Add `Incident` categorization model

[michplunkett]

Describe your changes

Add multi-label classification model to the scraper so that incidents categorized as Information can be more correctly utilized.

Checklist before requesting a review

• [x] The code runs successfully.

python -m incident_scraper build-model
Accuracy Score: 0.7074224021592442
Precision Score: 0.8922349427575909
Recall Score: 0.7796868203566768

GUESS WHO IS ABOVE 70%, MY DUDES!?

[michplunkett]

Any new additions to this need to wait until exam season is over. The thought of how this is fixed though, is the following:

• Save the vectorizer like the model is being saved via pickle
• Load the old, pre-fit vectorizer when not building
• Vectorize incoming strings and do the thing

🤞🏻 🤞🏻 🤞🏻 🤞🏻 🤞🏻

EDIT: IT WORRRRKED