feat(characters): add manual 2D speech preparation workshop by IAnMove · Pull Request #200 · IAnMove/hocuspocus

IAnMove · 2026-09-06T22:04:24Z

Summary

Manual 2D speech preparation workshop in Character Creator. Saves reviewed drafts without starting inference.

Workshop entry, draft recovery without rebasing another editor’s revision, and library save with workspace revision.
Face Rig child stays; this path does not POST generate/analyze/cleanup.
The panel test avoids HTMLElement-vs-null formatting that exhausted RAM (OOM 6 Sep). 13 directed tests, ~260 MiB peak RSS.

Does not merge the 3D scene PR or reopen Labs L0–L12.

Test plan

ui/tests/characterSpeechPreparationPanel.test.tsx (RAM-safe assertions)
draft recovery tests
required CI

Independent QA / Bugbot are not required.

Note

Medium Risk
Touches character-kit persistence and optimistic concurrency in the UI; mistakes could lose drafts or mishandle revision conflicts, though behavior is heavily tested and scene Face Rig defaults are unchanged.

Overview
Adds a Prepare 2D speech flow under Character Creator so users can pick or import a character base, review poses and mouth pieces via the existing Face Rig UI, and save explicitly to the workspace character-kit library—without running 3D orbit or inference.

The workshop is a lazy-loaded drawer (CharacterSpeechWorkshopEntry → CharacterSpeechPreparation) backed by useCharacterSpeechLibrary for load/save/upload, optimistic baseRevision on PATCH, and session draft recovery (characterSpeechDraft: validated, 2 MiB cap, per-workspace). Recovered drafts keep their original revision after a remote bump; save failures retain the draft with no auto-retry. CharacterKitFaceRigPanel gains allowModelActions={false} here so generate/cleanup/voice paths are disabled while manual placement, presets, and review stay available.

A mouth readiness checklist (speechPreparationReadiness) surfaces missing/pending/rejected/incompatible/approved per viseme. EN/ES copy and creator/library subtitles point users to the new entry. Docs add CHARACTER_SPEECH_WORKSHOP.md and refresh raster-plan CI/merge notes.

Coverage includes unit/panel tests (workspace races, draft recovery, save conflicts), a Playwright E2E that asserts no inference POSTs, and a panel-test fix that avoids DOM-heavy assertions that previously triggered OOM.

^{Reviewed by Cursor Bugbot for commit 7fa0c6b. Configure here.}

…-speech-workshop

cursor · 2026-09-06T22:04:34Z

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_d904fdb6-1c84-41ac-957d-2db15e2e954b)

github-actions · 2026-09-06T22:04:40Z

PR Review — Loreframe Studio

Risk: medium
Scope: 16 file(s); +1613/-16; React UI, docs

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

Findings

medium — Large pull request
1613 additions / 16 deletions. Reviewers will have an easier time with smaller, focused PRs.
medium — Very large file change (ui/tests/characterSpeechPreparationPanel.test.tsx)
ui/tests/characterSpeechPreparationPanel.test.tsx adds 430 lines. Consider splitting the PR.
low — UI changed — rebuild before merge
Run cd ui && npm run build (CI already does this). Pinokio Update rebuilds for end users; keep ui/dist untracked.

Changed files

added: docs/development/CHARACTER_SPEECH_WORKSHOP.md, ui/e2e/specs/character-speech-workshop.spec.ts, ui/src/features/characters/CharacterSpeechPreparation.tsx, ui/src/features/characters/CharacterSpeechWorkshopEntry.tsx, ui/src/features/characters/useCharacterSpeechLibrary.ts, ui/src/lib/characterSpeechDraft.ts, ui/src/lib/characterSpeechPreparation.ts, ui/tests/characterSpeechDraft.test.mjs, ui/tests/characterSpeechPreparation.test.mjs, ui/tests/characterSpeechPreparationPanel.test.tsx
modified: docs/development/CHARACTER_SPEECH_RASTER_PLAN.md, ui/src/features/characters/CharacterCreatorPanel.tsx, ui/src/features/characters/CharacterKitFaceRigPanel.tsx, ui/src/i18n/locales/en/characters.json, ui/src/i18n/locales/es/characters.json, ui/tests/characterCreator.test.tsx

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.

