Add autogenerated types dict to patch data

[Miepee]

Part 1 of not-sure-how-many that tries to add better typing/cleanup to the patcher data that is currently mostly just dict[Any, Any].

[duncathan]

seems as good a time as any to leave my general feedback here:

1. I strongly recommend using snake_case keys in the schema for better adherence to the python style guide in the generated TypedDict definitions
2. the KEY_WHATEVER constants are an antipattern imo. using normal string literals will be better for basically everything. if you want to rename a key, rename it in the generated file with your IDE's rename function and then update the schema
3. it's a bit surprising that so many keys in your schema aren't required. I find it's best to have as few optional fields as are necessary. I recommend using the ValidatorWithDefault tool we use in ODR, OPR, caver, etc. to automatically fill in defaults before validating the schema. then you can mark any field with a default as required and avoid messing around with get() and whatnot

[Miepee]

posted in discord but will post here for posterity too

seems as good a time as any to leave my general feedback here

generally agree with you on these these but would like to address them in seperate PRs.