Skip to content

DRAFT: remove migrated framework from the hub, reframe as umbrella#655

Open
rain-Brian wants to merge 3 commits into
mainfrom
seo/hub-remove-framework
Open

DRAFT: remove migrated framework from the hub, reframe as umbrella#655
rain-Brian wants to merge 3 commits into
mainfrom
seo/hub-remove-framework

Conversation

@rain-Brian

Copy link
Copy Markdown
Collaborator

⚠️ DRAFT for @zhmiao review — do not merge yet

Completes the repo split: removes from the Biodiversity hub the framework code that now lives in dedicated repos, and reframes the hub homepage as the ecosystem umbrella. Merge gate: do not merge until the PytorchWildlife PyPI release is cut from microsoft/Pytorch-Wildlife (PR #17 + the staged release), so pip install PytorchWildlife resolves there.

Removed (verified migrated + parity-checked)

Removed from hub Now lives at Parity
PytorchWildlife/ + setup.py + version.txt + MANIFEST.in microsoft/Pytorch-Wildlife complete superset after PR #17 (diff -rq clean)
PW_Bioacoustics/ microsoft/MegaDetector-Acoustic byte-for-byte identical, no hub drift
docs/base/ + docs/fine_tuning_modules/ (mkdocstrings API ref) PW docs site (PR #18) 32 pages migrated; removed mkdocstrings plugin + deps + nav

Reframed

  • docs/index.md is now the Microsoft Biodiversity umbrella homepage (was "Welcome to PyTorch-Wildlife"); points the framework to microsoft/Pytorch-Wildlife and leads with the ecosystem. Top nav group renamed PyTorch-WildlifeMicrosoft Biodiversity.

⚠️ Left in place and FLAGGED (NOT safe to auto-delete — needs your call)

Verification

  • mkdocs build --strict passes (removing mkdocstrings also cleared the pre-existing griffe warnings that used to block strict).
  • Crawlable-surface accounting: hub sitemap 57 → 27; the ~30 API pages moved to the PW site (6 → 40). Content relocated, not lost.
  • Homepage em-dash clean; SEO/docs + the migrated-package removal only.

Dependency chain

Stacked on hub content #654. Depends on PW #17 (release-readiness) + #18 (API docs) + the staged draft release being published first. See the release runbook on the PW draft release.

Part of resolving the hub/framework duplication (ADO Epic 506340). No reviewer set per current instruction.

Additive content for the cluster hub (homepage repositioning intentionally NOT
touched; see PR notes for the Biodiversity-vs-Pytorch-Wildlife positioning flag):

- New docs/ecosystem.md: the ecosystem directory + a "which project should I
  use?" decision guide, with topic-specific anchors linking out to each
  project's documentation site (cluster cohesion).
- New docs/ecosystem-standards.md: the shared documentation standards every
  ecosystem repo follows (structure, metadata, structured data, cross-linking,
  topic ownership) so new repos replicate the pattern.
- Add both to nav; surface the ecosystem + collaborators from the homepage.
- Add title front-matter to 10 hand-written pages; remove stale "SPARROW Studio"
  naming and em-dashes from touched pages.

Part of the cluster SEO-parity work (ADO Epic 506340).
…rella

Completes the repo split by removing from the Biodiversity hub the framework code
that now lives in dedicated repos, and reframes the hub homepage as the ecosystem
umbrella. DRAFT: do not merge until the PytorchWildlife PyPI release is cut from
microsoft/Pytorch-Wildlife (so pip install resolves there).

Removed (verified migrated + parity-checked):
- PytorchWildlife/ core package + setup.py + version.txt + MANIFEST.in
  -> microsoft/Pytorch-Wildlife (complete superset after that repo's release-readiness PR).
- PW_Bioacoustics/ -> microsoft/MegaDetector-Acoustic (byte-for-byte identical, no hub drift).
- docs/base/ + docs/fine_tuning_modules/ (mkdocstrings API reference) -> now hosted on the
  Pytorch-Wildlife docs site; removed the mkdocstrings plugin + deps + the Reference nav section.

Reframed:
- docs/index.md is now the Microsoft Biodiversity umbrella homepage (was 'Welcome to
  PyTorch-Wildlife'); it points the framework to microsoft/Pytorch-Wildlife and leads with
  the ecosystem. Top nav group renamed to Microsoft Biodiversity.

Left in place and FLAGGED for a team decision (NOT safe to auto-delete):
- PW_FT_classification/ has diverged from microsoft/MegaDetector-Classifier and carries
  hub-only fixes (security #628, GPU-crash #629); reconcile before removing.
- PW_FT_detection/ has no dedicated repo yet; needs a destination.
- demo/, requirements.txt, Dockerfile, Brewfile, archive/ left for review.

Verified: mkdocs build --strict passes; hub sitemap 57 -> 27 (the ~30 API pages moved to the
PW site, which grew 6 -> 40); homepage em-dash clean.
… Acoustic

Per SEO review: a hard removal of the ~32 hub API pages would 404 and leak link
signals. Replace the deletion outcome with mkdocs-redirects: each old hub API URL
(base/*, fine_tuning_modules/*) now serves a 0-second meta-refresh to its
Pytorch-Wildlife equivalent (Google treats instant meta-refresh as a permanent
redirect on static hosts). Redirect stubs are auto-excluded from the sitemap;
slugs are preserved 1:1. Also omit the private MegaDetector-Acoustic link from the
umbrella homepage until its Pages URL is stable.
@rain-Brian rain-Brian requested a review from zhmiao June 10, 2026 17:24
@rain-Brian rain-Brian changed the base branch from seo/hub-content to main June 25, 2026 22:08
@rain-Brian rain-Brian marked this pull request as ready for review June 25, 2026 23:43
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant