Add daemon restart, preview diagnostics, browse caching, and rate-limit pauses

This commit is contained in:
Fisto
2026-09-17 07:33:35 +02:00
parent cba7ae8143
commit dea1724eb3
13 changed files with 367 additions and 56 deletions
+9
View File
@@ -40,6 +40,15 @@ mobile draft at 390px.
- [x] C11 preview caching: Discord-hosted image/GIF previews (WebP 320px, hash-addressed, 5 MiB - [x] C11 preview caching: Discord-hosted image/GIF previews (WebP 320px, hash-addressed, 5 MiB
cap, one attempt) run after every walk in both modes. Live run `...-20260916090722`: cap, one attempt) run after every walk in both modes. Live run `...-20260916090722`:
`previewed: 2`; thumbs served over `/api/media/<scrape>/<item>/thumb` (200). `previewed: 2`; thumbs served over `/api/media/<scrape>/<item>/thumb` (200).
- [x] C36 Daemon restart, preview diagnostics, browse-time caching, and rate-limit pauses.
Settings → Server has a **Restart daemon** button (`POST /api/daemon/restart`: re-execs the
daemon, keeps the client running). The preview pass logs every failure with scrape/item/kind/
provider/reason plus a per-pass summary, and `GET /api/media/preview-failures` lists them.
Browsing no longer fetches Discord at all: the grid uses cached stills only (placeholders for
missing ones), GIF hover plays the cached animated GIF (kept as `anim_hash`, served from
`/anim`), and streaming only starts on explicit play. `429`s from previews/downloads raise a
cooldown (`VESKSLOP_MEDIA_RATELIMIT_PAUSE_MS`, default 60 s, or `Retry-After`), leaving rows
pending and pausing both walker and previews.
- [x] C35 Walk/preview cycling. A live scrape now alternates: walk 5 pages, preview up to one - [x] C35 Walk/preview cycling. A live scrape now alternates: walk 5 pages, preview up to one
batch (25, with the cooldown), walk on. Tiles fill in during a long walk instead of only at the batch (25, with the cooldown), walk on. Tiles fill in during a long walk instead of only at the
end, and preview traffic is spread through the walk. Configurable with end, and preview traffic is spread through the walk. Configurable with
+1
View File
@@ -181,6 +181,7 @@ Verified live on 2026-09-13: the web UI started the client under auto-Xvfb, the
| `VESKSLOP_MEDIA_PREVIEW_BATCH` | `25` | Previews per batch before the cooldown. | | `VESKSLOP_MEDIA_PREVIEW_BATCH` | `25` | Previews per batch before the cooldown. |
| `VESKSLOP_MEDIA_PREVIEW_BATCH_PAUSE_MS` | `60000` | Cooldown between preview batches. `0` disables batching. | | `VESKSLOP_MEDIA_PREVIEW_BATCH_PAUSE_MS` | `60000` | Cooldown between preview batches. `0` disables batching. |
| `VESKSLOP_MEDIA_WALK_CYCLE_PAGES` | `5` | Walk pages between preview cycles on a live scrape. `0` defers previews to the end. | | `VESKSLOP_MEDIA_WALK_CYCLE_PAGES` | `5` | Walk pages between preview cycles on a live scrape. `0` defers previews to the end. |
| `VESKSLOP_MEDIA_RATELIMIT_PAUSE_MS` | `60000` | Minimum pause after a `429` before media traffic resumes. |
| `VESKSLOP_FFPROBE` | `ffprobe` | Probe binary used to fill missing dimensions after a download. | | `VESKSLOP_FFPROBE` | `ffprobe` | Probe binary used to fill missing dimensions after a download. |
| `VESKSLOP_MEDIA_DOWNLOAD_MIN_MS` / `_MAX_MS` | `1000` / `3000` | Gap between downloaded files. | | `VESKSLOP_MEDIA_DOWNLOAD_MIN_MS` / `_MAX_MS` | `1000` / `3000` | Gap between downloaded files. |
| `VESKSLOP_AUDIO_DUCK` | `30` | Sink volume percent while others speak. | | `VESKSLOP_AUDIO_DUCK` | `30` | Sink volume percent while others speak. |
+19
View File
@@ -157,6 +157,25 @@ Previews run after every walk:
(`VESKSLOP_MEDIA_PREVIEW_BATCH`, `VESKSLOP_MEDIA_PREVIEW_BATCH_PAUSE_MS`), because bursts of (`VESKSLOP_MEDIA_PREVIEW_BATCH`, `VESKSLOP_MEDIA_PREVIEW_BATCH_PAUSE_MS`), because bursts of
thumbnail fetches trigger Discord and provider rate limits. A batch counts rows, not successes. thumbnail fetches trigger Discord and provider rate limits. A batch counts rows, not successes.
## Rate limits
Discord and the provider APIs throttle bursts. The engine detects limiting and holds all media
traffic:
- A `429` from a preview or download fetch pauses the pass; the row stays pending (never marked as a
permanent error) and is retried later.
- A walk page that fails with a `429` or a rate-limit message registers a cooldown, and every page
and preview waits it out before continuing.
- The pause is `VESKSLOP_MEDIA_RATELIMIT_PAUSE_MS` (default 60 s), or the server's `Retry-After`
when it is longer. Detection is logged as `rate limit detected; pausing media traffic until …`.
