Update app.py

app.py

@@ -523,6 +523,10 @@ async def _stream_fn(
     except StopIteration:
         update = history + [[message, None]]
         yield update, update
+    except Exception as e:
+        yield history, history
+        raise e
+
     try:
         async for response in generator:
             update = history + [[message, response]]
@@ -683,10 +687,14 @@ def vllm_generate_stream(
 
 
 # ! avoid saying 
-LANG_BLOCK_MESSAGE = """Sorry, the language you have asked is currently not supported. If you have questions in other supported languages, I'll be glad to help. \
-Please also consider clearing the chat box for a better experience."""
+# LANG_BLOCK_MESSAGE = """Sorry, the language you have asked is currently not supported. If you have questions in other supported languages, I'll be glad to help. \
+# Please also consider clearing the chat box for a better experience."""
+
+# KEYWORD_BLOCK_MESSAGE = "Sorry, I cannot fulfill your request. If you have any unrelated question, I'll be glad to help."
+
+LANG_BLOCK_MESSAGE = """Unsupported language."""
 
-KEYWORD_BLOCK_MESSAGE = "Sorry, I cannot fulfill your request. If you have any unrelated question, I'll be glad to help."
+KEYWORD_BLOCK_MESSAGE = "Invalid request."
 
 
 def _detect_lang(text):
@@ -696,7 +704,6 @@ def _detect_lang(text):
     try:
         dlang = detect_lang(text)
     except Exception as e:
-        # print(f'Error: {e}')
         if "No features in text." in str(e):
             return "en"
         else:
@@ -767,7 +774,12 @@ def chatml_format(message, history=None, system_prompt=None):
     return chatml_chat_convo_format(conversations, True, default_system=system_prompt)
 
 
-def debug_chat_response_stream_multiturn(*args, **kwargs):
+def debug_chat_response_stream_multiturn(message, history):
+    message_safety = safety_check(message, history=history)
+    if message_safety is not None:
+        # yield message_safety
+        raise gr.Error(message_safety)
+    
     message = "This is a debugging message"
     for i in range(len(message)):
         time.sleep(0.05)
@@ -787,13 +799,10 @@ def chat_response_stream_multiturn(
 ) -> str:
     global LOG_FILE, LOG_PATH
     if DEBUG:
-        yield from debug_chat_response_stream_multiturn()
+        yield from debug_chat_response_stream_multiturn(message, history)
         return
     from vllm import LLM, SamplingParams
     """Build multi turn
-    <bos>[INST] B_SYS SytemPrompt E_SYS Prompt [/INST] Answer <eos>
-    <bos>[INST] Prompt [/INST] Answer <eos>
-    <bos>[INST] Prompt [/INST]
 
     message is incoming prompt
     history don't have the current messauge
@@ -1424,7 +1433,8 @@ def launch_demo():
 
     if DEBUG:
         model_desc += "\n<br>!!!!! This is in debug mode, responses will copy original"
-        response_fn = debug_chat_response_echo
+        # response_fn = debug_chat_response_echo
+        response_fn = chat_response_stream_multiturn
         print(f'Creating in DEBUG MODE')
         if SAVE_LOGS:
             LOG_FILE = open(LOG_PATH, 'a', encoding='utf-8')