screentinker/server/lib/image-ops.js
screentinker 13c9c67335
Drop sharp: pure-JS image ops on a worker thread (#263)
* spike: replace sharp with pure-JS image ops (jimp + jsquash WASM)

Removes the last native dependency from the ingest path, so the server no longer
needs a per-platform/per-ABI prebuilt to thumbnail an image. Motivated by getting
the server onto hardware with no toolchain, but the ABI tax is paid on every
install — it is the same failure class lib/preflight-deps.js exists to explain.

lib/image-ops.js is the whole surface: metadata() and writeThumbnail(), which are
the only two things ingest ever asked sharp for.

Format parity holds. jpeg/png/gif/tiff/bmp are native to Jimp; webp and avif go
through @jsquash WASM, whose bundled .wasm must be compiled by hand because the
packages locate it with fetch(file://) and Node has no file:// fetch — the only
symptom otherwise is a bare "fetch failed". heic is unsupported, as it already
was: sharp advertises heif but its prebuilt libvips refuses HEVC.

#170 is preserved by a different mechanism. Jimp applies EXIF orientation at
decode and rewrites the tag to 1, so metadata() reports display dimensions and
imageDisplayDims() runs as a no-op instead of swapping W/H a second time. The
helper stays in the path so the rule keeps living in one place.

Verified: 1643/1643 tests pass, and ingest was exercised in a child process with
node_modules/sharp moved aside — jpeg, EXIF-rotated jpeg, png, webp, avif, gif
all measured and thumbnailed correctly, corrupt input still yields nulls with no
phantom thumbnail_path.

KNOWN BLOCKER, do not ship as-is: Jimp is pure JS on the main thread, where sharp
handed work to a libvips threadpool. A 12MP photo goes 65ms -> 1079ms, and the
event loop stalls for 1003ms of it (sharp: zero stalls). thumbnail-backfill.js
walks a whole library at boot, so this reproduces #240 exactly — blocked loop,
missed heartbeats, panels marked offline, reconnect churn. Needs a worker_thread
offload before this is viable; image-ops.js is the seam for it.

* Run image decoding on a worker thread

Fixes the blocker the previous commit shipped with. Pure-JS decoding costs ~1s of
solid CPU for a 12MP photo, and in-process that is not a slow upload but a stalled
event loop — no heartbeats, no socket traffic. thumbnail-backfill.js walks a whole
library at boot, so it reproduced #240 (blocked loop -> missed heartbeats -> panels
offline -> reconnect churn) from our own maintenance. sharp never did this because
libvips works on a threadpool.

image-ops.js is now a dispatcher over image-ops-worker.js; the work moved unchanged
to image-ops-core.js, so callers and their failure contract are untouched.

Measured on a 12MP photo: 1079ms wall with the loop stalled 1003ms, to 1881ms wall
for two ops with ZERO stalls and 185 timer ticks serviced. Wall time is worse and
that is fine — it is off the main thread now.

Design notes, all load-bearing:
  - ONE JOB AT A TIME. A decoded 12MP bitmap is ~48MB of RGBA; overlapping jobs
    multiply peak memory by queue depth, which is the wrong failure on the small
    targets this change exists to reach. Costs no throughput — the work is CPU-bound
    and one busy worker already saturates its core.
  - unref'd while idle, ref'd only in flight. Otherwise scripts/backfill-rotation-
    dims.js never exits and `node --test` hangs forever. Verified: a CLI-style run
    exits in 104ms, code 0.
  - decode failures reply as messages, so one bad upload cannot tear down the worker
    and take unrelated queued jobs with it.
  - in-process fallback if a thread cannot be had, warned rather than silent.

test/image-ops.test.js pins the loop-liveness property, which no functional test
would catch. Its thresholds were mutation-tested against the inline path: the first
version passed there too (4MP stalls only ~355ms, under a non-flaky threshold), so
the fixture is 12MP and the thresholds sit in the gap between the two behaviours —
worker ~90 ticks/~0ms, inline ~3 ticks/~897ms. It now fails inline, as a guard must.

