Enhance JSON review processing script with improved logging and error handling

Added detailed print statements to track the progress of label checks, JSON file processing, and error handling. Enhanced feedback for existing labels and improved error messages for various exceptions. Updated the main function to provide a summary of processed files and their statuses, ensuring better visibility into the processing workflow.

Author: stephen-rowan

Scripts/process_json_reviews.py

@@ -117,16 +117,25 @@ def ensure_labels_exist(repo: Any) -> None:
         "incorrect": "d73a4a",  # Red
     }
 
-    existing_labels = {label.name: label for label in repo.get_labels()}
+    print(f"  Checking for required labels...")
+    try:
+        existing_labels = {label.name: label for label in repo.get_labels()}
+        print(f"  ✓ Retrieved {len(existing_labels)} existing labels from repository")
+    except Exception as e:
+        print(f"  ✗ Failed to retrieve labels: {e}", file=sys.stderr)
+        raise
 
     for label_name, label_color in required_labels.items():
         if label_name not in existing_labels:
             try:
                 repo.create_label(label_name, label_color)
-                print(f"  ✓ Created label: {label_name}")
+                print(f"  ✓ Created label: {label_name} (color: {label_color})")
             except Exception as e:
                 # Label might already exist or we don't have permission
                 print(f"  ⚠ Could not create label '{label_name}': {e}", file=sys.stderr)
+                # Continue anyway - the label might have been created by another process
+        else:
+            print(f"  ✓ Label already exists: {label_name}")
 
 
 def create_github_issue(repo: Any, review_data: Dict[str, Any]) -> Optional[int]:
@@ -161,97 +170,142 @@ def create_github_issue(repo: Any, review_data: Dict[str, Any]) -> Optional[int]
 def process_json_file(file_path: Path, repo: Any) -> Tuple[bool, Optional[str]]:
     """Process a single JSON review file."""
     try:
+        print(f"  Reading file: {file_path}")
         # Read and parse JSON
-        with open(file_path, 'r') as f:
+        with open(file_path, 'r', encoding='utf-8') as f:
             review_data = json.load(f)
+        print(f"  ✓ Successfully parsed JSON")
 
         # Validate review
+        print(f"  Validating review data...")
         is_valid, error_msg = validate_review(review_data)
         if not is_valid:
+            print(f"  ✗ Validation failed: {error_msg}")
             return False, error_msg
+        print(f"  ✓ Validation passed")
 
         # Create GitHub Issue
+        print(f"  Creating GitHub Issue...")
         issue_number = create_github_issue(repo, review_data)
         if issue_number is None:
+            print(f"  ✗ Failed to create GitHub Issue")
             return False, "Failed to create GitHub Issue"
+        print(f"  ✓ Created GitHub Issue #{issue_number}")
 
         # Move file to processed directory
+        print(f"  Moving file to processed directory...")
         processed_path = PROCESSED_DIR / file_path.name
         os.makedirs(PROCESSED_DIR, exist_ok=True)
         shutil.move(str(file_path), str(processed_path))
+        print(f"  ✓ Moved to: {processed_path}")
 
         return True, None
 
     except json.JSONDecodeError as e:
+        print(f"  ✗ JSON decode error: {e}")
         return False, f"Invalid JSON: {e}"
+    except FileNotFoundError as e:
+        print(f"  ✗ File not found: {e}")
+        return False, f"File not found: {e}"
+    except PermissionError as e:
+        print(f"  ✗ Permission error: {e}")
+        return False, f"Permission error: {e}"
     except Exception as e:
+        print(f"  ✗ Unexpected error: {e}")
+        import traceback
+        traceback.print_exc()
         return False, f"Error processing file: {e}"
 
 
 def main() -> None:
     """Main function to process JSON review files."""
+    print("="*50)
+    print("Processing JSON Review Files")
+    print("="*50)
+    
     # Get GitHub token from environment
     github_token = os.environ.get("GITHUB_TOKEN")
     if not github_token:
         print("Error: GITHUB_TOKEN environment variable not set", file=sys.stderr)
         print("Note: For GitHub Actions, this is automatically set as GITHUB_TOKEN", file=sys.stderr)
         sys.exit(1)
+    else:
+        print(f"✓ GITHUB_TOKEN is set (length: {len(github_token)})")
 
     # Initialize GitHub client
     try:
+        print(f"\nConnecting to GitHub repository: {REPO_OWNER}/{REPO_NAME}")
         g = Github(github_token)
         repo = g.get_repo(f"{REPO_OWNER}/{REPO_NAME}")
+        print(f"✓ Successfully connected to repository")
     except Exception as e:
-        print(f"Error: Failed to connect to GitHub repository: {e}", file=sys.stderr)
+        print(f"✗ Error: Failed to connect to GitHub repository: {e}", file=sys.stderr)
         import traceback
         traceback.print_exc()
         sys.exit(2)
 
     # Ensure directories exist
+    print(f"\nChecking directories...")
     os.makedirs(REVIEWS_DIR, exist_ok=True)
     os.makedirs(PROCESSED_DIR, exist_ok=True)
+    print(f"✓ Reviews directory: {REVIEWS_DIR}")
+    print(f"✓ Processed directory: {PROCESSED_DIR}")
 
     # Find all JSON files in reviews directory (exclude processed directory)
+    print(f"\nScanning for JSON files in {REVIEWS_DIR}...")
+    all_json_files = list(REVIEWS_DIR.glob("*.json"))
     json_files = [
-        f for f in REVIEWS_DIR.glob("*.json")
+        f for f in all_json_files
         if f.parent.name != "processed"  # Exclude files in processed subdirectory
     ]
 
+    print(f"Found {len(all_json_files)} total JSON file(s)")
+    print(f"Found {len(json_files)} JSON file(s) to process (excluding processed/)")
+    
     if not json_files:
-        print("No JSON review files found in reviews/ directory")
+        print("\nNo JSON review files found in reviews/ directory")
+        print("Exiting successfully (no files to process)")
         return
 
-    print(f"Found {len(json_files)} JSON review file(s) to process")
-    
     # Process each file
     processed_count = 0
     failed_count = 0
     failed_files = []
 
     for json_file in json_files:
-        print(f"\nProcessing: {json_file.name}")
+        print(f"\n{'='*50}")
+        print(f"Processing: {json_file.name}")
+        print(f"{'='*50}")
         success, error_msg = process_json_file(json_file, repo)
 
         if success:
             processed_count += 1
+            print(f"✓ Successfully processed: {json_file.name}")
         else:
             failed_count += 1
             failed_files.append((json_file.name, error_msg))
-            print(f"  ✗ Failed: {error_msg}")
+            print(f"✗ Failed to process: {json_file.name}")
+            print(f"  Error: {error_msg}")
 
     # Summary
     print(f"\n{'='*50}")
     print(f"Processing Summary:")
+    print(f"{'='*50}")
     print(f"  ✓ Successfully processed: {processed_count}")
     print(f"  ✗ Failed: {failed_count}")
+    print(f"  Total files: {len(json_files)}")
 
     if failed_files:
         print(f"\nFailed files:")
         for filename, error in failed_files:
             print(f"  - {filename}: {error}")
 
     if failed_count > 0:
+        print(f"\n✗ Exiting with error code 1 due to {failed_count} failed file(s)")
         sys.exit(1)
+    else:
+        print(f"\n✓ All files processed successfully")
+        sys.exit(0)
 
 
 if __name__ == "__main__":