Update app.py

app.py

@@ -160,7 +160,6 @@ def load_embeddings() -> Optional[Dict[str, np.ndarray]]:
 
 def load_recipes_embeddings() -> Optional[Dict[str, np.ndarray]]:
     try:
-        # Locate or download the embeddings file
         embeddings_path = 'recipes_embeddings.safetensors'
         if not os.path.exists(embeddings_path):
             print("File not found locally. Attempting to download from Hugging Face Hub...")
@@ -169,21 +168,15 @@ def load_recipes_embeddings() -> Optional[Dict[str, np.ndarray]]:
                 filename="embeddings.safetensors",
                 repo_type="space"
             )
-        # Initialize a dictionary to store embeddings
+
+        # Using safe_open from safetensors to load embeddings
         embeddings = {}
-        # Open the safetensors file
+        from safetensors.numpy import safe_open
         with safe_open(embeddings_path, framework="pt") as f:
             keys = list(f.keys())
-            #print(f"Available keys in the .safetensors file: {keys}")  # Debugging info
-
-            # Iterate over the keys and load tensors
             for key in keys:
                 try:
-                    tensor = f.get_tensor(key)  # Get the tensor associated with the key
-                    if tensor.shape[0] != 384:  # Optional: Validate tensor shape
-                        print(f"Warning: Tensor for key {key} has unexpected shape {tensor.shape}")
-                    
-                    # Convert tensor to NumPy array
+                    tensor = f.get_tensor(key)
                     embeddings[key] = tensor.numpy()
                 except Exception as key_error:
                     print(f"Failed to process key {key}: {key_error}")
@@ -199,57 +192,6 @@ def load_recipes_embeddings() -> Optional[Dict[str, np.ndarray]]:
         print(f"Error loading embeddings: {e}")
         return None
 
-def load_documents_data(folder_path='downloaded_articles/downloaded_articles'):
-    """Load document data from HTML articles in a specified folder."""
-    try:
-        print("Loading documents data...")
-        # Check if the folder exists
-        if not os.path.exists(folder_path) or not os.path.isdir(folder_path):
-            print(f"Error: Folder '{folder_path}' not found")
-            return False
-        # List all HTML files in the folder
-        html_files = [f for f in os.listdir(folder_path) if f.endswith('.html')]
-        if not html_files:
-            print(f"No HTML files found in folder '{folder_path}'")
-            return False
-        documents = []
-        # Iterate through each HTML file and parse the content
-        for file_name in html_files:
-            file_path = os.path.join(folder_path, file_name)
-            try:
-                with open(file_path, 'r', encoding='utf-8') as file:
-                    # Parse the HTML file
-                    soup = BeautifulSoup(file, 'html.parser')
-                    # Extract text content (or customize this as per your needs)
-                    text = soup.get_text(separator='\n').strip()
-                    documents.append({"file_name": file_name, "content": text})
-            except Exception as e:
-                print(f"Error reading file {file_name}: {e}")
-        # Convert the list of documents to a DataFrame
-        data['df'] = pd.DataFrame(documents)
-        
-        if data['df'].empty:
-            print("No valid documents loaded.")
-            return False
-        print(f"Successfully loaded {len(data['df'])} document records.")
-        return True
-    except Exception as e:
-        print(f"Error loading documents data: {e}")
-        data['df'] = pd.DataFrame()
-        return False
-
-
-def load_data():
-    """Load all required data"""
-    embeddings_success = load_embeddings()
-    documents_success = load_documents_data()
-    recipes_embeddings_success = load_recipes_embeddings()
-    if not recipes_embeddings_success:
-        print("Warning: Failed to load embeddings, falling back to basic functionality")
-    return True
-
-# Initialize application
-print("Initializing application...")
 init_success = load_models() and load_data()
 
 def translate_text(text, source_to_target='ar_to_en'):
@@ -741,7 +683,7 @@ async def recipes_endpoint(profile: MedicalProfile):
 
         # Load recipe metadata from DataFrame
         file_path = 'recipes_metadata.xlsx'
-        df = pd.read_excel(file_path)
+        metadata_df = pd.read_excel(file_path)
 
         # Prepare the final recipes list
         recipes = []