fix

events/lollms_generation_events.py

@@ -45,7 +45,10 @@ def add_events(sio: socketio):
         lollmsElfServer.cancel_gen = True
         #kill thread
         ASCIIColors.error(f'Client {sid} requested cancelling generation')
-        client.generation_routine.cancel()
+        try:
+            client.generation_routine.cancel()
+        except Exception as ex:
+            pass
         lollmsElfServer.busy=False
         if lollmsElfServer.tts:
             lollmsElfServer.tts.stop()
@@ -59,7 +62,7 @@ def add_events(sio: socketio):
         client = lollmsElfServer.session.get_client(client_id)
         client.requested_stop=True
         print(f"Client {client_id} requested canceling generation")
-        lollmsElfServer.sio.emit("generation_canceled", {"message":"Generation is canceled."}, to=client_id)
+        await lollmsElfServer.sio.emit("generation_canceled", {"message":"Generation is canceled."}, to=client_id)
         lollmsElfServer.busy = False
 
 
@@ -102,11 +105,6 @@ def add_events(sio: socketio):
             )
 
             await lollmsElfServer.start_message_generation(message, message.id, client_id)
-
-            # lollmsElfServer.sio.sleep(0.01)
-            lollmsElfServer.busy = True
-            # tpe = threading.Thread(target=lollmsElfServer.start_message_generation, args=(message, message_id, client_id))
-            # tpe.start()
         else:
             lollmsElfServer.error("I am busy. Come back later.", client_id=client_id)
 
@@ -154,10 +152,6 @@ def add_events(sio: socketio):
 
             await lollmsElfServer.start_message_generation(message, message.id, client_id, force_using_internet=True)
 
-            # lollmsElfServer.sio.sleep(0.01)
-            lollmsElfServer.busy = True
-            # tpe = threading.Thread(target=lollmsElfServer.start_message_generation, args=(message, message_id, client_id))
-            # tpe.start()
         else:
             lollmsElfServer.error("I am busy. Come back later.", client_id=client_id)
 

lollms_core

@@ -1 +1 @@
-Subproject commit e5cd49c9c4deed4afe958e8c410876f29ae4f838
+Subproject commit 0c21805e55c06faed7b2db3d23d9cd0494617ded

lollms_webui.py

@@ -450,79 +450,6 @@ class LOLLMSWebUI(LOLLMSElfServer):
         cleaned_string = re.sub(pattern, "", cleaned_string)
         return cleaned_string
 
-    def make_discussion_title(self, discussion, client_id=None):
-        """
-        Builds a title for a discussion
-        """
-
-        # Get the list of messages
-        messages = discussion.get_messages()
-        discussion_messages = f"{self.system_full_header}Create a short title to this discussion\n--- discussion ---\n"
-        discussion_title = f"\n{self.ai_custom_header('assistant')}"
-
-        available_space = (
-            self.config.ctx_size
-            - 150
-            - self.model.count_tokens(discussion_messages)
-            - self.model.count_tokens(discussion_title)
-        )
-        # Initialize a list to store the full messages
-        full_message_list = []
-        # Accumulate messages until the cumulative number of tokens exceeds available_space
-        tokens_accumulated = 0
-        # Accumulate messages starting from message_index
-        for message in messages:
-            # Check if the message content is not empty and visible to the AI
-            if message.content != "" and (
-                message.message_type
-                <= MSG_OPERATION_TYPE.MSG_OPERATION_TYPE_SET_CONTENT_INVISIBLE_TO_USER.value
-                and message.message_type
-                != MSG_OPERATION_TYPE.MSG_OPERATION_TYPE_SET_CONTENT_INVISIBLE_TO_AI.value
-            ):
-
-                # Tokenize the message content
-                message_tokenized = self.model.tokenize(
-                    "\n"
-                    + self.config.discussion_prompt_separator
-                    + message.sender
-                    + ": "
-                    + message.content.strip()
-                )
-
-                # Check if adding the message will exceed the available space
-                if tokens_accumulated + len(message_tokenized) > available_space:
-                    break
-
-                # Add the tokenized message to the full_message_list
-                full_message_list.insert(0, message_tokenized)
-
-                # Update the cumulative number of tokens
-                tokens_accumulated += len(message_tokenized)
-
-        # Build the final discussion messages by detokenizing the full_message_list
-
-        for message_tokens in full_message_list:
-            discussion_messages += self.model.detokenize(message_tokens)
-        discussion_messages +"\n--- ---\n"
-        discussion_messages +f"\n{self.user_custom_header('user')}Your response should only contain the title without any comments or thoughts.\n"
-        discussion_messages += discussion_title
-        title = [""]
-
-        def receive(chunk: str, message_type: MSG_OPERATION_TYPE):
-            if chunk:
-                title[0] += chunk
-            antiprompt = self.personality.detect_antiprompt(title[0])
-            if antiprompt:
-                ASCIIColors.warning(f"\n{antiprompt} detected. Stopping generation")
-                title[0] = self.remove_text_from_string(title[0], antiprompt)
-                return False
-            else:
-                return True
-
-        self._generate(discussion_messages, 1024, client_id, receive)
-        title[0] = self.personality.remove_thinking_blocks(title[0])
-        ASCIIColors.info(f"TITLE:{title[0]}")
-        return title[0]
 
     def get_discussion_to(self, client_id, message_id=-1):
         messages = self.session.get_client(client_id).discussion.get_messages()