Files
hyperframes/skills/media-use/scripts/lib/manifest.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

192 lines
5.7 KiB
JavaScript

import {
readFileSync,
appendFileSync,
mkdirSync,
existsSync,
readdirSync,
openSync,
closeSync,
writeFileSync,
rmSync,
statSync,
} from "node:fs";
import { join } from "node:path";
const MANIFEST_FILE = "manifest.jsonl";
const INDEX_FILE = "index.md";
const TYPE_DIRS = {
bgm: "audio/bgm",
sfx: "audio/sfx",
voice: "audio/voice",
image: "images",
icon: "images",
logo: "images",
brand: "images",
video: "video",
grade: "luts",
lut: "luts",
};
export function mediaDir(projectDir) {
return join(projectDir, ".media");
}
export function manifestPath(projectDir) {
return join(mediaDir(projectDir), MANIFEST_FILE);
}
export function indexPath(projectDir) {
return join(mediaDir(projectDir), INDEX_FILE);
}
export function typeSubdir(type) {
const sub = TYPE_DIRS[type];
if (!sub) throw new Error(`unknown media type: ${type}`);
return sub;
}
export function typeDirPath(projectDir, type) {
return join(mediaDir(projectDir), typeSubdir(type));
}
export function readManifest(projectDir) {
const p = manifestPath(projectDir);
if (!existsSync(p)) return [];
const raw = readFileSync(p, "utf8");
const records = [];
for (const line of raw.split(/\r?\n/)) {
const trimmed = line.trim();
if (!trimmed) continue;
try {
records.push(JSON.parse(trimmed));
} catch {
// ponytail: skip malformed lines, don't crash
}
}
return records;
}
export function appendRecord(projectDir, record) {
const dir = mediaDir(projectDir);
mkdirSync(dir, { recursive: true });
const typeDir = typeDirPath(projectDir, record.type);
mkdirSync(typeDir, { recursive: true });
const p = manifestPath(projectDir);
const line = JSON.stringify(record) + "\n";
appendFileSync(p, line);
}
// Match prompts forgivingly. Agents rarely re-emit a byte-identical intent, so
// keying cache lookups on exact equality meant "Calm piano" and "calm piano"
// re-searched and re-downloaded. Normalize (trim, lowercase, collapse internal
// whitespace) on both sides; the raw prompt is still stored for audit.
export function normalizePrompt(prompt) {
return String(prompt ?? "")
.trim()
.toLowerCase()
.replace(/\s+/g, " ");
}
export function findByPrompt(projectDir, prompt, type) {
const key = normalizePrompt(prompt);
if (!key) return null;
const records = readManifest(projectDir);
return (
records.find(
(r) => normalizePrompt(r.provenance?.prompt) === key && (type == null || r.type === type),
) || null
);
}
export function findByEntity(projectDir, entity) {
const lower = entity.toLowerCase();
const records = readManifest(projectDir);
return records.find((r) => r.entity && r.entity.toLowerCase() === lower) || null;
}
export function nextId(projectDir, type) {
const records = readManifest(projectDir);
const prefix = type;
let max = 0;
for (const r of records) {
if (r.type !== type) continue;
const m = r.id?.match(new RegExp(`^${prefix}_(\\d+)$`));
if (m) max = Math.max(max, parseInt(m[1], 10));
}
return `${prefix}_${String(max + 1).padStart(3, "0")}`;
}
// Sync sleep (no busy-spin) for the allocation lock retry.
function sleepMs(ms) {
Atomics.wait(new Int32Array(new SharedArrayBuffer(4)), 0, 0, ms);
}
// Coarse per-project lock so concurrent resolves don't race on id allocation.
// ponytail: one lock file with a 15s stale-steal (a crashed holder can't wedge
// the project); fine for agent-scale concurrency — revisit if throughput needs
// finer locking. Date.now() is available here (a normal Node CLI, not a
// workflow DSL), so mtime-based staleness is safe.
const LOCK_STALE_MS = 15000;
const LOCK_TIMEOUT_MS = 20000;
function withLock(dir, fn) {
const lock = join(dir, ".lock");
const start = Date.now();
for (;;) {
try {
closeSync(openSync(lock, "wx")); // O_EXCL: atomic acquire
break;
} catch (err) {
if (err.code !== "EEXIST") throw err;
try {
if (Date.now() - statSync(lock).mtimeMs > LOCK_STALE_MS) {
rmSync(lock, { force: true }); // steal a stale lock from a dead holder
continue;
}
} catch {
continue; // lock vanished between check and stat — retry the acquire
}
if (Date.now() - start > LOCK_TIMEOUT_MS) {
throw new Error("media-use: timed out acquiring .media/.lock");
}
sleepMs(25);
}
}
try {
return fn();
} finally {
rmSync(lock, { force: true });
}
}
// Atomically allocate the next free id for `type` AND reserve its file, so a
// slow download/copy between allocation and appendRecord can't let a concurrent
// caller grab the same id (the MU-23 clobber). Under the lock we take the max id
// across BOTH the manifest and any already-reserved files in the type dir, then
// O_EXCL-create an empty placeholder at the target path; freeze/copy overwrites
// it. Returns { id, localPath }.
export function allocateId(projectDir, type, ext) {
mkdirSync(mediaDir(projectDir), { recursive: true });
const typeDir = typeDirPath(projectDir, type);
mkdirSync(typeDir, { recursive: true });
return withLock(mediaDir(projectDir), () => {
const re = new RegExp(`^${type}_(\\d+)`);
let max = 0;
for (const r of readManifest(projectDir)) {
if (r.type !== type) continue;
const m = r.id?.match(re);
if (m) max = Math.max(max, parseInt(m[1], 10));
}
for (const f of readdirSync(typeDir)) {
const m = f.match(re);
if (m) max = Math.max(max, parseInt(m[1], 10)); // skip ids reserved but not yet appended
}
const id = `${type}_${String(max + 1).padStart(3, "0")}`;
const localPath = `.media/${typeSubdir(type)}/${id}${ext}`;
writeFileSync(join(projectDir, localPath), "", { flag: "wx" }); // durable reservation
return { id, localPath };
});
}