ExtendDDG: Expanding Automated Dataset Description Generation with Supplemental Dataset Documentation

ExtendDDG is an extension of AutoDDG, developed by the NYU VIDA Lab, to supplement automated dataset description generation with a dataset's associated documentation, such as codebooks, methodology reports, and READMEs.

Quick Start

Requirements

• Python >=3.10
• AutoDDG

Run

uv sync

Run the example Python script.

cd examples
uv run python quick_start.py

Datasets

We have compiled a set of publicly available datasets that contain supplemental documentation to be used for testing. A summary table of these datasets and their associated documentation can be found here.

Development

Code Quality

uv run ruff check
uv run ruff format