aljaz/chatgpt-telegram-bot
1
1
Fork 0
mirror of https://github.com/aljazceru/chatgpt-telegram-bot.git synced 2025-12-23 15:44:54 +01:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'n3d1117:main' into user-budgets

Browse Source
This commit is contained in:
Alexander Hinzer
2023-03-15 01:18:03 +01:00
committed by GitHub
parent 60d1dba0a5 f54b6a46b2
commit d9f49440c2
5 changed files with 48 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
.env.example
View File

@@ -29,6 +29,20 @@ MAX_CONVERSATION_AGE_MINUTES=180
# Whether to answer to voice messages with the transcript or with a ChatGPT response of the transcript
VOICE_REPLY_WITH_TRANSCRIPT_ONLY=true
# How many chat completion choices to generate for each input message
N_CHOICES=1
# Number between 0 and 2. Higher values like 0.8 will make the output more random,
# while lower values like 0.2 will make it more focused and deterministic
TEMPERATURE=1.0
# The DALL·E generated image size
IMAGE_SIZE="512x512"
# Group trigger keyword, if set, the bot will only respond to messages that start with this keyword
# Useful for bots added to groups with privacy mode disabled
GROUP_TRIGGER_KEYWORD=""
# USD-price per 1000 tokens for cost information in usage statistics, see https://openai.com/pricing
TOKEN_PRICE=0.002
@@ -36,4 +50,4 @@ TOKEN_PRICE=0.002
IMAGE_PRICES="0.016,0.018,0.02"
# USD-price for 1 minute of audio transcription for cost information in usage statistics
TRANSCRIPTION_PRICE=0.006
TRANSCRIPTION_PRICE=0.006

7
Dockerfile
View File

@@ -1,7 +1,8 @@
FROM python:3.10-slim
RUN apt-get update
RUN apt-get install ffmpeg -y
RUN apt-get update && \
apt-get install ffmpeg -y && \
rm -rf /var/lib/apt/lists/*
RUN useradd -m appuser
USER appuser
@@ -9,7 +10,7 @@ WORKDIR /home/appuser/
ENV PATH="/home/appuser/.local/bin:$PATH"
RUN pip install --user pipenv
RUN pip install --user pipenv --no-cache-dir
WORKDIR /home/appuser/app
COPY . .