## Browsing cache
Browsing never calls Discord. The grid renders cached previews or local files only; videos without
a cached still show a labeled placeholder, and playback streams only when the user presses play.
GIF hover plays the animated GIF cached in `thumbs/` (the original is kept alongside the static
frame and referenced by `anim_hash`), never the provider, so scrolling cannot spam anyone.
## Job states ## Job states
`queued`, `running`, `paused`, `stopping`, `done`, `failed`, `pending-delete`, `deleting`, `deleted`. `queued`, `running`, `paused`, `stopping`, `done`, `failed`, `pending-delete`, `deleting`, `deleted`.
+3
View File
@@ -271,6 +271,9 @@ replaceable.
| POST | `/api/media/count` | `channelId`, `guildId?`, `after?`, `before?` | Discord search total for the channel or range. Returns `{ count }` or `{ count: null, reason }` (slopbot guild channels only) | | POST | `/api/media/count` | `channelId`, `guildId?`, `after?`, `before?` | Discord search total for the channel or range. Returns `{ count }` or `{ count: null, reason }` (slopbot guild channels only) |
| GET | `/api/media/errors` | none | Bot and scrape errors, newest first | | GET | `/api/media/errors` | none | Bot and scrape errors, newest first |
| POST | `/api/media/errors/clear` | none | Clear the error list | | POST | `/api/media/errors/clear` | none | Clear the error list |
| GET | `/api/media/preview-failures` | `limit?` | Rows whose preview failed, with the reason |
| GET | `/api/media/:scrapeId/:itemId/anim` | none | The cached animated GIF for hover playback |
| POST | `/api/daemon/restart` | none | Replace the daemon process with a fresh copy (the client keeps running) |
| POST | `/api/media/scrape/stop` | `id` | Stop a media scrape | | POST | `/api/media/scrape/stop` | `id` | Stop a media scrape |
| POST | `/api/media/scrape/resume` | `id` | Continue a paused scrape from its cursor | | POST | `/api/media/scrape/resume` | `id` | Continue a paused scrape from its cursor |
| POST | `/api/media/clean` | `ids?` | Delete the given scrapes, or every scrape when `ids` is absent | | POST | `/api/media/clean` | `ids?` | Delete the given scrapes, or every scrape when `ids` is absent |
+22 -1
View File
@@ -80,12 +80,14 @@ export interface DownloadOptions {
export interface DownloadOutcome { export interface DownloadOutcome {
id: number; id: number;
state: "downloaded" | "skipped" | "failed"; state: "downloaded" | "skipped" | "failed" | "ratelimited";
path?: string; path?: string;
hash?: string; hash?: string;
bytes?: number; bytes?: number;
/** File extension of the stored bytes, when known. */ /** File extension of the stored bytes, when known. */
ext?: string; ext?: string;
/** Seconds the server asked us to wait, when it is known. */
retryAfterMs?: number;
error?: string; error?: string;
} }
@@ -157,6 +159,16 @@ export async function downloadOne(
const response = await fetchImpl(media.url); const response = await fetchImpl(media.url);
if (!response.ok) { if (!response.ok) {
if (response.status === 429) {
const after = Number(response.headers.get("retry-after") ?? 0);
return {
id: media.id,
state: "ratelimited",
retryAfterMs: Number.isFinite(after) && after > 0 ? after * 1000 : undefined,
error: "http 429"
};
}
if (SKIP_STATUS.has(response.status)) { if (SKIP_STATUS.has(response.status)) {
return { id: media.id, state: "skipped", error: `http ${response.status}` }; return { id: media.id, state: "skipped", error: `http ${response.status}` };
} }
@@ -252,6 +264,15 @@ export async function downloadPreview(
try { try {
const response = await fetchImpl(target); const response = await fetchImpl(target);
if (response.status === 429) {
const after = Number(response.headers.get("retry-after") ?? 0);
return {
id: media.id,
state: "ratelimited",
retryAfterMs: Number.isFinite(after) && after > 0 ? after * 1000 : undefined,
error: "http 429"
};
}
if (!response.ok) return { id: media.id, state: "skipped", error: `http ${response.status}` }; if (!response.ok) return { id: media.id, state: "skipped", error: `http ${response.status}` };
const buffer = Buffer.from(await response.arrayBuffer()); const buffer = Buffer.from(await response.arrayBuffer());
+75 -4
View File
@@ -64,6 +64,8 @@ export interface ScrapeEngineOptions {
previews?: boolean; previews?: boolean;
/** Walk pages between preview cycles. 0 defers all previews to the end. Default 5. */ /** Walk pages between preview cycles. 0 defers all previews to the end. Default 5. */
walkCyclePages?: number; walkCyclePages?: number;
/** Pause after a rate limit is detected. Default 60 s. */
rateLimitPauseMs?: number;
/** Read dimensions and duration from a downloaded file. Defaults to ffprobe. */ /** Read dimensions and duration from a downloaded file. Defaults to ffprobe. */
probe?: ( probe?: (
path: string path: string
@@ -88,11 +90,17 @@ function oldestId(messages: CoreMessage[]): string | null {
return oldest; return oldest;
} }
function isRateLimited(error: unknown): boolean {
