Withdrawal Remedy Explorer

This repo contains the files needed to deploy the Withdrawal Remedy Explorer app on Streamlit Sharing.

Contents

• prepare_data.py was used to filter and transform the source data, edges.parquet and nodes.parquet, and generate the files read by the app: remedy_edges.parquet, source_nodes.parquet, and target_nodes.parquet.

• streamlit_app.py runs the app.

Instructions

General users of the Withdrawal Remedy Explorer app should use the app on Streamlit Sharing rather than through this repo.

For app developers, here are some instructions to get started.

Dependency management

This repo includes both uv.lock and requirements.txt. This is an unusual choice, motivated by the deployment on Streamlit Sharing (at the time of this decision, Streamlit did not have functionality to use uv for dependency management).

This provides two options for running the app locally:

1. Create a virtual environment using uv venv and install the dependencies with uv sync.
2. Create a Python virtual environment using your tool of choice and install the dependencies with pip install -r requirements.txt.

The big downside is keeping these files in sync. If you add dependencies using uv add, make sure to then run uv export --format requirements-txt > requirements.txt.

Running the app locally

Run streamlit run streamlit_app.py.

Updating input data

• To update the input data, replace edges.parquet and nodes.parquet and run python prepare_data.py. Note that the data must be in the same format as the current version of edges.parquet and nodes.parquet. For more information on how to update these files, contact the maintainers.
• This repo only contains the files necessary to run the app. The rest of the files and data associated with this project are stored in a separate, private repo. If you think you need access to those, contact the maintainers.