Added a script to restore punctuation

[Paul-HenriBJT]

Script to restore the punctuation. It uses https://github.com/oliverguhr/fullstop-deep-punctuation-prediction and community models derived from it.
Works with the following languages: Catalan, English, German, French, Spanish, Bulgarian, Italian, Polish, Dutch, Czech, Portuguese, Slovak and Slovenian.

We could maybe pass the path of the punctuation model as an arg to the main function.

[Vaibhavs10]

LGTM! - I'm using similar-ish models for an internal script.

[ylacombe]

Thanks for working on this @Paul-HenriBJT, I've left some tiny questions, but mostly for understanding and making sure it's optimized and easy to use !

[ylacombe]

Are there other ways to download this ?

[Paul-HenriBJT]

One other solution would be to remove the spacy models from the requirements.txt file and download them in the casing.py file using something like this:
from spacy.cli import download

MODELS = [
"ca_core_news_sm",
"en_core_web_sm",
"de_core_news_sm",
"fr_core_news_sm",
"es_core_news_sm",
"it_core_news_sm",
"pl_core_news_sm",
"nl_core_news_sm",
"pt_core_news_sm"
]

def setup_models():
for model in MODELS:
download(model)

if name == "main":
setup_models()

[ylacombe]

I'd like this better tbh