Add methods to generate notable move categories

[shrianshChari]

Partially fixes #6

[scheibo]

Thanks for attempting this, but theres quite a lot that needs to be fixed here.

1. You haven't added any tests, if you did you would see your sets disagree with those that already exist
2. You aren't actually using any of the tables youre computing
3. You cannot use @pkmn/dex outside of tests in this codebase (and why would you want to?)
4. Please do not open PRs before running the linter and test suite locally
5. This PR does not close Update the classifier to support modern Generations #6, there are many more tables that need to be replaced. You also seem to be continuing to work on the PR, please open the PR as a draft if it is actually not finished.

I've pushed some changes that should start you on a better footing. Note that Antar's sets might be incorrect, if so just change the test asserts to be like to be like:

 expect(classifier.computeGreaterSetupMoves(GEN)).toEqual([...classifier.GREATER_SETUP_MOVES, ...additionalMovesAnatarShoildHaveAdded]);

etc

[shrianshChari]

The other lists don't contain fields that I can use to filter the correct moves; I have hypothesized that I can try and filter based on the shortDesc of moves, items, and abilities but will leave that for other PRs.

[scheibo]

Great work

I have hypothesized that I can try and filter based on the shortDesc of moves, items, and abilities but will leave that for other PRs.

A clever idea, but I dont think we should rely on the descriptions as they are completely subjective