Update app.py

app.py

@@ -50,48 +50,87 @@ except PermissionError as e:
     
 document_store = {}
 
+# Check if FAISS index file exists, otherwise initialize it
 if os.path.exists(faiss_index_file):
     try:
         index = faiss.read_index(faiss_index_file)
         if index.ntotal > 0:
             print(f"FAISS index loaded with {index.ntotal} vectors.")
-            index.reset()
-            index = faiss.IndexIDMap(faiss.IndexFlatL2(384))
+            index.reset()  # Resetting the index if non-zero entries
+            index = faiss.IndexIDMap(faiss.IndexFlatL2(384))  # Reinitialize the index
         else:
-            index = faiss.IndexIDMap(faiss.IndexFlatL2(384))
+            index = faiss.IndexIDMap(faiss.IndexFlatL2(384))  # Initialize with flat L2 distance
     except Exception as e:
         print(f"Error loading FAISS index: {e}, reinitializing.")
-        index = faiss.IndexIDMap(faiss.IndexFlatL2(384))
+        index = faiss.IndexIDMap(faiss.IndexFlatL2(384))  # Initialize if reading fails
 else:
-    index = faiss.IndexIDMap(faiss.IndexFlatL2(384))
+    index = faiss.IndexIDMap(faiss.IndexFlatL2(384))  # Initialize if file doesn't exist
 
 # Function to upload document
 def upload_document(file_path, embed_model):
+    # Generate unique document ID
     doc_id = uuid.uuid4().int % (2**63 - 1)
+    
+    # Ensure the file is saved to the correct directory with secure handling
     file_location = os.path.join(UPLOAD_DIR, os.path.basename(file_path))
-    shutil.copy(file_path, file_location)
     
-    with open(file_location, "r", encoding="utf-8") as f:
-        text = f.read()
+    # Safely copy the file to the upload directory
+    shutil.copy(file_path, file_location)
     
-    vector = embed_model.encode(text).astype("float32")
-    index.add_with_ids(np.array([vector]), np.array([doc_id], dtype=np.int64))
-    document_store[doc_id] = {"path": file_location, "text": text}
+    # Read the content of the uploaded file
+    try:
+        with open(file_location, "r", encoding="utf-8") as f:
+            text = f.read()
+    except Exception as e:
+        print(f"Error reading file {file_location}: {e}")
+        return
     
-    faiss.write_index(index, faiss_index_file)
-    print(f"Document uploaded with doc_id: {doc_id}")
+    # Embed the text and add it to the FAISS index
+    try:
+        vector = embed_model.encode(text).astype("float32")
+        index.add_with_ids(np.array([vector]), np.array([doc_id], dtype=np.int64))
+        document_store[doc_id] = {"path": file_location, "text": text}
+        
+        # Save the FAISS index after adding the document
+        faiss.write_index(index, faiss_index_file)
+        print(f"Document uploaded with doc_id: {doc_id}")
+    except Exception as e:
+        print(f"Error during document upload: {e}")
 
+@app.route("/upload", methods=["POST"])
 @app.route("/upload", methods=["POST"])
 def handle_upload():
+    # Check if the request contains the file
     if "file" not in request.files:
         return jsonify({"error": "No file provided"}), 400
     
     file = request.files["file"]
+    
+    # Ensure the filename is safe and construct the full file path
     file_path = os.path.join(UPLOAD_DIR, file.filename)
-    file.save(file_path)
     
-    upload_document(file_path, bertmodel)
-    return jsonify({"message": "File uploaded successfully"})
+    # Ensure the upload directory exists and has correct permissions
+    try:
+        os.makedirs(UPLOAD_DIR, exist_ok=True)  # Ensure the directory exists
+        if not os.access(UPLOAD_DIR, os.W_OK):  # Check write permissions
+            os.chmod(UPLOAD_DIR, 0o777)
+    except PermissionError as e:
+        return jsonify({"error": f"Permission error with upload directory: {e}"}), 500
+
+    try:
+        # Save the file to the upload directory
+        file.save(file_path)
+    except Exception as e:
+        return jsonify({"error": f"Error saving file: {e}"}), 500
+    
+    # Process the document using the upload_document function
+    try:
+        upload_document(file_path, bertmodel)  # Assuming 'bertmodel' is defined elsewhere
+    except Exception as e:
+        return jsonify({"error": f"Error processing file: {e}"}), 500
+    
+    # Return success response
+    return jsonify({"message": "File uploaded and processed successfully"}), 200
 
 def get_model_and_tokenizer(model_id: str):
     """