use pydantic tagged unions to more quickly validate the result document

[williballenthin]

In our result document JSON document, we use unions in a few places, such as freeze.features.Feature:

capa/capa/features/freeze/features.py

Line 351 in c409b2b

Feature = Union[

I've learned that to validate incoming data against this union, pydantic will loop through each case and then use the best match. because there are lots of cases, pydantic does a lot of work.

we should be able to improve performance by using tagged unions. there's a bit of documentation here:
https://docs.pydantic.dev/latest/concepts/unions/#discriminated-unions

and a PyCon US '24 talk here: https://www.youtube.com/watch?v=Qvj5e9xtaSE

we should investigate using tagged unions and see if it improves performance of loading serialized result documents. granted, this isn't a common operation, but the data model would be more correct regardless.

[williballenthin]

I think this would do it:

Feature = Annotated[Union[
    OSFeature,
    ArchFeature,
    ...
    BasicBlockFeature,
], Field(discriminator='type')]