Potential switching from stringified IDs to using uuid for sanity checking.

[AbstractUmbra]

Whilst I'm personally a fan of not restricting types further than needed, I've hit a rare issue where a user provided bad input and it got me wondering.

In essence, we have let's say the Get Manga endpoint under GET /manga/:id. Path traversal in URLs is a common pitfall and this is what happened. We got input of . for :id. This means instead we hit GET /manga which is a collection endpoint versus the expected entity endpoint.

My proposed solution to this would be using uuid.UUID() at the library level and accepting both types as arguments, e.g. input_: str | UUID and then transforming all string input to UUID transparently, which would resolve this issue.

Currently I'm wondering if this is a non-issue and shouldn't be handled by the library, rather than by the consumers. More investigation is needed.