fix bug in context

delete required text caption with image

README.md

@@ -10,9 +10,6 @@
 <a href="https://t.me/chatgpt_karfly_bot?start=source=github" alt="Run Telegram Bot shield"><img src="https://img.shields.io/badge/RUN-Telegram%20Bot-blue?logo=data:image/svg+xml;base64,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" width="230"/></a>
 </p>
 
-<p align="center">
-<a href="https://github.com/karfly/chatgpt_telegram_bot/blob/main/static/donate/donate.md#%EF%B8%8F-donate" alt="Donate shield"><img src="https://img.shields.io/badge/-Donate-red?logo=undertale" width="100"/></a>
-</p>
 
 We all love [chat.openai.com](https://chat.openai.com), but... It's TERRIBLY laggy, has daily limits, and is only accessible through an archaic web interface.
 
@@ -25,6 +22,7 @@ You can deploy your own bot, or use mine: [@chatgpt_karfly_bot](https://t.me/cha
 - No request limits
 - Message streaming (watch demo)
 - GPT-4 and GPT-4 Turbo support
+- GPT-4 Vision support
 - Group Chat support (/help_group_chat to get instructions)
 - DALLE 2 (choose 👩‍🎨 Artist mode to generate images)
 - Voice message recognition
@@ -52,6 +50,7 @@ You can deploy your own bot, or use mine: [@chatgpt_karfly_bot](https://t.me/cha
 - *2 Mar 2023*: Added support of [ChatGPT API](https://platform.openai.com/docs/guides/chat/introduction).
 - *1 Aug 2023*: Added OpenAI API Base to config (useful while using OpenAI-compatible API like [LocalAI](https://github.com/go-skynet/LocalAI))
 - *15 Nov 2023*: Added support of [GPT-4 Turbo](https://help.openai.com/en/articles/8555510-gpt-4-turbo)
+- *2 Apt 2024*: Added [GPT-4 Vision](https://platform.openai.com/docs/guides/vision) support
 
 ## Bot commands
 - `/retry` – Regenerate last bot answer

bot/bot.py

@@ -218,18 +218,11 @@ async def _vision_message_handle_fn(
     try:
         # send placeholder message to user
         placeholder_message = await update.message.reply_text("...")
-        message = update.message.caption or update.message.text
+        message = update.message.caption or update.message.text or ''
 
         # send typing action
         await update.message.chat.send_action(action="typing")
 
-        if message is None or len(message) == 0:
-            await update.message.reply_text(
-                "🥲 You sent <b>empty message</b>. Please, try again!",
-                parse_mode=ParseMode.HTML,
-            )
-            return
-
         dialog_messages = db.get_dialog_messages(user_id, dialog_id=None)
         parse_mode = {"html": ParseMode.HTML, "markdown": ParseMode.MARKDOWN}[
             config.chat_modes[chat_mode]["parse_mode"]
@@ -271,7 +264,7 @@ async def _vision_message_handle_fn(
                 (n_input_tokens, n_output_tokens),
                 n_first_dialog_messages_removed,
             ) = gen_item
-            answer = current_model + " " + answer
+
             answer = answer[:4096]  # telegram message limit
 
             # update only when 100 new symbols are ready
@@ -416,7 +409,7 @@ async def message_handle(update: Update, context: CallbackContext, message=None,
             
             async for gen_item in gen:
                 status, answer, (n_input_tokens, n_output_tokens), n_first_dialog_messages_removed = gen_item
-                answer = current_model + " " + answer
+
                 answer = answer[:4096]  # telegram message limit
                     
                 # update only when 100 new symbols are ready
@@ -436,7 +429,7 @@ async def message_handle(update: Update, context: CallbackContext, message=None,
                 prev_answer = answer
             
             # update user data
-            new_dialog_message = {"user": _message, "bot": answer, "date": datetime.now()}
+            new_dialog_message = {"user": [{"type": "text", "text": _message}], "bot": answer, "date": datetime.now()}
 
             db.set_dialog_messages(
                 user_id,

bot/openai_utils.py

@@ -254,25 +254,32 @@ class ChatGPT:
         prompt = config.chat_modes[chat_mode]["prompt_start"]
 
         messages = [{"role": "system", "content": prompt}]
-        user_messages = {"role": "user", "content": []}
         
         for dialog_message in dialog_messages:
-            user_messages["content"].extend(dialog_message["user"])
+            messages.append({"role": "user", "content": dialog_message["user"]})
             messages.append({"role": "assistant", "content": dialog_message["bot"]})
-            
-
-        user_messages["content"].append({"type": "text", "text": message})
-
+                    
         if image_buffer is not None:
-            user_messages["content"].append(
+            messages.append(
                 {
-                    "type": "image",
-                    "image": self._encode_image(image_buffer),
+                    "role": "user", 
+                    "content": [
+                        {
+                            "type": "text",
+                            "text": message,
+                        },
+                        {
+                            "type": "image",
+                            "image": self._encode_image(image_buffer),
+                        }
+                    ]
                 }
+                
             )
+        else:
+            messages.append({"role": "user", "content": message})
 
-        response = messages + ([user_messages] if len(user_messages["content"]) > 0 else [])
-        return response
+        return messages
 
     def _postprocess_answer(self, answer):
         answer = answer.strip()