const anyError = error as { status?: number; message?: string } | undefined;
return anyError?.status === 429 || /\b429\b|rate.?limit/i.test(String(anyError?.message ?? error));
}
export class ScrapeEngine { export class ScrapeEngine {
private readonly queue: QueuedJob[] = []; private readonly queue: QueuedJob[] = [];
private readonly active = new Map<string, ScrapeMeta>(); private readonly active = new Map<string, ScrapeMeta>();
private readonly stopRequested = new Set<string>(); private readonly stopRequested = new Set<string>();
private readonly pageTimes: number[] = []; private readonly pageTimes: number[] = [];
private rateLimitUntil = 0;
private processingPromise: Promise<void> | null = null; private processingPromise: Promise<void> | null = null;
constructor(private readonly options: ScrapeEngineOptions) {} constructor(private readonly options: ScrapeEngineOptions) {}
@@ -264,6 +272,7 @@ export class ScrapeEngine {
break; break;
} }
await this.respectRateLimit();
const page = await this.fetchWithBackoff(meta, backend); const page = await this.fetchWithBackoff(meta, backend);
if (page.rateLimitedMs) await this.sleep(page.rateLimitedMs); if (page.rateLimitedMs) await this.sleep(page.rateLimitedMs);
@@ -366,6 +375,8 @@ export class ScrapeEngine {
return; return;
} }
await this.respectRateLimit();
const outcome = await downloadOne(row, targetDir, { const outcome = await downloadOne(row, targetDir, {
...options, ...options,
log: line => this.log(`${meta.id}: ${line}`) log: line => this.log(`${meta.id}: ${line}`)
@@ -394,6 +405,10 @@ export class ScrapeEngine {
meta.previewed = (meta.previewed ?? 0) + 1; meta.previewed = (meta.previewed ?? 0) + 1;
} }
} }
} else if (outcome.state === "ratelimited") {
// Leave the row pending and hold off: the server asked for a break.
this.registerRateLimit(outcome.retryAfterMs);
this.log(`${meta.id}: download #${row.id} rate limited (${outcome.error ?? "429"})`);
} else { } else {
store.markDownloadError(row.id, outcome.error ?? outcome.state); store.markDownloadError(row.id, outcome.error ?? outcome.state);
if (outcome.state === "failed") meta.downloadErrors = (meta.downloadErrors ?? 0) + 1; if (outcome.state === "failed") meta.downloadErrors = (meta.downloadErrors ?? 0) + 1;
@@ -412,6 +427,11 @@ export class ScrapeEngine {
destinationDir: meta.downloadDir ?? null destinationDir: meta.downloadDir ?? null
}; };
const pending = store.pendingPreviews(limit > 0 ? limit : 5000, true); const pending = store.pendingPreviews(limit > 0 ? limit : 5000, true);
let cached = 0;
let skipped = 0;
let failed = 0;
if (pending.length) this.log(`${meta.id}: preview pass: ${pending.length} pending`);
const batchSize = Math.max(1, options.previewBatchSize ?? 25); const batchSize = Math.max(1, options.previewBatchSize ?? 25);
const batchPauseMs = Math.max(0, options.previewBatchPauseMs ?? 60_000); const batchPauseMs = Math.max(0, options.previewBatchPauseMs ?? 60_000);
let batch = 0; let batch = 0;
@@ -432,6 +452,8 @@ export class ScrapeEngine {
return; return;
} }
await this.respectRateLimit();
// Audio has no image proxy: render a waveform strip with ffmpeg. // Audio has no image proxy: render a waveform strip with ffmpeg.
if (row.kind === "audio" && row.source !== "embed") { if (row.kind === "audio" && row.source !== "embed") {
const waveform = await remoteFrame(row.url, join(dir, "thumbs"), { waveform: true }); const waveform = await remoteFrame(row.url, join(dir, "thumbs"), { waveform: true });
@@ -444,6 +466,8 @@ export class ScrapeEngine {
meta.previewed = (meta.previewed ?? 0) + 1; meta.previewed = (meta.previewed ?? 0) + 1;
} else { } else {
store.markPreviewError(row.id, "waveform failed"); store.markPreviewError(row.id, "waveform failed");
failed++;
this.log(`${meta.id}: preview #${row.id} audio waveform failed (${row.url.slice(0, 60)})`);
} }
this.persist(dir, meta); this.persist(dir, meta);
await this.sleep(previewGapMs(options)); await this.sleep(previewGapMs(options));
@@ -458,6 +482,8 @@ export class ScrapeEngine {
const resolved = await resolveEmbedPreview(row, options); const resolved = await resolveEmbedPreview(row, options);
if (!resolved) { if (!resolved) {
store.markPreviewError(row.id, "no preview source"); store.markPreviewError(row.id, "no preview source");
skipped++;
this.log(`${meta.id}: preview #${row.id} ${row.kind}/${row.provider ?? "?"} has no preview source`);
this.persist(dir, meta); this.persist(dir, meta);
await this.sleep(previewGapMs(options)); await this.sleep(previewGapMs(options));
await afterRow(); await afterRow();
@@ -473,19 +499,22 @@ export class ScrapeEngine {
let bytes = outcome.bytes ?? 0; let bytes = outcome.bytes ?? 0;
let ext = outcome.ext ?? "webp"; let ext = outcome.ext ?? "webp";
// Animated GIFs are kept static in the grid; hovering plays the original. // Animated GIFs show a static frame in the grid; hovering plays the cached
// original from disk, so browsing never hotlinks the provider.
