fix: decode the post.uuid before comparing it with other uuids

[dmaljovec]

What?

We encountered a problem where the uuid of our posts looks like a hex string, but it is being compared with bytes on the kr_dir and kp objects. This bit of logic ensures we are comparing apples to apples by forcing the post.uuid to be decoded if it is a hex string and coercing it to a bytes object as comparing a str to a bytes will never result in equality.

Why?

The reason for this change, is that any update to our knowledge repo becomes a reindex of all posts which means our consumers have to wait to see their new posts. This should drastically reduce the amount of work per reindex operation.

Description of changeset:

• Change how post.uuid is compared to existing data.

Test Plan:

• Make an edit in a repository and ensure the resulting update_index that gets triggered does not rewrite everything.