Update README.md

Add support for GPT-4o

README.md

@@ -10,21 +10,19 @@
 <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.
 
 This repo is ChatGPT re-created as Telegram Bot. **And it works great.**
 
-You can deploy your own bot, or use mine: [@chatgpt_karfly_bot](https://t.me/chatgpt_karfly_bot)
+You can deploy your own bot, or use mine: [@jadvebot](https://t.me/jadvebot) (Our web: https://jadve.com)
 
 ## Features
 - Low latency replies (it usually takes about 3-5 seconds)
 - 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
@@ -108,7 +107,7 @@ You can be in this list:
 
 ## Contributors
 - Main contributor: @karfly
-- [Father.Bot](https://father.bot).
+- [Jadve AI](https://jadve.com).
 
 ## References
 1. [*Build ChatGPT from GPT-3*](https://learnprompting.org/docs/applied_prompting/build_chatgpt)

bot/bot.py

@@ -185,9 +185,9 @@ async def _vision_message_handle_fn(
     user_id = update.message.from_user.id
     current_model = db.get_user_attribute(user_id, "current_model")
 
-    if current_model != "gpt-4-vision-preview":
+    if current_model != "gpt-4-vision-preview" and current_model != "gpt-4o":
         await update.message.reply_text(
-            "🥲 Images processing is only available for <b>gpt-4-vision-preview</b> model. Please change your settings in /settings",
+            "🥲 Images processing is only available for <b>gpt-4-vision-preview</b> and <b>gpt-4o</b> model. Please change your settings in /settings",
             parse_mode=ParseMode.HTML,
         )
         return
@@ -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"]
@@ -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,
@@ -466,11 +459,14 @@ async def message_handle(update: Update, context: CallbackContext, message=None,
             await update.message.reply_text(text, parse_mode=ParseMode.HTML)
 
     async with user_semaphores[user_id]:
-        if current_model == "gpt-4-vision-preview" or update.message.photo is not None and len(update.message.photo) > 0:
-            logger.error('gpt-4-vision-preview')
-            if current_model != "gpt-4-vision-preview":
-                current_model = "gpt-4-vision-preview"
-                db.set_user_attribute(user_id, "current_model", "gpt-4-vision-preview")
+        if current_model == "gpt-4-vision-preview" or current_model == "gpt-4o" or update.message.photo is not None and len(update.message.photo) > 0:
+
+            logger.error(current_model)
+            # What is this? ^^^
+
+            if current_model != "gpt-4o" and current_model != "gpt-4-vision-preview":
+                current_model = "gpt-4o"
+                db.set_user_attribute(user_id, "current_model", "gpt-4o")
             task = asyncio.create_task(
                 _vision_message_handle_fn(update, context, use_new_dialog_timeout=use_new_dialog_timeout)
             )
@@ -876,4 +872,4 @@ def run_bot() -> None:
 
 
 if __name__ == "__main__":
-    run_bot()
+    run_bot()

bot/openai_utils.py

@@ -26,7 +26,7 @@ OPENAI_COMPLETION_OPTIONS = {
 
 class ChatGPT:
     def __init__(self, model="gpt-3.5-turbo"):
-        assert model in {"text-davinci-003", "gpt-3.5-turbo-16k", "gpt-3.5-turbo", "gpt-4", "gpt-4-1106-preview", "gpt-4-vision-preview"}, f"Unknown model: {model}"
+        assert model in {"text-davinci-003", "gpt-3.5-turbo-16k", "gpt-3.5-turbo", "gpt-4", "gpt-4o", "gpt-4-1106-preview", "gpt-4-vision-preview"}, f"Unknown model: {model}"
         self.model = model
 
