Add spaCy en-core-web-lg

[tomaz-suller]

spaCy models are distributed as wheels on GitHub releases, and so they are not made available in either Conda or PyPI, which is why this addition references wheel file in the release directly.

We use the model to pre-process narrative text before modelling.

[klwetstone]

@tomaz-suller There are a few releases of spacy-model-en-core-web-lg on conda, eg. here and here. It looks like version 3.7.x is not available, but versions 3.8 and 3.5 are. Would that work for your solution? Installing via conda is preferable if at all possible.

[tomaz-suller]

Thanks for the pointers. I was bit afraid of those conda packages since I didn't know their precedence, and since they are not an official form of distribution. You could say I'm a bit afraid of supply-chain attacks, even if for no reason.

Anyway, I've found a way around this by packing the model along with the submission, which doesn't require modifying the runtime, and I was able to submit that way, so I'm closing this PR.

For reference, instructions on how to do that are here.

[klwetstone]

Excellent! I'm glad it worked out.