1647/1647 pass. Ingest re-verified with node_modules/sharp moved aside.

* Measure and thumbnail an image from a single decode

Ingest asked for metadata() then writeThumbnail(), which decoded the file twice.
That pairing was free under sharp, whose .metadata() only parses the header, but
every decode here is a full one — ~1s for a 12MP photo — so the naive translation
doubled the most expensive thing on the ingest path.

image-ops.measureAndThumbnail() returns both from one decode. Full ingest of a
12MP photo: 2 decodes/~1.9s -> 1150ms, still with zero event-loop stalls.

The subtlety is the failure contract. In the two-call version width and height
were assigned BEFORE the thumbnail was attempted, so a failed thumbnail still left
usable dimensions on the row — the player needs them to letterbox. Merging naively
would have turned any thumbnail failure into total metadata loss. So a WRITE
failure is reported ({thumbnailWritten:false, thumbnailError}) with the dimensions
intact, and the caller sets thumbnail_path only when the write succeeded, keeping
the phantom-path discipline. A DECODE failure still throws — there is nothing to
report about an unreadable image.

backfill-rotation-dims.js deliberately keeps the separate calls: it probes every
image row but regenerates a thumbnail only for the few whose dimensions changed,
so pairing them there would decode files it has no reason to thumbnail.

Tests count decodes rather than timing them — an exact property, and a wall-clock
comparison would be flaky under load. The count filters for reads of the file under
test: Node's ESM loader also goes through fs.promises.readFile, so a raw call count
picks up jimp's and the WASM codecs' lazy loading and reads 30 instead of 1.

1649/1649 pass. Ingest re-verified across all 7 formats with sharp moved aside.

* Dockerfile: sharp is no longer a production dependency

--omit=dev now leaves it out entirely; better-sqlite3 is the only native module
the builder stage still needs a toolchain for.
2026-08-13 11:40:13 -05:00

