Crunchy is a generic python-based platform for constructing and executing automated and asynchronous (multi-threaded) processing workflows.

At its core, crunchy watches input directories for new files or directories that match user-defined patterns and, once their file size is stable (e.g., if data is being copied), launches custom processing scripts (termed workflows) using a pool of worker threads that facilitate parallelisation.

For an introduction into the structure of crunchy and how to use it to create and run customised workflows, please see the demonstration.ipynb file.

Documentation for crunchy can be found here:

https://samthiele.github.io/crunchy/crunchy.html

If crunchy has been useful for your work, please cite:

Thiele, S., Kirsch, M., Lorenz, S., Saffi, H., El Alami, S., Contreras Acosta, I.C., Madriz, Y., and Gloaguen, R., 2024. "Maximising the value of hyperspectral drill core scanning through real-time processing and analysis", Frontiers in Earth Science doi:10.3389/feart.2024.1433662