Telegram bot that reads messages, generates a first-order Markov chain from these, and then generates sentences based on the read messages, often to hilarious results.
Markov chain data is stored in a MongoDB database. The Telegram bot is implemented using simple-telegram-bot.
First, make sure that you have set up your Rust toolchain correctly. Check out the instructions in the official Rust documentation.
You need a Telegram bot token to actually run a bot. If you don't have one, get it from BotFather in Telegram.
Also make sure that you have a MongoDB database set up and running. Check out the MongoDB documentation for this.
Once you have all that, clone the repo if you haven't done so already, and
create an .env file with the needed values:
TELEGRAM_BOT_TOKEN=your-token-here
MONGODB_URL=192.168.56.101:27017
MONGODB_USER=someuser
MONGODB_PASSWORD=somepassword
MONGODB_DB=somedbname
Then cargo run normally, or cargo build depending on whether you want a
debug or a production build.
/markovgenerates a sentence based on the stored message information and sends it to the chat./markov_cleardeletes all message information for the current chat.- Any non-command message is split in words and stored as Markov chain
information to be used by the
/markovcommand.