Adjust TokenserverOrigin variants

[data-sync-user]

TokenserverOrigin was originally added to track the Python vs Rust implementations when both ran in parallel for the Rust version’s rollout: https://github.com/mozilla-services/syncstorage-rs/issues/1245

~~This tracking via metrics is no longer necessary and can be removed.~~

However, now that the Python version on the Tokenserver service no longer runs, this field happens to also identify tokens created from the Tokenserver purge script. This can also be useful to track, we should make it clearer what the origin is: e.g. “purge_script” instead of “python”.

┆Issue is synchronized with this Jira Task