Added switch for legacy-data, new parallel processing and progress bars

[Alessi0X]

As of late 2025, the openalex-snapshot features two major subfolders: data and legacy-data.

By default, the flatten script operates on data. Yet, for keeping the legacy and for reproducibility, one might need to refer to legacy data.

This modification of the flatten script allows the user, by changing the isDataLegacy variable, to drive the flattening script towards using the legacy-data subfolder (variable to be set True) or the default data subfolder (variable to set False).

Second, I have included a new script that does the flattening in parallel: the seven flatten_*() functions are executed in parallel thanks to the multiprocess library.

Lastly, the original serial flattening script had quite a lot of print() to monitor progress, which have been substituted with nicer tqdm progress bars.