A local web-based user-interface for exploratory feature selection analysis using the ReaFS library. Designed for interactive model diagnosis and feature ranking in regression tasks and part of the NaviCat catalyst discovery platform.
See pyproject.toml for complete dependency specifications. Most of the dependencies are installed with reafs.
# Clone or download this repo
cd reafs-dive
# Install dependencies with uv (creates virtual environment automatically)
uv syncThe reafs package is expected to be located in the parent directory. Modify pyproject.toml accordingly.
# Run the Dash app
uv run main.pyThen open http://localhost:8050 in your browser.
See the supplementary information of the main manuscript for a description.
The example folder holds data example/oa_data.zip ready for visualization.
This project was created as part of NCCR Catalysis (grant number 225147), a National Centre of Competence in Research funded by the Swiss National Science Foundation.
Schoepfer, A. A., Laplaza, R., Waser, J., & Corminboeuf, C. ACS Catal., 2026, 18114. https://doi.org/10.1021/acscatal.6c04292