[FT] Multilingual Extractiveness Metrics

[rolshoven]

Issue encountered

Currently, the Extractiveness metric uses the en_core_web_sm spaCy model, making it an English-only metric. For our evaluations, we need to use that metric with German, French, and Italian spaCy models.

Solution/Feature

Add a parameter to the Extractiveness metric, that defines the language, and load the corresponding spaCy model

As part of a community task I have been collaborating on, I've implemented the changes because they were required for our evaluations. I will open a PR right away.