Files
hyperframes/skills/media-use/scripts/lib/registry.mjs
T
Miguel ÁngelandClaude Opus 4.8 57b3c78987 feat(media-use): color grading — grade/lut resolve, smart-grade, grade-compare + compare (#2041)
* feat(media-use): color grading — grade/lut resolve, smart-grade, grade-compare CLI

Add color grading to media-use as first-class resolve types plus a faithful
comparison command. All local, offline, deterministic — no model, no GPU.

- resolve -t grade / -t lut: produce a data-color-grading block (or a frozen
  .cube). Look cascade: core preset (no file) -> bundled .cube library ->
  parametric buildCube. Emitted .cube is Rec.709 and validated against core's
  colorLuts constraints (LUT_3D_SIZE <= 64) before it is frozen.
- smart grade (grade --for <media>): ffmpeg signalstats -> adjust suggestion
  (exposure / contrast / white balance), surfaced with the measured evidence on
  stderr as a starting point; never auto-applied.
- hyperframes grade-compare: renders N candidate grades onto a reference frame
  through the real runtime shader into one labeled comparison PNG, so an agent
  picks a look without opening Studio. Prepends an "original" baseline cell by
  default (--no-baseline to omit). Shares the headless-capture pipeline with
  snapshot via capture/captureCompositionFrame.
- media-use SKILL: proactive "media opportunity pass" guidance (grounded
  signal -> offer, ask once, surface don't mutate).

Verified: media-use 116/116, grade-compare 7/7, snapshot 9/9, lint + format
clean, full build green, comparison renders end to end.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* test(cli): narrow grade-compare baseline assertion off unknown-typed grading

Assert the whole cell via toEqual instead of reaching into .grading.preset /
.grading.lut on the unknown-typed field, keeping the test typecheck-clean.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* feat(media-use): agent-authored LUTs via --params + validate --from cube; never-read-.cube guardrail

- resolve -t lut / -t grade --params '<json>': build a parametric .cube from
  explicit params (bypassing the intent cascade), validate, and freeze in one
  step. --intent becomes the optional description. Lets an agent commit a look
  it computed itself.
- --from <file.cube> now validates the ingested LUT for lut/grade types and
  rejects an invalid/oversized cube (no partial write) — the escape hatch for a
  LUT the agent generated with its own code.
- SKILL.md: hard rule to never read a .cube body into context (~size^3 lines,
  zero legible signal) — inspect via grade-compare (see it) or cube-validate
  (ok/size), read the manifest description for meaning; plus both authoring
  paths and the parametric-vs-film-stock ceiling note.

Verified: media-use 116/116, lint + format clean; smokes — --params builds a
valid frozen cube, grade --params returns a lut block, bad JSON and an oversized
--from cube are both rejected with no stray file.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* fix(cli): grade-compare validates referenced LUTs, warns on no-op cells, caps candidates

Bug-bash follow-ups — grade-compare silently accepted bad input:

- Validate LUT *content*, not just existence: each referenced .cube is parsed
  with core's parseCubeLut (now exported from @hyperframes/core) and rejected
  with a per-cell error ("LUT for \"<label>\" is not a valid .cube: ..."). A
  file that exists but isn't a valid cube no longer renders a silent no-op cell.
- Warn on inactive cells: a grading that normalizes to inactive (e.g. a
  malformed {lut:12345}) emits a stderr warning naming the cell; the
  auto-prepended "original" baseline is intentionally inactive and stays silent.
  stdout remains valid JSON.
- Cap candidates at 16 (excluding baseline): over-cap input renders the first N
  and reports {truncated:true, total:M} on stdout + a stderr note — no silent
  drop, no unbounded giant sheet.

Verified: grade-compare 10/10; non-cube LUT → clear error; {lut:12345} → warning
+ ok; 20 cells → cells=17 truncated total=20; valid runs unchanged. Lint/format
clean.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* feat(cli): general `hyperframes compare` visual-variant primitive

Generalize grade-compare's "render N variants → one labeled sheet → the agent
looks and picks" loop into a standalone command that works on ANY variation
(font, layout, motion, grade, whole compositions) — the tool never needs to
know what differs.

- `hyperframes compare <path...> [--at <sec>] [--labels a,b,c] [--out] [--cols]
  [--json]`: renders each agent-authored composition variant through the real
  runtime (captureCompositionFrame) and stitches one labeled comparison sheet +
  JSON ({ok, sheet, rendered, variants, truncated?/total?}). 2+ paths required;
  caps at 16 with loud truncation. It presents, it does not judge — choosing is
  the caller's job.
- Factored the shared "render a labeled set → contact sheet" path so compare,
  grade-compare, and snapshot all sit on it (no duplication). grade-compare is
  now the first color-specific specialization of this primitive.
- New pathArgs util + contactSheet test; hyperframes-cli SKILL documents compare
  as the agent's "see your own renders and choose" primitive.

