ENH/CLN: Rewrite `crowsetta.data` to be like `vocalpy.examples`

[NickleDave]

Not sure why I made the choices I did when implementing crowsetta.data originally -- it's too clever by half.

From the vocalpy.examples module I have now I don't see any reason to do what I did with having separate packages, the ability to download to a specified user directory, etc.

This makes it very annoying to actually get example data

Also, it was a mistake to name the examples as formats, as this means

• we only have one example per format
• or we give some example a not-very-meaningful name like simple-seq2

So the to-do list here is:

• [ ] literally rewrite the API to be like vocalpy.example / vocalpy.examples where we can call vocalpy.example('name') and get back either the annotation or the path the annotation file, and we can call vocalpy.examples.show
• [ ] give the examples very specific names and include the annotation format in the metadata that's listed when we call show

This is technically a breaking change that should require a version bump but ... I have a feeling it won't affect that many people, even if there were a million crowsetta users, no-one is, uh, relying on the data API in production, as it were. So I will just do this as part of 5.1