Restore the board-subfolder and Cluster Strength fixes #382 reverted

[lstein]

Indexing an InvokeAI board album skipped almost everything — "386 of 387 file(s) listed by InvokeAI were not found on disk and were skipped" against a backend at localhost:9090 with a correct InvokeAI root.

The resolution code had not been changed. It had been deleted.

What happened

PR #382 (the iPad fullscreen-panel fix) carried stale copies of files it never meant to touch, and squash-merging it rolled two already-merged PRs back to their exact pre-merge state:

Reverted PR	Files taken back
#378 — Resolve InvokeAI board media through their recorded subfolder	photomap/backend/invokeai_client.py, photomap/backend/routers/index.py, tests/backend/test_invokeai_client.py, tests/backend/test_invokeai_board_index.py
#376 — Pausing mid-edit no longer discards the Cluster Strength	photomap/frontend/static/javascript/umap.js, photomap/frontend/static/css/umap-floating-window.css, tests/backend/test_cluster_eps.py, tests/frontend/umap-eps-debounce.test.js (deleted outright), tests/frontend/umap-reindex-refresh.test.js

Every one of those files was byte-identical to its pre-merge content on master, tests included — which is why nothing failed. The tests that would have caught it went back with the code they covered.

The user-visible symptom is #378's: board albums went back to joining the bare filename to <invokeai_root>/outputs/{images,videos}, so on a backend whose subfolder strategy is not flat, almost nothing resolves. #376's loss is silent until someone pauses mid-edit over the Cluster Strength spinner.

What this does

Restores both commits verbatim — cherry-picked onto current master, no conflicts. #377's asyncio.to_thread hunk in index.py, which landed after the revert in an untouched region, is preserved.

No new logic: every line here has already been reviewed and merged once.

Verification

Live against the reporter's InvokeAI at localhost:9090, the same album that failed:

resolved: 386  missing: 1  video_api: True
videos: 296  images: 90
  /home/lstein/invokeai-main/outputs/images/user/d3d43dd3-….png
  /home/lstein/invokeai-main/outputs/videos/general/610600c3-….mp4

The inverse of the reported failure. The one remaining miss is a genuine gap rather than a resolution bug — InvokeAI lists a video whose subfolder resolves correctly and whose directory exists, but the file is not on disk.

Backend 844 passed, frontend 656 passed (633 before this PR, the difference being umap-eps-debounce.test.js coming back), ruff clean. Confirmed working in the app by the reporter.

Checked the seam between the restored #376 and #375, which was authored against the already-reverted tree: both floor the Cluster Strength at MIN_CLUSTER_EPS (0.01, the spinner's own min), so they agree.

Worth doing separately

Any branch opened before 2026-08-20 can carry the same staleness. This loop reports files a later commit took back to their pre-merge state:

for c in $(git log --format=%h -15); do
  for f in $(git show --name-only --format="" $c); do
    git cat-file -e $c^:"$f" 2>/dev/null &&
      git diff --quiet $c^ HEAD -- "$f" && echo "$c LOST: $f"
  done
done

(It misses files a commit created; a drop in the test count between PRs is the other tell.)

🤖 Generated with Claude Code

https://claude.ai/code/session_01KtGdMfK3k6z2tazjDjMFtE