IAnMove · 2026-09-06T22:05:03Z

Qué cambia

«Prepare 2D speech / Preparar habla 2D» dentro de Character Creator, lazy y
plegable sin perder el borrador al cerrar el desplegable.
Biblioteca existente por workspace; base importada con identidad nueva y
revisión pendiente, selección de pose, Face Rig y parches raster reutilizados.
Estado estructural de cada boca: falta/pendiente/rechazada/incompatible/aprobada.
No certificar calidad artística ni fonética por tener piezas aprobadas.
Guardado explícito con revisión optimista; errores conservan borrador y no
provocan reintentos/sobrescrituras. Respuestas tardías no cruzan workspaces.
Recuperación de borrador al navegar/cambiar workspace/remontar el panel,
conservando baseRevision original. Caché de sesión versionada/validada, 2 MiB,
con fallback en memoria; guardar o descartar limpia la recuperación.
En este taller, generación de imagen/voz y cleanup neural deshabilitados
tanto en botones como handlers. El editor de escenas conserva su modo previo.

Alcance y límites

Primera entrega manual. Rhubarb, MuseTalk, FLUX/Qwen con máscara y packs por vista
están descritos como pendientes, no instalados ni integrados. Los sprites propios
se asignan mediante las capas de escena existentes; el taller importa variantes
alineadas para parches. No se garantiza alineación automática ni boca expresiva.
Sin cambios de runtime/launchers, modelos, pesos, outputs o medios privados.
No reiniciar instancias compartidas ni publicar main.

Validación

979 tests UI aprobados, incluidos 22 específicos del taller y recuperación.
Python: 2160 aprobados, 1 omitido; CPU/BLAS limitados, sin GPU/proveedores.
Lint, tipos, i18n y guards clean/deps/documentación/marca aprobados.
Ratchet contra development c2794f6c, build/budget y 19 E2E aprobados.
Navegación real → revisión → salir/remontar → recuperar → guardar → recargar,
con API simulada y sin inferencia. Entrada gzip 316154/327680 bytes.
Integración local de esa base: 7 tests del contrato Labs entrante aprobados.
No hay cambios Python de este PR; suite Python completa anterior conservada.
Score 53.0/100: indicador, no certificado; ratchet pasa con avisos de
complejidad (tres casos >=15 y Face Rig 84→85). Sin modificar baseline.
Primera tanda E2E: 18/19; el nuevo test cambiaba de sección antes de concluir
la navegación lazy. Se comprueba ahora desmontaje real, sin sleeps/retries.
Repetición completa: 19/19. Revisión independiente del ajuste completada.
Luna revisó la integración y detectó pérdida de borrador al navegar; corregida
y re-revisada sin nuevos bloqueos en 6713ce7c. El helper delegado fue revisado
por el principal (incluido reemplazo de caché al agotarse cuota), no se cuenta
como revisión independiente por su propio autor.
Incidente local documentado: un assert comparaba HTMLElement con null y su
formateador de errores provocó OOM. Corregido esperando la recarga y comparando
primitivas. Repetición específica aprobada bajo cgroup 1,5 GiB/swap 0; suites
posteriores limitadas a 3 GiB, una tanda a la vez, no GPU ni proveedores.

Validación de código local en b663912d; el commit posterior solo registra
evidencia documental. CI remoto, Cursor y merge se comprobarán por separado.

CI verde no es validación artística con generación real. La aprobación del
usuario de la identidad MiniMax se conserva; el pack hablando queda pendiente.

github-actions · 2026-09-06T22:08:42Z

Code health

Quality score: 53.0/100

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

Component	Weight	Current	Change
Cyclomatic health	45%	53.7	+0.0
File concentration	25%	59.5	+0.1
Oversized-file debt	20%	34.9	+0.2
Modularity	10%	69.6	+0.5

Change vs PR base: +0.1 points.

