Update app.py

app.py

@@ -42,32 +42,31 @@ def predict(input, history=[]):
     # Check if the user input is a question
     is_question = "?" in input
 
-
-    # tokenize the new input sentence
-    new_user_input_ids = tokenizer.encode(input + tokenizer.eos_token, return_tensors='pt')
-
-     # Convert history tensor to a list
-    history_list = history.tolist() if isinstance(history, torch.Tensor) else history
+    if is_question: 
+        sql_encoding = sql_tokenizer(table=table, query=input + sql_tokenizer.eos_token, return_tensors="pt")
+        sql_outputs = sql_model.generate(**sql_encoding)
+        response = sql_tokenizer.batch_decode(sql_outputs, skip_special_tokens=True)
+
+        '''
+        bot_input_ids = torch.cat([torch.LongTensor(history), sql_encoding], dim=-1)
+        history = sql_model.generate(bot_input_ids, max_length=1000, pad_token_id=sql_tokenizer.eos_token_id).tolist()
+        response = sql_tokenizer.decode(history[0]).split("<|endoftext|>")
+        response = [(response[i], response[i + 1]) for i in range(0, len(response) - 1, 2)]
+        '''
+    else:
+        # tokenize the new input sentence
+        new_user_input_ids = tokenizer.encode(input + tokenizer.eos_token, return_tensors='pt')
     
-    # append the new user input tokens to the chat history
-    bot_input_ids = torch.cat([torch.LongTensor(history_list), new_user_input_ids], dim=-1)
-      
-    # generate a response
-    history = model.generate(bot_input_ids, max_length=1000, pad_token_id=tokenizer.eos_token_id).tolist()
-    response_dialog = tokenizer.decode(history[0])
-
-    # Use the SQL model to generate a response
-    encoding = sql_tokenizer(table=table, query=response_dialog, return_tensors="pt")
-    outputs = sql_model.generate(**encoding)
-    response_sql = sql_tokenizer.batch_decode(outputs, skip_special_tokens=True)
-
-    # Add the SQL model's response to the chat history
-    history.extend(response_sql)
-  
-    # convert the tokens to text, and then split the responses into the right format
-    response = tokenizer.decode(history[0]).split("<|endoftext|>")
-    response = [(response[i], response[i + 1]) for i in range(0, len(response) - 1, 2)]  # convert to tuples of list
-
+        # append the new user input tokens to the chat history
+        bot_input_ids = torch.cat([torch.LongTensor(history), new_user_input_ids], dim=-1)
+          
+        # generate a response
+        history = model.generate(bot_input_ids, max_length=1000, pad_token_id=tokenizer.eos_token_id).tolist()
+    
+        # convert the tokens to text, and then split the responses into the right format
+        response = tokenizer.decode(history[0]).split("<|endoftext|>")
+        response = [(response[i], response[i + 1]) for i in range(0, len(response) - 1, 2)]  # convert to tuples of list
+        
     return response, history