Adding Stanza for other languages than Swedish

[heatherleaf]

The Stanford NLP documentation recommends changing to Stanza when working in Python:

We are actively developing a Python package called Stanza, with state-of-the-art NLP performance enabled by deep learning. Besides, this package also includes an API for starting and making requests to a Stanford CoreNLP server. It is the recommended way to use Stanford CoreNLP in Python. (https://stanfordnlp.github.io/CoreNLP/other-languages.html#python)

And since Stanza comes with pre-trained models for many languages, I suggest we move to Stanza as the default annotator.

[anne17]

We started working on this a long time ago but never really finished. Things left to do:

• handling orphans
• implement batch-size-optimisations and gpu restrictions
• try to harmonise better with the Swedish stanza code (might be wise to break out some functionality into a util module)
• test for more languages than English (module might not work for languages that don't support named-entity recognition)