ParlaTI

Chatbot based on Knowledge base from ti.ch.

Setup

poetry shell
poetry install

Crawl data

This will run scrapy and fetch the data, that will be converted and saved as Markdown.

python -m scrapy runspider crawler.py

Generate embeddings

This ingest script will fetch all the documents in documents folder and compute the embeddings. Embeddings will be store in ChromaDB and persisted in the db folder.

python ingest.py

Test your data

Modify the file as you wish, just test your data with:

python query.py

Run the bot

This will run your telegram bot. Be sure to have a valid telegram token.