Fix review/rejoin treating a selected stale take as current

[cursor]

Resumen ejecutivo

Esta primera sección está pensada para project managers y revisores no técnicos. Mantén el detalle técnico completo más abajo.

Qué cambia

Si el director ya había elegido una toma y luego regenera el fotograma inicial, guardar una nota o pulsar Rejoin ya no trata ese vídeo viejo como actual.

Para qué sirve

Evita montar la película final con un plano que ya no corresponde a la imagen de partida.

Impacto para el usuario

Rejoin y el guardado de revisión siguen respetando la toma elegida para reproducirla, pero piden regenerar el vídeo cuando esa toma quedó obsoleta.

Riesgo

• Medio
• Bajo
• Alto

Estado

• Listo para revisión
• En desarrollo
• Bloqueado por CI o revisión
• Requiere migración o acción manual

Summary

_backfill_clip_video_attempts promoted selected_video_filename to video_filename and also set video_stale=False. That is correct for playback authority, but it erased an intentional stale mark after rerun_clip_image.

load_pipeline_state (GET, Rejoin) and _update_saved_pipeline (prompt edits; also the projection #406 now persists) then treated the old take as current. Rejoin assembled a movie whose start frame no longer matched the video.

The backfill still uses the selection for playback. It no longer clears video_stale. Explicit take changes (_select_take, select_clip_video_attempt, video rerun) still set video_stale=False.

Overview

1. User reruns a clip video (this writes selected_video_filename).
2. User reruns the start image → disk has video_stale=True while the selected take remains.
3. Opening the desk, saving notes, or clicking Rejoin loaded the projection and cleared stale.
4. Rejoin concatenated the outdated take against the new still.

After this change, step 3 keeps video_stale=True and Rejoin refuses until the video is regenerated.

Detailed changes

Backend

• app/services/director_pipeline.py: when a selected take exists, copy it to video_filename only.

UI and Wizard

Unchanged. The desk already keeps stale when re-persisting the current take (re-persisting the current take does not clear a stale clip).

Data, provenance and compatibility

No migration. Existing checkpoints that already lost video_stale stay as they are; new loads no longer drop a present flag.

Files and ownership

• app/services/director_pipeline.py — root-cause one-liner in the GET/persist projection.
• tests/test_director_cancellation.py — Rejoin + notes persist with selected+stale.
• tests/test_director_review.py — review notes keep stale and the existing approval tag.
• Left untouched: _select_take / video rerun still clear stale when the user picks or generates a new take.

Validation

• Date (UTC): 2026-09-12

• Base SHA: 60e867a

• Head SHA: acf3c51

• Validation scope: focused

• python scripts/verify_clean_repo.py — PASS

• python -m compileall -q app/services/director_pipeline.py app/services/director_review.py — PASS

• Focused Python tests: pytest -q tests/test_director_review.py (5 passed) and test_rejoin_rejects_stale_video_even_when_a_take_is_selected plus image-rerun / history-rejoin cases (all passed)

• cd ui && npm run i18n:check — N/A (no UI files)

• UI tests — N/A

• cd ui && npm run lint -- --max-warnings=0 — N/A

• cd ui && npm run build — N/A

• git diff --check — PASS

• E2E/smoke checks: N/A — no video generation; temp files and mocked concatenate only

Code quality

• Score: pending CI
• Complexity trend: pending CI
• Production LOC trend: pending CI
• Regression versus main: pending CI
• Ratchet: pending CI

CI and review

• CI of this HEAD: pending

• Independent agent review of this HEAD: pending

• Human merge click (operational, not code review): pending

• Reviewed at current HEAD

• Earlier review; HEAD has changed since (stale)

• Pending

• Unavailable

Coste de la tarea

• Tests simulados: 0 tokens externos
• Tests reales: N/A
• Llamadas LLM externas: 0
• Tokens de prompt: N/A
• Tokens de respuesta: N/A
• Tokens totales: N/A
• Generaciones de imágenes/audio/vídeo: 0
• Tiempo transcurrido: N/A
• Proveedores/modelos: N/A

Notes and limitations

Does not restore video_stale on checkpoints that already persisted false after a notes save on #406 or a prompt edit. Does not change the rule that choosing a different take clears stale.

Follow-up work

• #406 still awaits review (sidecar-recovered takes). This fix is on development so that PR inherits it when it rebases or when both merge.
• #405 (Approve after notes blur) and #408 (cut-paper mouth before first word) remain open; not this bug.

Checklist

• The executive summary is understandable without reading the code.
• The detailed Summary/Overview has not been removed or shortened.
• Tests and their actual results are recorded.
• Generated assets, secrets and local-only files are not committed.
• Required CI and Cursor/Bugbot review are complete, or the PR is clearly marked as waiting for them.

  

[github-actions]

PR Review — Loreframe Studio

Risk: low
Scope: 3 file(s); +65/-1; backend services

Automated review from scripts/analyze_pr.py. This is a heuristic pass (no LLM) so humans still own the merge decision.

Findings

• No heuristic issues. Still run the CI checklist below.

Changed files

• modified: app/services/director_pipeline.py, tests/test_director_cancellation.py, tests/test_director_review.py

CONTRIBUTING checklist

• python scripts/verify_clean_repo.py
• python -m compileall -q app/services app/launch.py scripts
• cd ui && npm run build if the UI changed
• No weights, CivitAI sidecars, or generated guides
• Stays local-first (no required accounts / telemetry)

Posted by the repo PR review workflow. Re-runs on each push to the PR.

[github-actions]

Code health

Quality score: 63.8/100

Higher is better. The score is a trend dashboard; the independent ratchet below remains the CI gate.

Component	Weight	Current	Change
Cyclomatic health	45%	59.2	+0.0
File concentration	25%	73.7	+0.0
Oversized-file debt	20%	49.4	+0.0
Modularity	10%	88.3	+0.0

Change vs PR base: +0.0 points.

Metric	Value
Production LOC	313,652
Production files	1,052
Test LOC	127,872
Functions measured	21,367
Functions complexity ≥ 15	906
Maximum complexity	674
Policy	code-health-policy-v1
HEAD	3cdfd27c6ad71538f02fedcd9e85cf918f77a7cb
Base	60e867a6aae832d5b56c5ebd44b99867bda55b26
UI measurement	complete

Markdown, JSON catalogs and tests are out of this table. Only app/ runtime + ui/src TS/JS count.

Most complex functions

Complexity	Where
674	app/wgp.py:7230 generate_video
382	ui/src/stores/useStore.ts:4129 Async method 'startGeneration'
355	app/_launch_runtime.py:23674 _run_generation
308	app/wgp.py:12372 generate_video_tab
269	ui/src/components/Sidebar/SceneAnimatorPanel.tsx:508 Function 'SceneAnimatorPanel'
264	ui/src/stores/useStore.ts:8684 Async method 'loadSettingsFromOutput'
258	app/services/director/planners/short_film.py:3433 ShortFilmPlanner._plan_story_driven
248	app/services/director_pipeline.py:13614 _run_video_generation
245	app/services/director_pipeline.py:7714 _run_pipeline
233	ui/src/features/agent/agentActions.ts:1152 Function 'parseAction'
226	app/services/director_pipeline.py:6543 update_comic_preview
225	ui/src/features/agent/agentActions.ts:2998 Async function 'executeAgentActions'

Trend vs baseline

Metric	Δ
Production LOC	+2
Test LOC	+62
Functions ≥ 15	+0
Maximum complexity	+0

Warnings

• production LOC increased by +2
• hotspot app/services/director_pipeline.py increased by +2 lines

Ratchet passed.