diff options
Diffstat (limited to 'names')
| -rw-r--r-- | names/compose-libretranslate.yml | 26 | ||||
| -rw-r--r-- | names/server.js | 17 | ||||
| -rw-r--r-- | names/translate.js | 82 |
3 files changed, 125 insertions, 0 deletions
diff --git a/names/compose-libretranslate.yml b/names/compose-libretranslate.yml new file mode 100644 index 0000000..80ed9c6 --- /dev/null +++ b/names/compose-libretranslate.yml @@ -0,0 +1,26 @@ +# LibreTranslate for post translation (coinos v3). Reached only by the +# names registrar over the docker network (POST /translate there); no port +# on the host, no web UI. Models live in the lt-data volume. +services: + libretranslate: + image: libretranslate/libretranslate:v1.9.6 + container_name: libretranslate + environment: + LT_LOAD_ONLY: en,es,fr,de,pt,id,tr,vi,ru,ar,ur,hi,bn,ja,ko,zh,it,nl,pl,uk,fa,he,th,sv,cs,el,ro,hu,fi,da,nb,ca + LT_DISABLE_WEB_UI: "true" + LT_CHAR_LIMIT: "5000" + LT_THREADS: "4" + LT_UPDATE_MODELS: "true" + volumes: + - lt-data:/home/libretranslate/.local + networks: [net] + deploy: + resources: + limits: + memory: 8g + restart: unless-stopped +volumes: + lt-data: +networks: + net: + external: true diff --git a/names/server.js b/names/server.js index 6cd82a1..463ab8e 100644 --- a/names/server.js +++ b/names/server.js @@ -28,6 +28,7 @@ import { mnemonicToSeedSync } from '@scure/bip39'; import { lnBackend } from '../bridge/ln.js'; import { ArkManager } from '../src/ark/manager.js'; import { fetchPreview } from './preview.js'; +import { translate, languages } from './translate.js'; import { decodeAddress } from '../src/ark/proto.js'; import { decodeNoffer } from '../src/noffer.js'; import { decodeBolt11 } from '../src/ark/lightning.js'; @@ -1036,6 +1037,22 @@ Bun.serve({ }); } + // A post in another language, in the reader's (LibreTranslate behind a + // cache — see translate.js). POST { text, target, source? } → + // { text, source }; GET /translate/languages lists what it can do. + if (url.pathname === '/translate/languages' && req.method === 'GET') { + try { return new Response(JSON.stringify(await languages()), { headers: { 'content-type': 'application/json', 'access-control-allow-origin': '*', 'cache-control': 'public, max-age=3600' } }); } + catch { return json({ error: 'translation unavailable' }, 503); } + } + if (url.pathname === '/translate' && req.method === 'POST') { + if (!rateOk(ip, 60)) return json({ error: 'rate limited' }, 429); + let body = null; + try { body = await req.json(); } catch {} + if (!body || typeof body.text !== 'string' || typeof body.target !== 'string') return json({ error: 'bad request' }, 400); + try { return json(await translate(body.text, body.target, body.source)); } + catch (e) { return json({ error: e.message || 'translation failed' }, e.status || 503); } + } + if (url.pathname === '/health') { return json({ ok: true, domain: DOMAIN, names: Object.keys(state.names).length, diff --git a/names/translate.js b/names/translate.js new file mode 100644 index 0000000..7d6181d --- /dev/null +++ b/names/translate.js @@ -0,0 +1,82 @@ +// Post translation for coinos v3: a thin, cached front for a LibreTranslate +// container on the same docker network. Notes are public, so nothing private +// passes through here; the cache keys on the text itself, so a popular post +// is translated once whoever asks. +import { createHash } from 'node:crypto'; +import { splitProse, hasWords, translatePieces } from '../src/translate-split.js'; + +const LT = process.env.LT_URL || 'http://libretranslate:5000'; +const MAX_CHARS = 5000; +const CACHE_MAX = 20000; +const BUSY_MAX = 4; // LibreTranslate is CPU-bound: a few at once, the rest wait their turn +const cache = new Map(); // sha256(target|text) -> { text, source } +let busy = 0; +const waiting = []; +let langs = null, langsAt = 0; + +async function slot() { + if (busy < BUSY_MAX) { busy++; return; } + await new Promise((r) => waiting.push(r)); + busy++; +} +function release() { busy--; const next = waiting.shift(); if (next) next(); } + +// The languages the container has models for: { code: [targets] } +export async function languages() { + if (langs && Date.now() - langsAt < 3600_000) return langs; + const r = await fetch(LT + '/languages', { signal: AbortSignal.timeout(5000) }); + if (!r.ok) throw new Error('languages ' + r.status); + const list = await r.json(); + langs = Object.fromEntries(list.map((l) => [l.code, l.targets || []])); + if (langs['zh-Hans']) langs.zh = langs['zh-Hans']; // the app says 'zh' + langsAt = Date.now(); + return langs; +} + +// The app's locale codes ↔ LibreTranslate's ('zh' is Simplified there). +const norm = (c) => String(c || '').toLowerCase().split(/[-_]/)[0]; +const toLt = (c) => (c === 'zh' ? 'zh-Hans' : c); + +export async function translate(text, target, source = 'auto') { + text = String(text || ''); + target = norm(target); + source = source && source !== 'auto' ? norm(source) : 'auto'; + if (!text.trim()) return { text: '', source }; + if (text.length > MAX_CHARS) throw Object.assign(new Error('too long'), { status: 413 }); + const have = await languages(); + if (!have[toLt(target)]) throw Object.assign(new Error('unsupported language'), { status: 422 }); + const key = createHash('sha256').update(target + '|' + source + '|' + text).digest('hex'); + const hit = cache.get(key); + if (hit) { cache.delete(key); cache.set(key, hit); return hit; } + const prose = splitProse(text).filter((p) => p.prose && hasWords(p.prose)); + if (!prose.length) return { text, source: source === 'auto' ? null : source }; + await slot(); + try { + const call = async (path, body) => { + const r = await fetch(LT + path, { + method: 'POST', headers: { 'content-type': 'application/json' }, + body: JSON.stringify(body), signal: AbortSignal.timeout(30_000), + }); + const out = await r.json().catch(() => ({})); + if (!r.ok) throw Object.assign(new Error(out.error || path + ' ' + r.status), { status: r.status === 400 ? 422 : 502 }); + return out; + }; + let from = source, confidence = null; + if (from === 'auto') { + const d = await call('/detect', { q: prose.map((p) => p.prose).join('\n') }); + from = norm(d[0]?.language) || null; + confidence = d[0]?.confidence ?? null; + if (!from) throw Object.assign(new Error('unknown language'), { status: 422 }); + } + if (from === target) return { text, source: from, confidence }; + if (!have[toLt(from)]) throw Object.assign(new Error('unsupported language'), { status: 422 }); + const translated = await translatePieces(text, async (qs) => { + const r = await call('/translate', { q: qs, source: toLt(from), target: toLt(target), format: 'text' }); + return Array.isArray(r.translatedText) ? r.translatedText : [r.translatedText]; + }); + const out = { text: translated, source: from, confidence }; + cache.set(key, out); + if (cache.size > CACHE_MAX) cache.delete(cache.keys().next().value); + return out; + } finally { release(); } +} |