148 lines
5.9 KiB
JavaScript

'use strict';
/*
* Image operations, off the main thread.
*
* WHY THIS EXISTS: image-ops-core is pure JavaScript, so unlike the native sharp it replaced —
* which handed work to a libvips threadpool — its CPU cost lands on whatever thread calls it. A
* 12MP photo measures at ~1.0s of solid, uninterruptible main-thread work. That is not a slow
* upload, it is a stalled event loop: no heartbeats, no socket traffic, nothing. lib/thumbnail-
* backfill.js walks an entire content library at boot, so in-process it reproduces #240 exactly
* (blocked loop -> missed heartbeats -> panels marked offline -> reconnect churn), arriving from
* our own maintenance. The same reasoning already moved this file's video branch from
* execFileSync to execFile; this is that fix for the image branch.
*
* The work is therefore hosted on a worker thread and this module is the only entry point.
*
* ONE JOB AT A TIME, deliberately. Decoding holds a full RGBA bitmap — a 12MP photo is ~48MB — so
* letting jobs overlap multiplies peak memory by the queue depth, which is exactly the wrong
* failure on the small targets this whole change is meant to reach. Serialized, the ceiling is one
* image regardless of how many uploads land at once. It also costs nothing in throughput: the work
* is CPU-bound, and a single busy worker already saturates the core it runs on.
*
* The worker is unref'd while idle so it never holds the process open — scripts/backfill-rotation-
* dims.js is a CLI that must exit, and `node --test` would otherwise hang forever — and ref'd only
* while a job is in flight, so an in-progress thumbnail cannot be cut short by the process exiting.
*/
const path = require('path');
const WORKER_PATH = path.join(__dirname, 'image-ops-worker.js');
const IDLE_SHUTDOWN_MS = 60_000; // release the decoder heap (jimp + the WASM codecs) when quiet
let worker = null;
let idleTimer = null;
let inFlight = null; // { id, resolve, reject } — at most one, by design
let inlineOnly = false; // set if a worker cannot be created at all; see runInline
let nextId = 1;
const queue = [];
function clearIdleTimer() {
if (idleTimer) { clearTimeout(idleTimer); idleTimer = null; }
}
function scheduleIdleShutdown() {
clearIdleTimer();
if (!worker || inFlight || queue.length) return;
idleTimer = setTimeout(() => {
idleTimer = null;
if (worker && !inFlight && !queue.length) { const w = worker; worker = null; w.terminate(); }
}, IDLE_SHUTDOWN_MS);
idleTimer.unref?.();
}
// Reject everything outstanding. Called when the worker dies underneath us — a crash means OOM or
// a bug, not a bad image (image-ops-worker catches decode failures and replies normally), so there
// is nothing to usefully retry and callers already treat a rejection as "no metadata".
function failAll(reason) {
const dead = [inFlight, ...queue].filter(Boolean);
inFlight = null;
queue.length = 0;
for (const job of dead) job.reject(new Error(reason));
}
function ensureWorker() {
if (worker) return worker;
const { Worker } = require('worker_threads');
worker = new Worker(WORKER_PATH);
worker.unref();
worker.on('message', (msg) => {
const job = inFlight;
if (!job || job.id !== msg.id) return; // a reply from a terminated generation; ignore
inFlight = null;
if (msg.ok) job.resolve(msg.result); else job.reject(new Error(msg.error));
pump();
});
worker.on('error', (err) => { worker = null; failAll(`image worker failed: ${err.message}`); });
worker.on('exit', (code) => {
worker = null;
if (inFlight || queue.length) failAll(`image worker exited (code ${code})`);
});
return worker;
}
function pump() {
if (inFlight) return;
if (!queue.length) { worker?.unref(); scheduleIdleShutdown(); return; }
clearIdleTimer();
inFlight = queue.shift();
const w = ensureWorker();
w.ref(); // a job is running: hold the process open until it finishes
w.postMessage(inFlight.job);
}
// Last resort: if worker_threads cannot give us a thread at all, do the work in-process rather
// than refuse to thumbnail. Stalls the loop — that is the bug this module exists to avoid — so it
// is announced rather than silent.
async function runInline(job) {
const core = require('./image-ops-core');
return job.op === 'metadata'
? core.metadata(job.src)
: core.writeThumbnail(job.src, job.dest, job.width, job.quality);
}
function submit(job) {
if (inlineOnly) return runInline(job);
job.id = nextId++;
return new Promise((resolve, reject) => {
try {
ensureWorker();
} catch (err) {
inlineOnly = true;
console.warn(`[image-ops] no worker thread (${err.message}) — decoding in-process, which blocks the event loop`);
return resolve(runInline(job));
}
queue.push({ id: job.id, job, resolve, reject });
pump();
});
}
/* Display dimensions, shaped like the sharp metadata callers destructure. See image-ops-core. */
function metadata(src) {
return submit({ op: 'metadata', src });
}
/* Write a JPEG thumbnail `width` px wide, aspect preserved. Rotation is implicit in the decode. */
function writeThumbnail(src, dest, width, quality = 70) {
return submit({ op: 'writeThumbnail', src, dest, width, quality });
}
/*
* Both of the above from ONE decode -> { width, height, orientation, thumbnailWritten,
* thumbnailError }. Prefer this wherever both are wanted: a decode here is the full ~1s of a 12MP
* photo, not sharp's cheap header parse, so the pair costs double. See image-ops-core.
*/
function measureAndThumbnail(src, dest, width, quality = 70) {
return submit({ op: 'measureAndThumbnail', src, dest, width, quality });
}
/* Drop the worker now rather than waiting out the idle timer. For shutdown paths and tests. */
async function shutdown() {
clearIdleTimer();
const w = worker;
worker = null;
if (w) await w.terminate();
}
module.exports = { metadata, writeThumbnail, measureAndThumbnail, shutdown };