Credit Scorecard Builder

Tools for building and exploring clinical risk scorecards. The main entry point is the Streamlit UI in UserInterface/credit_score_app.py.

Quickstart

1) Create an environment

python3 -m venv venv
source venv/bin/activate

2) Install dependencies

From the repository root (CreditScorecardBuilder):

python3 -m pip install -U pip
python3 -m pip install -r requirements.txt

requirements.txt installs the local llm_utils package via ./llm_utils.

3) Configure data directory

The app looks for data in /data (Docker) or via a devcontainer mount. If running locally, set a path explicitly before launch by editing config/config.py and setting Config._USER_DATA_DIR to a valid directory.

4) Run the Streamlit app

streamlit run UserInterface/credit_score_app.py

Documentation

Preview locally

mkdocs serve