Metric	Value
Production LOC	257,712
Production files	628
Test LOC	88,730
Functions measured	16,410
Functions complexity ≥ 15	818
Maximum complexity	670
Policy	code-health-policy-v1
HEAD	`4067954dc78e78ffe594232ccd11d0653450f15c`
Base	`c2794f6ce80d4148936fb474e4566eec5e82e8d1`
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
670	`app/wgp.py:7164` `generate_video`
376	`ui/src/stores/useStore.ts:4029` `Async method 'startGeneration'`
355	`app/_launch_runtime.py:23505` `_run_generation`
308	`app/wgp.py:12298` `generate_video_tab`
275	`ui/src/components/Sidebar/SceneAnimatorPanel.tsx:475` `Function 'SceneAnimatorPanel'`
271	`ui/src/stores/useStore.ts:8572` `Async method 'loadSettingsFromOutput'`
258	`app/services/director/planners/short_film.py:3433` `ShortFilmPlanner._plan_story_driven`
248	`app/services/director_pipeline.py:13607` `_run_video_generation`
245	`app/services/director_pipeline.py:7712` `_run_pipeline`
244	`ui/src/features/agent/agentActions.ts:1145` `Function 'parseAction'`
226	`app/services/director_pipeline.py:6541` `update_comic_preview`
225	`ui/src/features/agent/agentActions.ts:2892` `Async function 'executeAgentActions'`

Trend vs baseline

Metric	Δ
Production LOC	+509
Test LOC	+832
Functions ≥ 15	+3
Maximum complexity	+0

Warnings

production LOC increased by +509
functions at complexity >= 15 increased by +3
complexity hotspot ui/src/features/characters/CharacterKitFaceRigPanel.tsx rose 84 -> 85

Ratchet passed.

IAnMove · 2026-09-06T22:12:22Z

Cierre verificado: MERGED en development, a0bf91c; padres c2794f6 + 7fa0c6b y árbol idéntico al HEAD validado. CI required y jobs de producto PASS. Cursor no ejecutado por cuota (dispensa previa del usuario); Independent QA neutral, no confundir con publisher PASS. Revisión local cruzada de agentes documentada arriba. Taller manual integrado; sincronización con audio real y prueba artística pendientes. No se han reiniciado instancias compartidas ni publicado main; tampoco se han subido modelos o medios.

CURRENT_WORK now treats the manual 2D speech workshop as merged on a0bf91c. The remaining limit is artistic validation, not the RAM-safe panel tests already in that PR.

IAnMove added 5 commits September 6, 2026 23:35

feat(characters): add manual 2d speech preparation workshop

4cfc6cd

fix(characters): recover speech drafts without bypassing revision checks

6713ce7

test(characters): await navigation before checking speech draft recovery

27d68e7

Merge remote-tracking branch 'origin/development' into feat/character…

b663912

…-speech-workshop

docs(characters): record bounded speech workshop validation

7fa0c6b

IAnMove added a commit that referenced this pull request Sep 6, 2026

docs: note speech workshop PR #200 in current work

b97ab6a

IAnMove merged commit a0bf91c into development Sep 6, 2026
10 checks passed

IAnMove added a commit that referenced this pull request Sep 6, 2026

docs: restack current-work archive onto development after #200

3c1b77d

IAnMove mentioned this pull request Sep 6, 2026

fix(labs): distinguish Series shot numbers from a concrete attemptId #201

Merged

Navigation Menu

feat(characters): add manual 2D speech preparation workshop - #200

feat(characters): add manual 2D speech preparation workshop#200
IAnMove merged 5 commits into
developmentfrom
feat/character-speech-workshop

IAnMove commented Sep 6, 2026 •

edited by cursor Bot

Loading

Uh oh!

cursor Bot commented Sep 6, 2026

Uh oh!

github-actions Bot commented Sep 6, 2026

Uh oh!

IAnMove commented Sep 6, 2026

Uh oh!

github-actions Bot commented Sep 6, 2026

Uh oh!

Uh oh!

IAnMove commented Sep 6, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

IAnMove commented Sep 6, 2026 • edited by cursor Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Test plan

Uh oh!

cursor Bot commented Sep 6, 2026

Bugbot couldn't run - usage limit reached

Uh oh!

github-actions Bot commented Sep 6, 2026

PR Review — Loreframe Studio

Findings

Changed files

CONTRIBUTING checklist

Uh oh!

IAnMove commented Sep 6, 2026

Qué cambia

Alcance y límites

Validación

Uh oh!

github-actions Bot commented Sep 6, 2026

Code health

Quality score: 53.0/100

Most complex functions

Trend vs baseline

Warnings

Uh oh!

Uh oh!

IAnMove commented Sep 6, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

IAnMove commented Sep 6, 2026 •

edited by cursor Bot

Loading