Recipe details need to be in a separate config file

[kasra-keshavarz]

The recipe details, including keywords, paths, scripts, and parallelization details, need to stay out of the main extract-dataset.sh script and instead reside in a configuration file.

This simplifies development by other working groups.