let animHash: string | undefined;
if (ext === "gif") { if (ext === "gif") {
const poster = await gifPoster(join(dir, "thumbs", `${hash}.gif`), join(dir, "thumbs")); const poster = await gifPoster(join(dir, "thumbs", `${hash}.gif`), join(dir, "thumbs"));
if (poster) { if (poster) {
rmSync(join(dir, "thumbs", `${hash}.gif`), { force: true }); animHash = hash;
hash = poster.hash; hash = poster.hash;
bytes = poster.bytes; bytes = poster.bytes;
ext = "webp"; ext = "webp";
} }
} }
store.markPreview(row.id, { hash, bytes, ext }); store.markPreview(row.id, { hash, bytes, ext, animHash });
meta.previewed = (meta.previewed ?? 0) + 1; meta.previewed = (meta.previewed ?? 0) + 1;
cached++;
// Embed thumbnails also land in the custom destination, so the download has a // Embed thumbnails also land in the custom destination, so the download has a
// viewable poster for embeds while the media tab keeps its local copy. // viewable poster for embeds while the media tab keeps its local copy.
@@ -506,18 +535,57 @@ export class ScrapeEngine {
meta.previewed = (meta.previewed ?? 0) + 1; meta.previewed = (meta.previewed ?? 0) + 1;
} else { } else {
store.markPreviewError(row.id, "too large"); store.markPreviewError(row.id, "too large");
skipped++;
this.log(`${meta.id}: preview #${row.id} gif too large and no frame (${row.url.slice(0, 60)})`);
} }
} else if (outcome.state === "ratelimited") {
// Leave the row pending and hold off: the server asked for a break.
this.registerRateLimit(outcome.retryAfterMs);
this.log(`${meta.id}: preview #${row.id} rate limited (${outcome.error ?? "429"})`);
} else if (outcome.state === "failed") { } else if (outcome.state === "failed") {
// Transient network failure: leave the row pending for the next pass. // Transient network failure: leave the row pending for the next pass.
this.log(`${meta.id}: preview ${row.id} failed: ${outcome.error ?? "unknown"}`); failed++;
this.log(
`${meta.id}: preview #${row.id} ${row.kind}/${row.provider ?? "?"} failed (transient): ` +
`${outcome.error ?? "unknown"} (${row.url.slice(0, 60)})`
);
} else { } else {
store.markPreviewError(row.id, outcome.error ?? outcome.state); store.markPreviewError(row.id, outcome.error ?? outcome.state);
skipped++;
this.log(
`${meta.id}: preview #${row.id} ${row.kind}/${row.provider ?? "?"} skipped: ` +
`${outcome.error ?? outcome.state} (${row.url.slice(0, 60)})`
);
} }
this.persist(dir, meta); this.persist(dir, meta);
await this.sleep(previewGapMs(options)); await this.sleep(previewGapMs(options));
await afterRow(); await afterRow();
} }
const remaining = store.pendingPreviews(5000, true).length;
if (pending.length) {
this.log(
`${meta.id}: preview pass done: ${cached} cached, ${skipped} skipped, ${failed} failed, ` +
`${remaining} still pending`
);
}
}
/** A rate limit was seen: hold all walk and preview traffic for a while. */
private registerRateLimit(retryAfterMs?: number): void {
const pause = Math.max(this.options.rateLimitPauseMs ?? 60_000, retryAfterMs ?? 0);
const until = this.now() + pause;
if (until > this.rateLimitUntil) {
this.rateLimitUntil = until;
this.log(`rate limit detected; pausing media traffic until ${new Date(until).toISOString()}`);
}
}
private async respectRateLimit(): Promise<void> {
const wait = this.rateLimitUntil - this.now();
if (wait > 0) await this.sleep(wait);
} }
private async fetchWithBackoff(meta: ScrapeMeta, backend: ScrapeBackend): Promise<FetchPage> { private async fetchWithBackoff(meta: ScrapeMeta, backend: ScrapeBackend): Promise<FetchPage> {
@@ -530,6 +598,9 @@ export class ScrapeEngine {
return await backend.fetchPage(meta.channelId, meta.cursor, pageSize); return await backend.fetchPage(meta.channelId, meta.cursor, pageSize);
} catch (error) { } catch (error) {
meta.errors++; meta.errors++;
if (isRateLimited(error)) {
this.registerRateLimit((error as { retryAfterMs?: number }).retryAfterMs);
}
if (attempt >= maxRetries) throw error; if (attempt >= maxRetries) throw error;
const retryAfter = (error as { retryAfterMs?: number }).retryAfterMs; const retryAfter = (error as { retryAfterMs?: number }).retryAfterMs;
+38 -1
View File
@@ -135,6 +135,7 @@ export interface MediaRow {
file_hash: string | null; file_hash: string | null;
preview_hash: string | null; preview_hash: string | null;
preview_ext: string | null; preview_ext: string | null;
anim_hash: string | null;
file_bytes: number | null; file_bytes: number | null;
preview_bytes: number | null; preview_bytes: number | null;
file_path: string | null; file_path: string | null;
