Fix: use per request cache on tus metadata

[fenos]

What kind of change does this PR introduce?

Bug fix

What is the current behavior?

The metadata cache is stored in Map in memory.
When running multiple instances creates issues as it never gets invalidated

What is the new behavior?

Introducing a RequestCacheMetadata which is scoped to a single request.
This reduces the calls to fetch the metadata to 1 call per request instead of 8 per request.

[coveralls]

Pull Request Test Coverage Report for Build 7583015849

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.2%) to 82.976%

---

Totals
Change from base Build 7552611398:	0.2%
Covered Lines:	7882
Relevant Lines:	9393

---

💛 - Coveralls

[fenos]

Will wait for:

• @tus/server: Metadata on namingFunction + propagate error tus/tus-node-server#558
• @s3/store: allow providing custom cache implementation tus/tus-node-server#552

to remove the ported s3Store (to have the workaround for the S3 cache), and use the S3Store from the @tus/s3-store package.

Additionally, we can also remove the Metadata.parse call present on the namingFunction since it will be now be part of the @tus/server api 🎉

[github-actions]

🎉 This PR is included in version 0.46.1 🎉

The release is available on GitHub release

Your semantic-release bot 📦🚀