chat-with-markdown

Chat with markdown

Ingest a folder of .md files and create RAG system

TODO: everything below is in question, better use https://trieve.ai

Setup

Install pip

python3 -m pip install –upgrade pip

Install virtualenv

pip3 install virtualenv

Create virtualenv

virtualenv --python=python3.9 venv

Requirements

python

pip install -r requirements.txt

LLM inference service

This setup uses a cloudflare workers account with AI workers

In this case, you need to define the following environment variables:

• API_TOKEN
• API_BASE_URL

Get started

1. Place the markdown wikis in a folder called input_dir
2. Run python parse_mds.py --input_dir=input_dir/[NAME] where [NAME] is the root directory of the wiki you want to use. This will create a pickle file into a .wikis_pkl directory.
3. Run streamlit run app.py