Repo for Q&A bot with chat interface and custom data ingestion with vector database. Built using LangChain
and Streamlit
.
Make sure you have at least python 3.8
and install dependencies by running pip(3) install -r requirements.txt
You can check out the configuration file in src/cfg/default.cfg
. It currently has one sample site vero.fi
which scrapes the Finnish tax office website for Finnish tax regulation, guidelines and other related information.
Right now we only support sitemap ingestion, but more methods will be added in the future.
- Run
python(3) src/ingest_data.py --site vero.fi [--debug]
(debug switch will only scrape a tiny portion of the site so testing can be rapid)
- Export your OpenAI API key as an environment variable
export OPENAI_API_KEY=YOUR_KEY_HERE
- Run
streamlit run src/app.py
- Open your browser at
localhost:8501?site=vero.fi