Fix API response to include file content and improve file handling

app.py

@@ -15,9 +15,15 @@ app = Flask(__name__)
 CORS(app)
 
 # Configure logging
-logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+    handlers=[
+        logging.StreamHandler(sys.stdout),
+        logging.FileHandler('/tmp/mineruapi.log')
+    ]
+)
 logger = logging.getLogger()
-logger.addHandler(logging.StreamHandler(sys.stdout))
 
 # Constants
 UPLOAD_FOLDER = os.environ.get('UPLOAD_FOLDER', '/tmp/pdf_uploads')
@@ -401,7 +407,14 @@ def convert_pdf():
         markdown_path = os.path.join(session_dir, f"{base_filename}.md")
         json_path = os.path.join(session_dir, f"{base_filename}.json")
         
-        # Define file URLs
+        # Define alternate locations to search for generated files
+        alternate_locations = [
+            os.path.dirname(input_path),  # Check PDF directory
+            os.path.join(OUTPUT_FOLDER, base_filename),  # Check output folder with base filename
+            session_dir  # Default session directory
+        ]
+        
+        # URLs for file download
         markdown_url = f"/download/{session_id}/{base_filename}.md"
         json_url = f"/download/{session_id}/{base_filename}.json"
         
@@ -465,31 +478,70 @@ def convert_pdf():
                         "command": " ".join(cmd)
                     }), 500
             
-            # Check if output files exist
+            # Check if files were generated
             md_exists = os.path.exists(markdown_path)
             json_exists = os.path.exists(json_path)
             
-            # If files don't exist in the expected location, try to find them in the session directory
+            # Search in alternate locations if files not found
+            if not md_exists or not json_exists:
+                logging.info("Files not found in primary location, searching alternate locations...")
+                
+                for location in alternate_locations:
+                    if not md_exists:
+                        alt_md_path = os.path.join(location, f"{base_filename}.md")
+                        if os.path.exists(alt_md_path):
+                            # Copy to session directory for download access
+                            try:
+                                shutil.copy(alt_md_path, markdown_path)
+                                md_exists = True
+                                logging.info(f"Found and copied Markdown file from {alt_md_path}")
+                            except:
+                                # If copy fails, update path to point to original
+                                markdown_path = alt_md_path
+                                md_exists = True
+                                logging.info(f"Found Markdown file at alternate location: {alt_md_path}")
+                    
+                    if not json_exists:
+                        alt_json_path = os.path.join(location, f"{base_filename}.json")
+                        if os.path.exists(alt_json_path):
+                            # Copy to session directory for download access
+                            try:
+                                shutil.copy(alt_json_path, json_path)
+                                json_exists = True
+                                logging.info(f"Found and copied JSON file from {alt_json_path}")
+                            except:
+                                # If copy fails, update path to point to original
+                                json_path = alt_json_path
+                                json_exists = True
+                                logging.info(f"Found JSON file at alternate location: {alt_json_path}")
+            
+            # If still not found, check for any markdown or JSON files in the directory
             if not md_exists:
-                # Look for any markdown file in the session directory
                 md_files = [f for f in os.listdir(session_dir) if f.endswith('.md')]
                 if md_files:
                     markdown_path = os.path.join(session_dir, md_files[0])
                     md_exists = True
-                    logging.info(f"Found markdown file at alternate location: {markdown_path}")
+                    logging.info(f"Found Markdown file at alternate location: {markdown_path}")
             
             if not json_exists:
-                # Look for any JSON file in the session directory
                 json_files = [f for f in os.listdir(session_dir) if f.endswith('.json')]
                 if json_files:
                     json_path = os.path.join(session_dir, json_files[0])
                     json_exists = True
                     logging.info(f"Found JSON file at alternate location: {json_path}")
             
-            if not md_exists:
+            if md_exists:
+                with open(markdown_path, 'r', encoding='utf-8') as f:
+                    markdown_content = f.read()
+            else:
+                markdown_content = None
                 logging.warning(f"Markdown file not found at {markdown_path}")
             
-            if not json_exists:
+            if json_exists:
+                with open(json_path, 'r', encoding='utf-8') as f:
+                    json_content = json.load(f)
+            else:
+                json_content = None
                 logging.warning(f"JSON file not found at {json_path}")
             
             return jsonify({
@@ -497,6 +549,8 @@ def convert_pdf():
                 "session_id": session_id,
                 "markdown_url": markdown_url if md_exists else None,
                 "json_url": json_url if json_exists else None,
+                "markdown_content": markdown_content,
+                "json_content": json_content,
                 "message": "PDF conversion completed"
             })
             
@@ -530,14 +584,41 @@ def convert_pdf():
 def download_file(session_id, filename):
     # Validate the session ID and filename
     if not all(c.isalnum() or c == '-' for c in session_id):
+        logging.warning(f"Invalid session ID format: {session_id}")
         abort(400, "Invalid session ID")
     
+    # Sanitize filename
+    if not all(c.isalnum() or c in ['-', '_', '.'] for c in filename):
+        logging.warning(f"Invalid filename format: {filename}")
+        abort(400, "Invalid filename")
+    
+    # First try the primary location
     base_path = os.path.join(OUTPUT_FOLDER, session_id)
     file_path = os.path.join(base_path, filename)
     
+    # If not found, try alternate locations
     if not os.path.exists(file_path):
-        abort(404, "File not found")
+        logging.info(f"File not found at primary location: {file_path}")
+        
+        # Try searching in other common locations
+        alternate_locations = [
+            OUTPUT_FOLDER,
+            os.path.join(OUTPUT_FOLDER, os.path.splitext(filename)[0]),
+            os.path.join(UPLOAD_FOLDER, session_id),
+            os.path.join(tempfile.gettempdir(), f"minerupdf_{session_id}")
+        ]
+        
+        for location in alternate_locations:
+            alt_path = os.path.join(location, filename)
+            if os.path.exists(alt_path):
+                logging.info(f"Found file at alternate location: {alt_path}")
+                file_path = alt_path
+                break
+        else:
+            logging.warning(f"File not found in any location: {filename}")
+            abort(404, "File not found")
     
+    logging.info(f"Serving file: {file_path}")
     return send_file(file_path, as_attachment=True)
 
 if __name__ == '__main__':