@@ -173,6 +174,7 @@ CREATE TABLE IF NOT EXISTS media (
file_hash TEXT, file_hash TEXT,
preview_hash TEXT, preview_hash TEXT,
preview_ext TEXT, preview_ext TEXT,
anim_hash TEXT,
file_bytes INTEGER, file_bytes INTEGER,
preview_bytes INTEGER, preview_bytes INTEGER,
file_path TEXT, file_path TEXT,
@@ -217,6 +219,7 @@ export class ScrapeStore {
if (!names.has("download_error")) this.db.exec("ALTER TABLE media ADD COLUMN download_error TEXT"); if (!names.has("download_error")) this.db.exec("ALTER TABLE media ADD COLUMN download_error TEXT");
if (!names.has("preview_error")) this.db.exec("ALTER TABLE media ADD COLUMN preview_error TEXT"); if (!names.has("preview_error")) this.db.exec("ALTER TABLE media ADD COLUMN preview_error TEXT");
if (!names.has("preview_ext")) this.db.exec("ALTER TABLE media ADD COLUMN preview_ext TEXT"); if (!names.has("preview_ext")) this.db.exec("ALTER TABLE media ADD COLUMN preview_ext TEXT");
if (!names.has("anim_hash")) this.db.exec("ALTER TABLE media ADD COLUMN anim_hash TEXT");
if (!names.has("play_url")) this.db.exec("ALTER TABLE media ADD COLUMN play_url TEXT"); if (!names.has("play_url")) this.db.exec("ALTER TABLE media ADD COLUMN play_url TEXT");
if (!names.has("aspect_ratio")) this.db.exec("ALTER TABLE media ADD COLUMN aspect_ratio REAL"); if (!names.has("aspect_ratio")) this.db.exec("ALTER TABLE media ADD COLUMN aspect_ratio REAL");
} }
@@ -355,7 +358,16 @@ export class ScrapeStore {
.all(limit) as unknown as MediaRow[]; .all(limit) as unknown as MediaRow[];
} }
markPreview(id: number, preview: { hash: string; bytes: number; ext?: string }): void { markPreview(id: number, preview: { hash: string; bytes: number; ext?: string; animHash?: string }): void {
if (preview.animHash) {
this.db
.prepare(
"UPDATE media SET preview_hash = ?, preview_bytes = ?, preview_ext = ?, anim_hash = ?, preview_error = NULL WHERE id = ?"
)
.run(preview.hash, preview.bytes, preview.ext ?? "webp", preview.animHash, id);
return;
}
this.db this.db
.prepare( .prepare(
"UPDATE media SET preview_hash = ?, preview_bytes = ?, preview_ext = ?, preview_error = NULL WHERE id = ?" "UPDATE media SET preview_hash = ?, preview_bytes = ?, preview_ext = ?, preview_error = NULL WHERE id = ?"
@@ -385,6 +397,31 @@ export class ScrapeStore {
this.db.prepare(`UPDATE media SET ${sets.join(", ")} WHERE id = ?`).run(...values, id); this.db.prepare(`UPDATE media SET ${sets.join(", ")} WHERE id = ?`).run(...values, id);
} }
/** Rows whose preview failed, with the reason, newest first. */
previewFailures(
limit = 50
): Array<{ id: number; kind: MediaKind; provider: string | null; url: string; error: string }> {
const rows = this.db
.prepare(
"SELECT id, kind, provider, url, preview_error FROM media WHERE preview_error IS NOT NULL ORDER BY id DESC LIMIT ?"
)
.all(limit) as unknown as Array<{
id: number;
kind: MediaKind;
provider: string | null;
url: string;
preview_error: string;
}>;
return rows.map(row => ({
id: row.id,
kind: row.kind,
provider: row.provider,
url: row.url,
error: row.preview_error
}));
}
/** The message a media row came from, for URL refreshes. */ /** The message a media row came from, for URL refreshes. */
sourceMessageId(mediaId: number): string | null { sourceMessageId(mediaId: number): string | null {
const row = this.db const row = this.db
+73
View File
@@ -877,6 +877,79 @@ void (async () => {
assert.equal(existsSync(join(dir, `${poster?.hash}.webp`)), true); assert.equal(existsSync(join(dir, `${poster?.hash}.webp`)), true);
}); });
await test("markPreview stores the animated gif reference", () => {
const dir = join(root, "anim-store");
mkdirSync(dir, { recursive: true });
const store = new ScrapeStore(dir);
store.insertPage([
message("1200", { attachments: [attachment({ filename: "loop.gif" })] })
]);
const row = store.queryMedia({ kind: "gif" })[0];
store.markPreview(row.id, { hash: "static-hash", bytes: 10, ext: "webp", animHash: "anim-hash" });
const updated = store.getMedia(row.id);
assert.equal(updated?.preview_hash, "static-hash");
assert.equal(updated?.anim_hash, "anim-hash");
store.close();
});
await test("rate limits pause media traffic", async () => {
const backend = new FakeBackend([
{
messages: [
message("1300", {
attachments: [
attachment({ filename: "a.png", contentType: "image/png" }),
attachment({ filename: "b.png", contentType: "image/png" })
]
})
]
}
]);
const sleeps: number[] = [];
const engine = new ScrapeEngine({
rootDir: root,
pageSize: 10,
pageMinMs: 0,
pageMaxMs: 0,
sleep: async (ms: number) => {
sleeps.push(ms);
},
random: () => 0,
