.get('report', ''),
-        
-#         # Metadata
-#         'evaluation_date': datetime.datetime.now().isoformat(),
-#         'leaderboard_version': 1
-#     }
-    
-#     # Add to dataframe
-#     new_row_df = pd.DataFrame([new_entry])
-#     updated_df = pd.concat([df, new_row_df], ignore_index=True)
-    
-#     # Sort by quality score (descending)
-#     updated_df = updated_df.sort_values('quality_score', ascending=False).reset_index(drop=True)
-    
-#     # Save updated leaderboard
-#     if save_leaderboard(updated_df):
-#         print(f"Added '{model_name}' to leaderboard")
-#         return updated_df
-#     else:
-#         print("Failed to save leaderboard")
-#         return df
-
-
 def add_model_to_leaderboard(
     model_name: str,
     author: str,
@@ -284,6 +202,64 @@ def add_model_to_leaderboard(
 
     return updated_df
 
+def prepare_leaderboard_display(df: pd.DataFrame) -> pd.DataFrame:
+    """Prepare leaderboard for display by formatting and selecting appropriate columns."""
+    
+    if df.empty:
+        return df
+    
+    # Select columns for display (exclude detailed_metrics and validation_report)
+    display_columns = [
+        'model_name', 'author', 'submission_date', 'model_type',
+        'quality_score', 'bleu', 'chrf', 
+        'rouge1', 'rougeL',
+        'total_samples', 'language_pairs_covered', 'google_pairs_covered',
+        'coverage_rate'
+    ]
+    
+    # Only include columns that exist
+    available_columns = [col for col in display_columns if col in df.columns]
+    display_df = df[available_columns].copy()
+    
+    # Format numeric columns
+    numeric_format = {
+        'quality_score': '{:.4f}',
+        'bleu': '{:.2f}',
+        'chrf': '{:.4f}',
+        'rouge1': '{:.4f}',
+        'rougeL': '{:.4f}',
+        'coverage_rate': '{:.1%}',
+    }
+    
+    for col, fmt in numeric_format.items():
+        if col in display_df.columns:
+            display_df[col] = display_df[col].apply(lambda x: fmt.format(float(x)) if pd.notnull(x) else "0.0000")
+    
+    # Format submission date
+    if 'submission_date' in display_df.columns:
+        display_df['submission_date'] = pd.to_datetime(display_df['submission_date']).dt.strftime('%Y-%m-%d %H:%M')
+    
+    # Rename columns for better display
+    column_renames = {
+        'model_name': 'Model Name',
+        'author': 'Author', 
+        'submission_date': 'Submitted',
+        'model_type': 'Type',
+        'quality_score': 'Quality Score',
+        'bleu': 'BLEU',
+        'chrf': 'ChrF',
+        'rouge1': 'ROUGE-1',
+        'rougeL': 'ROUGE-L',
+        'total_samples': 'Samples',
+        'language_pairs_covered': 'Lang Pairs',
+        'google_pairs_covered': 'Google Pairs',
+        'coverage_rate': 'Coverage'
+    }
+    
+    display_df = display_df.rename(columns=column_renames)
+    
+    return display_df
+
 def get_leaderboard_stats(df: pd.DataFrame) -> Dict:
     """Get summary statistics for the leaderboard."""