34
README.md
View File

@@ -6,26 +6,22 @@
A [Telegram bot](https://core.telegram.org/bots/api) that integrates with OpenAI's _official_ [ChatGPT](https://openai.com/blog/chatgpt/) APIs to provide answers. Ready to use with minimal configuration required.
## Screenshots
![demo.pdf](https://github.com/n3d1117/chatgpt-telegram-bot/files/10900400/demo.pdf)
![demo](https://user-images.githubusercontent.com/11541888/225114786-0d639854-b3e1-4214-b49a-e51ce8c40387.png)
## Features
- [x] Support markdown in answers
- [x] Reset conversation with the `/reset` command
- [x] Typing indicator while generating a response
- [x] Access can be restricted by specifying a list of allowed users
- [x] Docker support
- [x] Proxy support
- [x] (NEW!) Support multiple answers via the `n_choices` configuration parameter
- [x] (NEW!) Customizable model parameters (see [configuration](#configuration) section)
- [x] (NEW!) See token usage after each answer
- [x] (NEW!) Multi-chat support
- [x] Docker and Proxy support
- [x] (NEW!) Image generation using DALL·E via the `/image` command
- [x] (NEW!) Transcribe audio and video messages using Whisper (may require [ffmpeg](https://ffmpeg.org))
- [x] (NEW!) Automatic conversation summary to avoid excessive token usage (fixes [#34](https://github.com/n3d1117/chatgpt-telegram-bot/issues/34))
- [x] (NEW!) Group chat support with inline queries
- To use this feature, enable inline queries for your bot in BotFather via the `/setinline` [command](https://core.telegram.org/bots/inline)
- [x] (NEW!) Track token usage per user (stored in /token_usage/`user_id`.json)
- [x] (NEW!) Get personal token usage statistics and cost per day/month via the `/stats` command
- [x] (NEW!) Track token usage per user (stored in /token_usage/`user_id`.json) - by [@AlexHTW](https://github.com/AlexHTW)
- [x] (NEW!) Get personal token usage statistics and cost per day/month via the `/stats` command - by [@AlexHTW](https://github.com/AlexHTW)
- [x] (NEW!) Group chat trigger keyword - by [@tracking](https://github.com/tracking)
## Additional features - help needed!
- [ ] Add stream support ([#43](https://github.com/n3d1117/chatgpt-telegram-bot/issues/43))
@@ -56,6 +52,10 @@ MAX_TOKENS=2000 # Defaults to 1200
MAX_HISTORY_SIZE=15 # Defaults to 10
MAX_CONVERSATION_AGE_MINUTES=120 # Defaults to 180 (3h)
VOICE_REPLY_WITH_TRANSCRIPT_ONLY=false # Defaults to true
N_CHOICES=1 # Defaults to 1
TEMPERATURE=1.0 # Defaults to 1.0
IMAGE_SIZE="256x256" # Defaults to 512x512
GROUP_TRIGGER_KEYWORD="@bot" # Defaults to "" (no keyword required)
TOKEN_PRICE=0.002 # Defaults to 0.002, current price: https://openai.com/pricing
IMAGE_PRICES="0.016,0.018,0.02" # Defaults to OpenAI Dall-E pricing for sizes 256x256,512x512,1024x1024
TRANSCRIPTION_PRICE=0.006 # Defaults to minute price of OpenAI Whisper of 0.006
@@ -70,6 +70,10 @@ TRANSCRIPTION_PRICE=0.006 # Defaults to minute price of OpenAI Whisper of 0.006
* `MAX_HISTORY_SIZE`: Max number of messages to keep in memory, after which the conversation will be summarised to avoid excessive token usage ([#34](https://github.com/n3d1117/chatgpt-telegram-bot/issues/34))
* `MAX_CONVERSATION_AGE_MINUTES`: Maximum number of minutes a conversation should live, after which the conversation will be reset to avoid excessive token usage
* `VOICE_REPLY_WITH_TRANSCRIPT_ONLY`: Whether to answer to voice messages with the transcript only or with a ChatGPT response of the transcript ([#38](https://github.com/n3d1117/chatgpt-telegram-bot/issues/38))
* `N_CHOICES`: Number of answers to generate for each input message
* `TEMPERATURE`: Number between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic
* `IMAGE_SIZE`: The DALL·E generated image size. Allowed values: "256x256", "512x512", or "1024x1024"
* `GROUP_TRIGGER_KEYWORD`: If set, the bot will only respond to messages that start with this keyword. This is useful for bots added to groups with privacy mode disabled. **Note**: by default, *no keyword* is required (`""`)
* `TOKEN_PRICE`: USD-price per 1000 tokens for cost information in usage statistics
@@ -80,23 +84,13 @@ TRANSCRIPTION_PRICE=0.006 # Defaults to minute price of OpenAI Whisper of 0.006
# 'gpt-3.5-turbo' or 'gpt-3.5-turbo-0301'
'model': 'gpt-3.5-turbo',
# Number between 0 and 2. Higher values like 0.8 will make the output more random,
# while lower values like 0.2 will make it more focused and deterministic. Defaults to 1
'temperature': 1,
# How many answers to generate for each input message. Defaults to 1
'n_choices': 1,
# Number between -2.0 and 2.0. Positive values penalize new tokens based on whether
# they appear in the text so far, increasing the model's likelihood to talk about new topics. Defaults to 0
'presence_penalty': 0,
# Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing
# frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. Defaults to 0
'frequency_penalty': 0,
# The DALL·E generated image size. 256x256, 512x512, or 1024x1024. Defaults to 512x512
'image_size': '512x512'
'frequency_penalty': 0
}
```
</details>

16
main.py
View File

@@ -33,27 +33,20 @@ def main():
'max_conversation_age_minutes': int(os.environ.get('MAX_CONVERSATION_AGE_MINUTES', 180)),
'assistant_prompt': os.environ.get('ASSISTANT_PROMPT', 'You are a helpful assistant.'),
'max_tokens': int(os.environ.get('MAX_TOKENS', 1200)),
'n_choices': int(os.environ.get('N_CHOICES', 1200)),
'temperature': float(os.environ.get('TEMPERATURE', 1.0)),
'image_size': os.environ.get('IMAGE_SIZE', '512x512'),
# 'gpt-3.5-turbo' or 'gpt-3.5-turbo-0301'
'model': 'gpt-3.5-turbo',
# Number between 0 and 2. Higher values like 0.8 will make the output more random,
# while lower values like 0.2 will make it more focused and deterministic.
'temperature': 1,
# How many chat completion choices to generate for each input message.
'n_choices': 1,
# Number between -2.0 and 2.0. Positive values penalize new tokens based on whether
# they appear in the text so far, increasing the model's likelihood to talk about new topics.
'presence_penalty': 0,
# Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing
# frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.
'frequency_penalty': 0,
# The DALL·E generated image size
'image_size': '512x512'
'frequency_penalty': 0
}
telegram_config = {
@@ -63,6 +56,7 @@ def main():
'monthly_guest_budget': float(os.environ.get('MONTHLY_GUEST_BUDGET', '100.0')),
'proxy': os.environ.get('PROXY', None),
'voice_reply_transcript': os.environ.get('VOICE_REPLY_WITH_TRANSCRIPT_ONLY', 'true').lower() == 'true',
'group_trigger_keyword': os.environ.get('GROUP_TRIGGER_KEYWORD', ''),
'token_price': float(os.environ.get('TOKEN_PRICE', 0.002)),
'image_prices': [float(i) for i in os.environ.get('IMAGE_PRICES',"0.016,0.018,0.02").split(",")],
'transcription_price': float(os.environ.get('TOKEN_PRICE', 0.002)),

11
telegram_bot.py
View File

@@ -258,9 +258,18 @@ class ChatGPT3TelegramBot:
logging.info(f'New message received from user {update.message.from_user.name}')
chat_id = update.effective_chat.id
prompt = update.message.text
if self.is_group_chat(update):
trigger_keyword = self.config['group_trigger_keyword']
if prompt.startswith(trigger_keyword):
prompt = prompt[len(trigger_keyword):].strip()
else:
logging.warning(f'Message does not start with trigger keyword, ignoring...')
return
await context.bot.send_chat_action(chat_id=chat_id, action=constants.ChatAction.TYPING)
response, total_tokens = self.openai.get_chat_response(chat_id=chat_id, query=update.message.text)
response, total_tokens = self.openai.get_chat_response(chat_id=chat_id, query=prompt)
# add chat request to users usage tracker
self.usage[update.message.from_user.id].add_chat_tokens(total_tokens, self.config['token_price'])