log: () => {},
previews: true,
rateLimitPauseMs: 30_000,
download: {
fetchImpl: async () =>
new Response("slow down", { status: 429, headers: { "retry-after": "10" } }),
minGapMs: 0,
maxGapMs: 0,
previewMinGapMs: 0,
previewMaxGapMs: 0,
previewBatchSize: 100,
sleep: async () => {}
}
});
const started = await engine.start({ channelId: "160", guildId: null, mode: "links", backend });
await engine.processQueue();
const store = new ScrapeStore(join(root, started.id));
for (const row of store.queryMedia({ kind: "image" })) {
assert.equal(row.preview_hash, null);
assert.equal(row.preview_error, null, "rate limited rows stay pending");
}
store.close();
assert.ok(
sleeps.some(ms => ms > 25_000 && ms <= 30_000),
"the engine paused after the rate limit"
);
});
await test("previews run in cycles during the walk", async () => { await test("previews run in cycles during the walk", async () => {
const events: string[] = []; const events: string[] = [];
const pages: Array<{ messages: CoreMessage[] }> = [ const pages: Array<{ messages: CoreMessage[] }> = [
+2 -1
View File
@@ -463,7 +463,8 @@
<span class="vk-settings-actions"> <span class="vk-settings-actions">
<button id="core-start" class="btn btn-primary vk-small-btn" type="button">Start</button> <button id="core-start" class="btn btn-primary vk-small-btn" type="button">Start</button>
<button id="core-stop" class="btn btn-secondary vk-small-btn" type="button">Stop</button> <button id="core-stop" class="btn btn-secondary vk-small-btn" type="button">Stop</button>
<button id="core-restart" class="btn btn-secondary vk-small-btn" type="button">Restart</button> <button id="core-restart" class="btn btn-secondary vk-small-btn" type="button">Restart client</button>
<button id="daemon-restart" class="btn btn-secondary vk-small-btn" type="button">Restart daemon</button>
</span> </span>
</div> </div>
<div id="core-status" class="vk-settings-card"></div> <div id="core-status" class="vk-settings-card"></div>
+14 -45
View File
@@ -443,13 +443,6 @@
media.replaceChildren(node); media.replaceChildren(node);
} }
/** Only real video files can be framed; embed pages (YouTube etc.) never are. */
function canFrameVideo(item) {
const direct = /\.(mp4|webm|m4v)(\?|$)/i;
if (item.playUrl && direct.test(item.playUrl)) return true;
return item.source !== "embed" && direct.test(item.url ?? "");
}
function playIcon(item) { function playIcon(item) {
const play = el("span", "vk-media-play", "▶"); const play = el("span", "vk-media-play", "▶");
play.title = "Play"; play.title = "Play";
@@ -472,11 +465,8 @@
const media = el("div", "vk-media-media"); const media = el("div", "vk-media-media");
const imageSource = item.localUrl || item.thumbUrl; const imageSource = item.localUrl || item.thumbUrl;
const fallbackSource = item.source !== "embed" ? item.url : null; const fallbackSource = item.source !== "embed" ? item.url : null;
const streamPath = `/api/media/${item.scrapeId}/${item.itemId}/stream`; // Hover plays the cached animated GIF from the archive, never the provider.
const animatedGif = const animatedGif = item.kind === "gif" && item.animUrl ? item.animUrl : null;
item.kind === "gif" && item.remotePreviewUrl && /\.gif(\?|$)/i.test(item.remotePreviewUrl)
? item.remotePreviewUrl
: null;
if (imageSource) { if (imageSource) {
const img = document.createElement("img"); const img = document.createElement("img");
@@ -496,23 +486,6 @@
} }
media.append(img); media.append(img);
} else if (item.kind === "video" && canFrameVideo(item)) {
// No thumbnail: take the first frame through the same-origin stream endpoint,
// which refreshes expired URLs and avoids cross-origin blocking.
const video = document.createElement("video");
video.preload = "metadata";
video.muted = true;
video.playsInline = true;
video.src = streamPath;
video.addEventListener(
"loadedmetadata",
() => {
if (video.duration) video.currentTime = Math.min(0.1, video.duration / 2);
},
{ once: true }
);
video.addEventListener("error", () => placeholder(media, item), { once: true });
media.append(video);
} else if (fallbackSource && /^(https?:|\/)/.test(fallbackSource)) { } else if (fallbackSource && /^(https?:|\/)/.test(fallbackSource)) {
const img = document.createElement("img"); const img = document.createElement("img");
img.loading = "lazy"; img.loading = "lazy";
@@ -1231,22 +1204,6 @@
{ once: true } { once: true }
); );
media.append(video); media.append(video);
} else if (item.kind === "video" && canFrameVideo(item)) {
// No still available: show the first frame through the stream endpoint.
