refactor: Catalog update script

[muchzill4]

Changes

• Sever the update scripts coupling with topo internal
• Refactor template update scripts to separate GitHub source metadata, GitHub fetching, parsing, and writing concerns
• Make GITHUB_TOKEN optional, warning about possible rate limiting instead of failing immediately
• Resolve the catalog output path relative to the repo root so the script works from subdirectories

Going forward / known issues

• Decide what to do when a single template fails to update (right now it will be dropped)
• Validate the written catalog is schema compatible (we used to check x-topo.name, but we should be stricter)
• Organise catalog.*.json so that it's not internal to topo
• Source template list not from a const, but a file sibling to above
• Rename ref to sha everywhere and actually store sha's instead of main
• Setup CI job to run updater periodically

[awphi]

One last thing I was thinking about not changed here though - should the generated catalog JSON's schema link point to the schema at the ref from which it was generated?

[muchzill4]

One last thing I was thinking about not changed here though - should the generated catalog JSON's schema link point to the schema at the ref from which it was generated?

The plan is to use versioned schema eventually, I think pinning to a concrete ref will happen then.