Update app.py

app.py

@@ -66,32 +66,49 @@ def load_google_sheet_data(sheet_id, service_account_key_base64):
         return [], [], torch.tensor([])
 
     try:
-        # Decode the base64 key
+        print("Decoding base64 key...")
         key_bytes = base64.b64decode(service_account_key_base64)
         key_dict = json.loads(key_bytes)
+        print("Base64 key decoded and parsed.")
 
-        # Authenticate using the service account key
+        print("Authenticating with service account...")
         from google.oauth2 import service_account
         creds = service_account.Credentials.from_service_account_info(key_dict)
         client = gspread.authorize(creds)
+        print("Authentication successful.")
 
+        print(f"Opening sheet with key '{sheet_id}'...")
+        # *** IMPORTANT: If your sheet is NOT the first sheet, change 'sheet1'
+        # *** For example, if your sheet is named 'Data', use:
+        # sheet = client.open_by_key(sheet_id).worksheet("Data")
         sheet = client.open_by_key(sheet_id).sheet1
         print(f"Successfully opened Google Sheet with ID: {sheet_id}")
+
+        print("Getting all records from the sheet...")
         sheet_data = sheet.get_all_records()
+        print(f"Retrieved {len(sheet_data)} raw records from sheet.")
 
         if not sheet_data:
             print(f"Warning: No data records found in Google Sheet with ID: {sheet_id}")
             return [], [], torch.tensor([])
 
+        print("Filtering data for 'Service' and 'Description' columns...")
         filtered_data = [row for row in sheet_data if row.get('Service') and row.get('Description')]
+        print(f"Filtered down to {len(filtered_data)} records.")
+
         if not filtered_data:
             print("Warning: Filtered data is empty after checking for 'Service' and 'Description'.")
+            # Check if headers exist at all if filtered_data is empty but sheet_data isn't
+            if sheet_data and ('Service' not in sheet_data[0] or 'Description' not in sheet_data[0]):
+                 print("Error: 'Service' or 'Description' headers are missing or misspelled in the sheet.")
             return [], [], torch.tensor([])
 
-        if not filtered_data or 'Service' not in filtered_data[0] or 'Description' not in filtered_data[0]:
-             print("Error: Filtered Google Sheet data must contain 'Service' and 'Description' columns.")
+        # Re-checking column existence on filtered_data (redundant after filter but safe)
+        if 'Service' not in filtered_data[0] or 'Description' not in filtered_data[0]:
+             print("Error: Filtered Google Sheet data must contain 'Service' and 'Description' columns. This should not happen if filtering worked.")
              return [], [], torch.tensor([])
 
+
         services = [row["Service"] for row in filtered_data]
         descriptions = [row["Description"] for row in filtered_data]
         print(f"Loaded {len(descriptions)} entries from Google Sheet for embedding.")
@@ -106,7 +123,6 @@ def load_google_sheet_data(sheet_id, service_account_key_base64):
         print(f"An error occurred while accessing the Google Sheet: {e}")
         return [], [], torch.tensor([])
 
-
 def load_llm_model(model_id, hf_token):
     """Loads the LLM in full precision (for CPU)."""
     print(f"Loading model {model_id} in full precision...")