ChatSummaryBot

ChatSummaryBot is a simple bot that summarizes chat messages in group chats. Currently, it supports Telegram, but it can be easily extended to other platforms.

Features

• Summarizes chat messages from Telegram channels.
• Supports multiple summarization models: Ollama and Groq.
• Configurable via a YAML file.
• Generates summaries using emojis, bullet points, and other visual elements.
• Provides daily summaries and highlights active participants.
• Extracts and summarizes links and attachments.
• Bots generate messages and send them every 24 hours.
• Supports channels with subtopics (subforms), processing each subtopic separately and generating summaries per topic.
• Ability to publish summaries to a dedicated Telegram channel:
  • Summary from each search channel is posted as a new topic in the target channel.
  • Requires summary_channel_id key in the configuration file.

Requirements

• Python 3.12 or higher. Other version of Python 3 may work, but not tested.
• Telegram API credentials (API ID and API Hash)
• Supported summarization models (Ollama or Groq)

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/ChatSummaryBot.git
   cd ChatSummaryBot

2. Create a virtual environment and activate it:

   python -m venv venv # or python3.12 -m venv venv 
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install the required dependencies:

   pip install -r requirements.txt

Configuration

1. Copy the example configuration file and update it with your details:

   cp config.example.yaml config.yaml

2. Edit config.yaml and fill in the required fields:

   session_name: chatSummary
   api_id: "YOUR_API_ID"
   api_hash: "YOUR_API_HASH"
   model_provider: ollama  # or groq
   model_name: llama3.2  # or your preferred model
   max_length: 4000
   summarization_frequency: 24  # in hours
   output_dir: summaries
   channels:
     - id: -1001235860760
       name: IT_TALKS_BENELUX
       language: russian
       context: "This is a channel for software engineers working in the Benelux region, discussing various topics. The channel contains a lot of non-serious discussion, but it's known to be useful for finding a job or getting real insights."
       filters:
         keywords: ["tech", "benelux"]
   
     - id: -1001157394889
       name: LEETCODE_HEROES
       filters:
         keywords: ["leetcode", "interview"]

Model Configuration

Groq Cloud

This option is recommended because it allows for larger models, and Groq cloud offers a free tier that is sufficient for personal use cases.

1. Sign up for a Groq Cloud account and obtain your API key.
2. Set the GROQ_API_KEY environment variable:

   export GROQ_API_KEY="YOUR_GROQ_API_KEY"

Ollama Locally

1. Install the Ollama model locally by following the instructions on the Ollama website.
2. Ensure the model is running and accessible on your local machine.

brew install ollama
ollama pull llama3.2
ollama serve

Usage

1. Run the summarizer bot. Omit GROQ_API_KEY if you are using the Ollama model.

   GROQ_API_KEY=1231 python summarizeChat.py

2. The bot will start processing the configured channels and generate summaries based on the specified frequency.

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

This project is licensed under the MIT License. See the LICENSE file for details.