Verified: 26/26 across compare + grade-compare + snapshot + contactSheet (no
regressions); compare renders 3 variants into one visibly-distinct labeled
sheet; 2+-path error path clean; lint/format clean.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* fix(ci): green the skills CI — skip ffmpeg tests when absent, oxfmt markdown

The "Test: skills" CI job runs bare `node --test` with no ffmpeg on PATH (by
design — skills tests are meant to be node-builtin-only). The grade-analyzer +
smart-grade tests shell to ffmpeg and were failing there with ENOENT. Guard
them to skip when ffmpeg isn't on PATH; they still run locally / where it is.

Also oxfmt README.md + hyperframes/media-use SKILL.md (the whole-repo
`oxfmt --check .` Format job caught markdown left unformatted by the rebase
conflict resolution).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* fix(ci): skip core-conformance test when tsx is unavailable

The "Test: skills" CI job installs no deps, so the normalizeHfColorGrading
conformance test (which imports core's TS via `node --import tsx`) failed there.
Guard it to skip when tsx can't resolve; runs locally / in the deps-installed
Test job. Completes the skills-CI greening (the ffmpeg guards handled the rest).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* fix(cli): escape grade-compare src double-quotes (CodeQL XSS) + Windows-safe compare test

- grade-compare built `<img src="...">` (double-quoted) with the single-quote
  escaper, leaving `"` unescaped — a `"` in the frame path could break out
  (CodeQL: incomplete HTML attribute sanitization). Use escapeXml for src.
- compare label test hard-coded POSIX paths that can't match on Windows; assert
  the derived labels (the subject); path resolution is covered elsewhere.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* refactor(media-use): generate LUT library from params (drop committed .cube files)

The 3 bundled .cube files were 733 lines each (2,199 total) and were themselves
buildCube output — pure repo bloat. Replace with compact per-look params in
luts/index.json, generated on resolve; add an optional `url` for future scanned
LUTs to be CDN-hosted + downloaded on demand (freezeUrl) instead of committed.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* feat(media-use): serve library LUTs from CDN on-demand (static.heygen.ai/luts), params fallback

Looks now carry a CDN `url` (hosted at s3://heygen-public/luts → static.heygen.ai/luts/<id>.cube);
resolve downloads + validates + freezes on demand, like bgm/image. `params` stays
as the deterministic offline fallback (--local-only, or if the download fails), so
resolution is never blocked on the network. Provider prefers url, falls back to params.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

* fix(media-use): address #2041 review — atomic LUT writes, compare telemetry, follow-ups

- Atomic .cube writes: library provider (url + params) and the parametric
  generator now write to a .tmp path, validate, then rename, so a crash can
  never orphan an invalid .cube at the final path (was validate-after-write).
- track("media_use_resolve") now emits provenance.via (url/params-fallback/params).
- grade-compare + compare: --timeout flag (was hardcoded 5000) and a
  media_use_compare event (cells, truncated, total, render_ready_timed_out);
  openSettledCompositionPage now surfaces the render-ready timeout.
- compare staging skips node_modules/.git; --for gets an upfront existence check.
- Rec.709 luma comment; HYPERFRAMES_ANALYZE_TIMEOUT_MS override; measured note
  uses basename; LUT s3 hosting moved from index.json into luts/README.md.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H5k87mPZ4d6yiFwcWSb8Vv

