Better bucket regexes

[TheTechromancer]

Summary

Switch storage bucket regexes from positional capture groups to named groups, and surface region info where the hostname carries it.

Previously, every STORAGE_BUCKET_HOSTNAME regex used a fixed contract: group 1 = bucket name, group 2 = trailing domain. That breaks down once some providers expose region in the URL and others don't — consumers can't tell which group means what without per-provider logic. With named groups, consumers just call match.groupdict() and read name / region (the latter only present when meaningful).

This is a breaking change for downstream consumers indexing by group number, so the artifact is published as cloud_providers_v3.json and the package version is bumped to 10.0.0. cloud_providers_v2.json is left in place, frozen, so legacy consumers keep resolving until they migrate.

Changes

• 6 providers updated to use (?P<name>…) (and (?P<region>…) where applicable):
  • Amazon — split into 3 patterns covering legacy global (bucket.s3.amazonaws.com), dash-style (bucket.s3-us-west-2.amazonaws.com), and dot-style (bucket.s3.eu-central-1.amazonaws.com).
  • DigitalOcean — region (nyc3, sfo2, …) broken out of the trailing domain.
  • Hetzner — rewritten to capture region (fsn1, nbg1, hel1); the previous regex matched the wrong URL shape.
  • Google / Microsoft / Cloudflare — name-only (no region in their bucket URLs).
• Artifact rename: cloud_providers_v2.json → cloud_providers_v3.json across cloudcheck_update, src/lib.rs (signature URL + cache path), scripts/update_readme_table.py, the daily-update workflow, and the README.
• Version bump: pyproject.toml and Cargo.toml → 10.0.0.
• Tests: parametrized coverage in test_cloudcheck.py exercising real hostnames per provider and verifying groupdict(), plus a sweep that asserts every provider's regexes compile and expose a name group.

Test plan

• uv run pytest test_cloudcheck.py -k "regex or import_provider" — the new + existing regex tests pass.
• uv run python -m cloudcheck_update.cli — regenerates cloud_providers_v3.json with the new regex strings round-tripped intact.
• After merge, the daily-update workflow publishes cloud_providers_v3.json to stable; the network-dependent test_lookup_* tests resume passing once the file is live at the new URL.

Relevant:

• Add Hetzner Object Storage bucket-cataloger module (#1876) bbot#3073