     async def send_message(self, message, dialog_messages=[], chat_mode="assistant"):
@@ -37,7 +37,7 @@ class ChatGPT:
         answer = None
         while answer is None:
             try:
-                if self.model in {"gpt-3.5-turbo-16k", "gpt-3.5-turbo", "gpt-4", "gpt-4-1106-preview", "gpt-4-vision-preview"}:
+                if self.model in {"gpt-3.5-turbo-16k", "gpt-3.5-turbo", "gpt-4", "gpt-4o", "gpt-4-1106-preview", "gpt-4-vision-preview"}:
                     messages = self._generate_prompt_messages(message, dialog_messages, chat_mode)
 
                     r = await openai.ChatCompletion.acreate(
@@ -78,7 +78,7 @@ class ChatGPT:
         answer = None
         while answer is None:
             try:
-                if self.model in {"gpt-3.5-turbo-16k", "gpt-3.5-turbo", "gpt-4", "gpt-4-1106-preview"}:
+                if self.model in {"gpt-3.5-turbo-16k", "gpt-3.5-turbo", "gpt-4","gpt-4o", "gpt-4-1106-preview"}:
                     messages = self._generate_prompt_messages(message, dialog_messages, chat_mode)
 
                     r_gen = await openai.ChatCompletion.acreate(
@@ -138,7 +138,7 @@ class ChatGPT:
         answer = None
         while answer is None:
             try:
-                if self.model == "gpt-4-vision-preview":
+                if self.model == "gpt-4-vision-preview" or self.model == "gpt-4o":
                     messages = self._generate_prompt_messages(
                         message, dialog_messages, chat_mode, image_buffer
                     )
@@ -186,7 +186,7 @@ class ChatGPT:
         answer = None
         while answer is None:
             try:
-                if self.model == "gpt-4-vision-preview":
+                if self.model == "gpt-4-vision-preview" or self.model == "gpt-4o":
                     messages = self._generate_prompt_messages(
                         message, dialog_messages, chat_mode, image_buffer
                     )
@@ -254,25 +254,36 @@ 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_url",
+                            "image_url" : {
+                              
+                                "url": f"data:image/jpeg;base64,{self._encode_image(image_buffer)}",
+                                "detail":"high"
+                            }
+                        }
+                    ]
                 }
+                
             )
+        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()
@@ -296,6 +307,9 @@ class ChatGPT:
         elif model == "gpt-4-vision-preview":
             tokens_per_message = 3
             tokens_per_name = 1
+        elif model == "gpt-4o":
+            tokens_per_message = 3
+            tokens_per_name = 1
         else:
             raise ValueError(f"Unknown model: {model}")
 
@@ -347,4 +361,4 @@ async def generate_images(prompt, n_images=4, size="512x512"):
 
 async def is_content_acceptable(prompt):
     r = await openai.Moderation.acreate(input=prompt)
-    return not all(r.results[0].categories.values())
+    return not all(r.results[0].categories.values())

config/models.yml

@@ -1,4 +1,4 @@
-available_text_models: ["gpt-3.5-turbo", "gpt-3.5-turbo-16k", "gpt-4-1106-preview", "gpt-4-vision-preview", "gpt-4", "text-davinci-003"]
+available_text_models: ["gpt-3.5-turbo", "gpt-3.5-turbo-16k", "gpt-4-1106-preview", "gpt-4-vision-preview", "gpt-4", "text-davinci-003", "gpt-4o"]
 
 info:
   gpt-3.5-turbo:
@@ -65,6 +65,18 @@ info:
       smart: 5
       fast: 4
       cheap: 3
+  gpt-4o:
+    type: chat_completion
+    name: GPT-4o
+    description: GPT-4o is a special variant of GPT-4 designed for optimal performance and accuracy. Suitable for complex and detailed tasks.
+
+    price_per_1000_input_tokens: 0.03
+    price_per_1000_output_tokens: 0.06
+
+    scores:
+      smart: 5
+      fast: 2
+      cheap: 2
 
   text-davinci-003:
     type: completion
@@ -85,4 +97,4 @@ info:
 
   whisper:
     type: audio
-    price_per_1_min: 0.006
+    price_per_1_min: 0.006