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-08 22:20:16 -04:00

163 lines
7.1 KiB
JavaScript

// Provider registry — the v2 contract.
//
// Each media type maps to an ORDERED list of provider entries. Providers are
// tried in order; the first to return a non-null result wins, which keeps
// resolution deterministic (same request -> same provider -> same file ->
// reproducible renders). heygen-CLI is always first for the types it serves.
//
// An entry exposes any of three capability methods — search / generate /
// process — plus { name }. media-use holds no keys; each external tool owns its
// own auth. Providers, by type:
// - heygen CLI: catalog + TTS, first for every type it serves (sub creds)
// - mflux: local FLUX-class image gen, spec-selected to the machine's RAM
// (free, private, offline once cached)
// - codex CLI: image gen on the user's ChatGPT sub — the better-quality upsell
// and the fallback when no local model fits
// - Kokoro (via the hyperframes CLI): local voiceover, free/private default
// for the voice type, ahead of the paid HeyGen TTS upsell
//
// Generation is local-first, cloud-upsell. `ctx.provider` forces one provider
// (e.g. "make an image with codex").
import { bgmProvider } from "./bgm-provider.mjs";
import { sfxProvider } from "./sfx-provider.mjs";
import { imageProvider, iconProvider } from "./image-provider.mjs";
import { brandProvider } from "./brand-provider.mjs";
import {
svglSearch,
simpleIconsSearch,
githubAvatarSearch,
faviconSearch,
} from "./logo-provider.mjs";
import { heygenTtsGenerate } from "./voice-provider.mjs";
import { localTtsGenerate } from "./tts-local-provider.mjs";
import { codexImageGenerate } from "./codex-provider.mjs";
import { mfluxImageGenerate } from "./mflux-provider.mjs";
// Provider markers: `network` = hits a remote service (skipped by --local-only).
// `paid` = costs wallet credits (documentation for the agent's cost judgment,
// X4: agent-initiated paid should confirm). HeyGen catalog SEARCH is free;
// HeyGen TTS now costs credits, so it is the paid upsell behind local Kokoro.
const A = (name, caps) => ({ name, ...caps }); // local, free
const N = (name, caps) => ({ name, network: true, ...caps }); // remote, free
const P = (name, caps) => ({ name, network: true, paid: true, ...caps }); // remote, paid
// heygen-CLI first (and currently only). All remote providers are skipped by --local-only.
const REGISTRY = {
bgm: [N("heygen.audio.sounds", { search: bgmProvider.search })],
sfx: [N("heygen.audio.sounds", { search: sfxProvider.search })],
image: [
N("heygen.asset.search", { search: imageProvider.search }),
// Catalog miss -> generate. Local first (best FLUX-class model the machine's
// RAM can run, spec-selected; free, private, kept under --local-only), then
// the codex CLI on the user's ChatGPT sub as the better-quality upsell and
// the fallback when no local model fits.
A("mflux.local", { generate: mfluxImageGenerate }),
N("codex.image_gen", { generate: codexImageGenerate }),
],
icon: [N("heygen.asset.search", { search: iconProvider.search })],
logo: [
// Official brand marks. Tiers verified by a 54-brand stress test (100%
// cascade hit); HeyGen asset search is deliberately absent — it returns
// generic look-alike icons for brand queries. All free, all network →
// --local-only leaves only the cache rungs.
N("svgl", { search: svglSearch }),
N("simple-icons", { search: simpleIconsSearch }),
N("github.avatar", { search: githubAvatarSearch }),
N("favicon.ddg", { search: faviconSearch }),
],
voice: [
// Local Kokoro first (free, private, on-device via the hyperframes CLI, kept
// under --local-only), then HeyGen TTS as the higher-quality paid upsell and
// the fallback when Kokoro is not set up.
A("kokoro.local", { generate: localTtsGenerate }),
P("heygen.tts", { generate: heygenTtsGenerate }),
],
brand: [
// Local design spec, not heygen — reads frame.md / design.md tokens.
A("design_spec", { search: brandProvider.search }),
],
grade: [
// Local deterministic cascade handled by resolve.mjs so grade records can
// carry an inline block as well as an optional frozen .cube file.
A("color_grade.local", { search: async () => null, generate: async () => null }),
],
lut: [
// Lower-level local LUT generation/freezing path handled by resolve.mjs.
A("cube_lut.local", { search: async () => null, generate: async () => null }),
],
};
function listFor(type) {
const list = REGISTRY[type];
if (!list) throw new Error(`unknown media type: ${type}`);
return list;
}
/** Ordered providers for a type. */
export function getProviders(type) {
return listFor(type);
}
/** All declared media types. */
export function listTypes() {
return Object.keys(REGISTRY);
}
/** Provider names available for a type, in cascade order (for --provider validation). */
export function providerNamesFor(type) {
return listFor(type).map((p) => p.name);
}
/**
* Does an override token (full name like "codex.image_gen" or a prefix like
* "codex") match any provider declared for the type? Same match rule as
* runProviders, so validation and dispatch never disagree.
*/
export function providerMatches(type, want) {
return providerNamesFor(type).some((n) => n === want || n.startsWith(`${want}.`));
}
/**
* Back-compat shim for the v1 single-provider API. Returns the first declared
* provider for the type (tagged with `type`); throws for an unknown type.
* Kept for v1 callers only — new code should use getProviders/runCapability.
*/
export function getProvider(type) {
const first = listFor(type)[0] || {};
return { ...first, type };
}
/**
* Run a capability across an explicit ordered provider list. Tries each in
* order, returns the first non-null result, skips providers that don't expose
* the capability. Pure over its input — the unit-testable core of the cascade.
*
* Offline guard: a `network` provider is skipped when `ctx.localOnly` is set —
* unconditionally, even under a `ctx.provider` override. --local-only is a hard
* safety flag: it must never make a network call. Forcing a network provider
* while offline yields a clean miss (the caller explains the conflict), never a
* silent network request.
* Provider override: `ctx.provider` (a full name like "codex.image_gen" or a
* prefix like "codex") pins resolution to matching providers only — this is how
* a user "make an image WITH codex" forces the upsell instead of taking the
* free-first default.
*/
export async function runProviders(providers, capability, intent, ctx) {
const want = ctx?.provider;
for (const p of providers) {
if (want && p.name !== want && !p.name.startsWith(`${want}.`)) continue;
if (p.network && ctx?.localOnly) continue; // --local-only wins, even over --provider
const fn = p[capability];
if (typeof fn !== "function") continue;
const res = await fn(intent, ctx);
if (res) return res;
}
return null;
}
/** Run a capability over the providers for a type (deterministic, heygen-first). */
export async function runCapability(type, capability, intent, ctx) {
return runProviders(getProviders(type), capability, intent, ctx);
}