const video = document.createElement("video");
video.preload = "metadata";
video.muted = true;
video.playsInline = true;
video.src = streamUrl;
video.addEventListener(
"loadedmetadata",
() => {
if (video.duration) video.currentTime = Math.min(0.1, video.duration / 2);
},
{ once: true }
);
media.append(video);
media.append(viewerPlayIcon(item));
} else if ((item.kind === "image" || item.kind === "gif") && /^(https?:|\/)/.test(source)) { } else if ((item.kind === "image" || item.kind === "gif") && /^(https?:|\/)/.test(source)) {
const img = document.createElement("img"); const img = document.createElement("img");
img.src = source; img.src = source;
@@ -1254,6 +1211,7 @@
media.append(img); media.append(img);
} else { } else {
media.append(el("span", "vk-media-card-kind", item.kind.toUpperCase())); media.append(el("span", "vk-media-card-kind", item.kind.toUpperCase()));
if (item.kind === "video") media.append(viewerPlayIcon(item));
} }
meta.replaceChildren(); meta.replaceChildren();
@@ -1490,6 +1448,17 @@
document.querySelector("#media-clean")?.addEventListener("click", openCleanModal); document.querySelector("#media-clean")?.addEventListener("click", openCleanModal);
document.querySelector("#daemon-restart")?.addEventListener("click", async () => {
if (!window.confirm("Restart the daemon? The page will disconnect for a few seconds.")) return;
try {
await window.VK.api.post("/api/daemon/restart", {});
} catch {
// The daemon exits mid-response on purpose.
}
setStatus("daemon restarting; reload the page in a few seconds");
});
window.addEventListener("resize", () => { window.addEventListener("resize", () => {
if (state.view) layoutTop(); if (state.view) layoutTop();
}); });
+33 -4
View File
@@ -9,6 +9,7 @@
* VESKSLOP_SOCKET=/path/to/vesktop-puppet.sock MCP_PORT=8788 pnpm control:bridge * VESKSLOP_SOCKET=/path/to/vesktop-puppet.sock MCP_PORT=8788 pnpm control:bridge
*/ */
import { spawn } from "node:child_process";
import { config as loadEnv } from "dotenv"; import { config as loadEnv } from "dotenv";
import { createMcpExpressApp } from "@modelcontextprotocol/sdk/server/express.js"; import { createMcpExpressApp } from "@modelcontextprotocol/sdk/server/express.js";
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js"; import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
@@ -286,6 +287,7 @@ const webui = startWebUi({
transcripts, transcripts,
preferences, preferences,
media, media,
restartDaemon,
connectAudioDevices: () => autoConnectAudio(requireCore()) connectAudioDevices: () => autoConnectAudio(requireCore())
}); });
@@ -1004,11 +1006,38 @@ app.listen(PORT, HOST, (error?: Error) => {
if (stale) console.warn(`[daemon] ${stale}`); if (stale) console.warn(`[daemon] ${stale}`);
}); });
function shutdown() { function shutdown(keepCore = false) {
stopSpeakingPoll(); stopSpeakingPoll();
core?.close(); core?.close();
void Promise.allSettled([coreProcess.dispose(4000), audio.dispose()]).finally(() => process.exit(0));
const tasks: Array<Promise<unknown>> = [audio.dispose()];
if (!keepCore) tasks.push(coreProcess.dispose(4000));
void Promise.allSettled(tasks).finally(() => process.exit(0));
} }
process.on("SIGINT", shutdown); /** Replace this daemon with a fresh copy, leaving the client running. */
process.on("SIGTERM", shutdown); let restartingDaemon = false;
function restartDaemon(): void {
if (restartingDaemon) return;
restartingDaemon = true;
const quote = (value: string) => `'${value.replace(/'/g, `'\\''`)}'`;
const args = [...process.execArgv, ...process.argv.slice(1)];
const command = `${quote(process.execPath)} ${args.map(quote).join(" ")}`;
console.log("[daemon] restart requested: spawning a replacement");
const child = spawn("sh", ["-c", `sleep 2; exec ${command}`], {
detached: true,
stdio: "ignore",
cwd: process.cwd(),
env: { ...process.env, VESKSLOP_DAEMON_RESTART: "1" }
});
child.unref();
setTimeout(() => shutdown(true), 300);
}
process.on("SIGINT", () => shutdown());
process.on("SIGTERM", () => shutdown());
+51
View File
@@ -123,6 +123,8 @@ export interface MediaItemResult {
remotePreviewUrl: string | null; remotePreviewUrl: string | null;
/** Why no preview exists, when a fetch failed (for example "http 404"). */ /** Why no preview exists, when a fetch failed (for example "http 404"). */
previewError: string | null; previewError: string | null;
/** Local animated GIF for hover playback, when the original was cached. */
animUrl: string | null;
localUrl: string | null; localUrl: string | null;
thumbUrl: string | null; thumbUrl: string | null;
provider: string | null; provider: string | null;
@@ -175,6 +177,7 @@ export class MediaService {
log: line => this.deps.log?.(line), log: line => this.deps.log?.(line),
