Preprocessing from arrow file to load an HF dataset

[TevenLeScao]

This adds an option to launch preprocessing from an HF dataset (loaded from an arrow file for now as that's the use-case on JZ) rather than just jsonlines.

[TevenLeScao]

Update: this version is ~4 times slower with the change from column format to row format, compared to a hacky version that just gets the column for a certain feature. Including it as _hack.py.