[paynejd]

TBv3 keeps pins as overview highlights and lists follows as personal shortcuts. TBv2's My Pins list (GET /users/<me>/pins/?includeCreatorPins=true) showed a person's own pins plus the org pins they created, so this one-time data migration turns each person's list into follows.

What changed

• New data migration core/pins/migrations/0005_pins_to_follows.py. For every pin it follows the pinned org, source or collection for the person whose My Pins list showed it: the pin's user for a user pin, the pin's creator for an org pin.
• Skips a pin when the resource is gone or deleted (is_active=False, which the follow API refuses too), or when the person can no longer view it. Viewing follows CanViewConceptDictionary: public, staff, the repo's owner, or a member of the org that owns it. A repo version counts as public only while its HEAD is. Org pins also need the person to still see the pinning org, as Pin.get_visible_to required for that list.
• Follows are inserted with bulk_create(..., ignore_conflicts=True), not UserProfile.follow(), so no Followed events are recorded and an existing follow stays as it is. Pins are untouched. The reverse is a no-op, since by then people may have changed their follows.
• Dependencies: pins.0004, users.0037, orgs.0022, sources.0046, collections.0071, contenttypes.0002. Historical models only.

How I verified it

• PinsToFollowsMigrationTest runs the migration function with historical models (MigrationLoader(...).project_state(...)) on a dataset covering: user pins, org pins (follower = creator), an existing follow (kept, no duplicate), a deleted source, an inactive collection, private repos and a private org the follower lost access to (skipped) or can still see (followed), a staff creator, a public version of a private HEAD, an org pin whose private org the creator can't see, and a second run (adds nothing). It checks pins and the Event count are unchanged, and that each person's follows equal their includeCreatorPins=true list less the deleted resources. 8 follows (7 new) on that dataset.
• Mutation check: dropping the creator rule, the pinning-org check, the HEAD check, the is_active filter or the owner check each fails the test.
• python manage.py test core.pins core.users --keepdb (107 tests, OK); pylint -j2 core 10.00; migrate, migrate pins 0004 and migrate pins run cleanly; makemigrations --dry-run shows only the known local MapProject.encoder_model drift, not included.

The web side (oclweb3) is a separate PR.

Part of OpenConceptLab/ocl_issues#2947

🤖 Generated with Claude Code

https://claude.ai/code/session_01TjM9dGow2KV1Gwr2dWQdph

[paynejd]

Codex review, pass 1 (codex-cli 0.163.0, commit 833ac23)

Verdict: clean, no findings.

findings.json

{
  "reviewer": "codex-cli 0.163.0",
  "perspective": "defects",
  "mode": "full",
  "commit": "833ac23d",
  "verdict": "clean",
  "prior": [],
  "findings": []
}

[paynejd]

Simplification review (claude subagent, commit 833ac23)

Verdict: clean, nothing to cut. No commits since Codex pass 1, so no follow-up Codex check is needed.

simplify.json

{
  "reviewer": "claude subagent",
  "perspective": "simplify",
  "mode": "full",
  "commit": "833ac23d",
  "verdict": "clean",
  "prior": [],
  "findings": []
}