onError: (scrapeId, message) => this.recordError("scrape", `${scrapeId}: ${message}`), onError: (scrapeId, message) => this.recordError("scrape", `${scrapeId}: ${message}`),
walkCyclePages: number(process.env.VESKSLOP_MEDIA_WALK_CYCLE_PAGES, 5), walkCyclePages: number(process.env.VESKSLOP_MEDIA_WALK_CYCLE_PAGES, 5),
rateLimitPauseMs: number(process.env.VESKSLOP_MEDIA_RATELIMIT_PAUSE_MS, 60_000),
download: { download: {
maxFileBytes: number(process.env.VESKSLOP_MEDIA_MAX_FILE_BYTES, 50 * 1024 * 1024), maxFileBytes: number(process.env.VESKSLOP_MEDIA_MAX_FILE_BYTES, 50 * 1024 * 1024),
previewMaxBytes: number(process.env.VESKSLOP_MEDIA_PREVIEW_MAX_BYTES, 5 * 1024 * 1024), previewMaxBytes: number(process.env.VESKSLOP_MEDIA_PREVIEW_MAX_BYTES, 5 * 1024 * 1024),
@@ -559,6 +562,17 @@ export class MediaService {
}; };
} }
/** Absolute path for a cached animated GIF, or null. */
getMediaAnim(scrapeId: string, itemId: number): { path: string; contentType: string } | null {
const item = this.getMediaItem(scrapeId, itemId);
if (!item?.anim_hash) return null;
return {
path: join(this.rootDir, scrapeId, "thumbs", `${item.anim_hash}.gif`),
contentType: "image/gif"
};
}
/** Absolute path for a cached preview, or null. */ /** Absolute path for a cached preview, or null. */
getMediaThumb(scrapeId: string, itemId: number): { path: string; contentType: string } | null { getMediaThumb(scrapeId: string, itemId: number): { path: string; contentType: string } | null {
const item = this.getMediaItem(scrapeId, itemId); const item = this.getMediaItem(scrapeId, itemId);
@@ -653,6 +667,42 @@ export class MediaService {
return totals; return totals;
} }
/** Preview failures across scrapes with their reasons, newest scrape first. */
previewFailures(
limit = 50
): Array<{
scrapeId: string;
itemId: number;
kind: MediaKind;
provider: string | null;
url: string;
error: string;
}> {
const results: Array<{
scrapeId: string;
itemId: number;
kind: MediaKind;
provider: string | null;
url: string;
error: string;
}> = [];
for (const meta of this.engine.list()) {
if (results.length >= limit) break;
const store = new ScrapeStore(scrapeDir(this.rootDir, meta.id));
try {
for (const row of store.previewFailures(limit - results.length)) {
results.push({ scrapeId: meta.id, itemId: row.id, ...row });
}
} finally {
store.close();
}
}
return results;
}
/** Distinct posters across every scrape, biggest first. */ /** Distinct posters across every scrape, biggest first. */
listAuthors(): Array<{ id: string | null; name: string | null; count: number }> { listAuthors(): Array<{ id: string | null; name: string | null; count: number }> {
const totals = new Map<string, { id: string | null; name: string | null; count: number }>(); const totals = new Map<string, { id: string | null; name: string | null; count: number }>();
@@ -801,6 +851,7 @@ export class MediaService {
playUrl: row.play_url, playUrl: row.play_url,
remotePreviewUrl: row.preview_url, remotePreviewUrl: row.preview_url,
previewError: row.preview_error, previewError: row.preview_error,
animUrl: row.anim_hash ? `/api/media/${meta.id}/${row.id}/anim` : null,
localUrl: localUrl:
row.file_path && !meta.downloadDir row.file_path && !meta.downloadDir
? `/api/media/${meta.id}/${row.id}/file` ? `/api/media/${meta.id}/${row.id}/file`
+27
View File
@@ -75,6 +75,8 @@ export interface WebUiDeps {
transcripts: TranscriptStore; transcripts: TranscriptStore;
preferences: PreferencesStore; preferences: PreferencesStore;
media: MediaService; media: MediaService;
/** Replace the daemon process with a fresh copy, keeping the client running. */
restartDaemon: () => void;
connectAudioDevices: () => Promise<{ connected: boolean; message: string }>; connectAudioDevices: () => Promise<{ connected: boolean; message: string }>;
} }
@@ -267,6 +269,22 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
}) })
); );
app.post(
"/api/daemon/restart",
wrap(async (_req, res) => {
res.json({ ok: true, result: { restarting: true } });
deps.restartDaemon();
})
);
app.get(
"/api/media/preview-failures",
wrap(async (req, res) => {
const limit = clampLimit(req.query.limit, 50, 500);
res.json({ ok: true, result: deps.media.previewFailures(limit) });
})
);
app.post( app.post(
"/api/core/restart", "/api/core/restart",
wrap(async (req, res) => { wrap(async (req, res) => {
@@ -764,6 +782,15 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
}) })
); );
app.get(
"/api/media/:scrapeId/:itemId/anim",
wrap(async (req, res) => {
const file = deps.media.getMediaAnim(String(req.params.scrapeId), Number(req.params.itemId));
if (!file || !existsSync(file.path)) throw new HttpError(404, "anim not found");
res.type(file.contentType).sendFile(file.path);
})
);
app.get( app.get(
"/api/media/:scrapeId/:itemId/file", "/api/media/:scrapeId/:itemId/file",
wrap(async (req, res) => { wrap(async (req, res) => {