diff --git a/.agents/skills/lyrics-lookup/SKILL.md b/.agents/skills/lyrics-lookup/SKILL.md new file mode 100644 index 0000000..17b4f6a --- /dev/null +++ b/.agents/skills/lyrics-lookup/SKILL.md @@ -0,0 +1,93 @@ +--- +name: lyrics-lookup +description: Find and inject lyrics for songs saved on worship.hesed.sbs that have none — LRCLIB first (free, often synced), then an agy web search (untimed), then local faster-whisper transcription. Use when the user asks to add/fill/inject lyrics, says a song has no lyrics, or asks to transcribe saved songs. For replacing lyrics that are already there but wrong, use lyrics-regenerate instead. +--- + +# Add lyrics to songs that don't have any + +Shared lyrics live on the server (`video_notes`, kind `lyrics`) and every user of +worship.hesed.sbs sees the same ones. Every save keeps the previous version as a +revision (`video_note_revs`), so nothing is destroyed — but **never overwrite +existing lyrics from this skill**; that is `lyrics-regenerate`'s job, and it takes +a backup first. + +## Sources, cheapest and best first + +| Order | Source | Timed? | Cost | Where | +|---|---|---|---|---| +| 1 | **LRCLIB** | usually synced | free, no key | server-side: `POST /api/notes//lyrics/web` | +| 2 | **agy web search** (Genius/AZLyrics/hymnary…) | no — untimed | flat-rate agy | `web_lyrics.py --agy` | +| 3 | **faster-whisper** on this machine | synced (word timings) | free, ~0.65× real time | `auto_lyrics.py` | + +Never reach for ElevenLabs Scribe or any credit-billed ASR — the user asked for +free transcription only. + +## Credentials + +Writing needs admin. Either env var works for every script: + +```bash +export YTP_ADMIN_PASSWORD='…' # the /admin password, see the ytplayer-admin-password memory +export YTP_TOKEN='ytp_…' # or an API token minted at /admin +``` + +The password is also vault secret `YTPLAYER_ADMIN_PASSWORD` +(`~/development/.secrets/ytplayer-admin.env`) — use it through +`vault__secret_exec` when you must not print it. + +## Run it + +```bash +cd ~/development/personal/ytplayer + +# 1) LRCLIB for everything that has no lyrics, then agy for the leftovers +YTP_ADMIN_PASSWORD=… python3 scripts/lyrics/web_lyrics.py --missing --agy + +# just look, change nothing +YTP_ADMIN_PASSWORD=… python3 scripts/lyrics/web_lyrics.py --missing --dry-run + +# 2) transcribe what the web doesn't have (needs the venv below) +YTP_ADMIN_PASSWORD=… ~/.local/share/lyrics-asr/.venv/bin/python \ + scripts/lyrics/auto_lyrics.py --missing +``` + +`--missing` lists songs from `GET /api/admin/media` and keeps only the ones with +`lyricsLines == 0`. Use `--ids A,B,C` to aim at specific videos. Ids are YouTube +ids (11 chars) or uploads (`upl_<12 hex>`). + +Whisper venv, once: + +```bash +uv venv ~/.local/share/lyrics-asr/.venv --python 3.12 +~/.local/share/lyrics-asr/.venv/bin/pip install -r scripts/lyrics/requirements.txt +``` + +The same transcriber also runs unattended as the `lyrics-worker` container in +`docker-compose.yml` (`auto_lyrics.py --watch 300 --state …`), which is what gives +newly downloaded songs lyrics "at their own pace". It authenticates with +`LYRICS_WORKER_TOKEN`. + +## What lands in the database + +```jsonc +{ "lines": [{ "t": 12.4, "text": "Holy You are", "kind": "line" }], // t null = untimed + "tags": ["from LRCLIB (synced)"], // provenance — always tag + "offset": 0 } +``` + +Tags in use: `from LRCLIB (synced)` / `(plain text)`, `from the web (untimed) — check and Tap-sync`, +`auto-transcribed`, `from the file (synced)` (embedded in an admin upload). +The tag is how later runs tell machine lyrics from published ones — keep it accurate. + +## Things that bite + +- **LRCLIB 503/429** on bursts. `lrclib_regen.py:http_json` already retries with + backoff; if you write new code against LRCLIB, copy it. Keep ~0.4 s between calls. +- **Wrong-artist matches.** A common title ("Still") matches another genre's song. + `artist_ok()` in `lrclib_regen.py` is the verification — port it rather than + trusting a title+duration hit. +- **Karaoke / minus-one tracks have no vocals.** Whisper returns noise; the script + calls them instrumental below `--min-words 25` and skips them. That's correct. +- **Untimed lyrics are fine.** The app shows them as a plain scrolling list and the + user can Tap-sync them in the admin lyric editor. +- **Never redistribute.** These are third-party lyrics in a private library. diff --git a/.agents/skills/lyrics-regenerate/SKILL.md b/.agents/skills/lyrics-regenerate/SKILL.md new file mode 100644 index 0000000..0ee6859 --- /dev/null +++ b/.agents/skills/lyrics-regenerate/SKILL.md @@ -0,0 +1,98 @@ +--- +name: lyrics-regenerate +description: Back up and replace lyrics on worship.hesed.sbs that are wrong, mistimed or misheard — typically machine transcripts — with published lyrics from LRCLIB. Use when the user says lyrics are incorrect/off/mistimed, names songs whose lyrics are bad, or asks to re-check songs on LRCLIB. Always backs up first and never writes an unverified artist match. +--- + +# Replace bad lyrics from LRCLIB + +Machine transcripts (Whisper, Scribe) mishear words and drift out of time. +LRCLIB's published lyrics are usually correct and often **synced**. This skill +swaps them in — backup first, artist verified, one song at a time. + +Script: `scripts/lyrics/lrclib_regen.py`. + +## The rule: back up before you touch anything + +The server keeps every previous version as a revision, and `/admin` can restore +one, but the script **also** writes an offline JSON backup of the current lyrics +of every song it will consider — on a dry run too. Do not skip it, do not write +your own one-off loop that lacks it. + +``` +/mnt/c/Users/josh/Documents/ytplayer-lyrics-backup-.json +``` + +(`--backup-dir` or `YTP_BACKUP_DIR` to move it; on the devbox it falls back to `~/`.) + +## Run it + +```bash +cd ~/development/personal/ytplayer +export YTP_ADMIN_PASSWORD='…' # or YTP_TOKEN=ytp_… + +# 1) ALWAYS dry-run first and read every line of the output +python3 scripts/lyrics/lrclib_regen.py --tagged auto-transcribed + +# 2) apply once the matches look right +python3 scripts/lyrics/lrclib_regen.py --tagged auto-transcribed --apply +``` + +Picking the songs: + +| Flag | Picks | +|---|---| +| `--ids A,B,C` | exactly those videos | +| `--tagged auto-transcribed` | songs whose lyrics carry that tag (machine transcripts) | +| `--since '2026-09-19 01:45' --until '2026-09-19 02:00'` | songs whose lyrics were **saved** in that window | +| *(none)* | every song the server has lyrics for | + +`--since/--until` is the one to reach for when the user says *"the songs that got +lyrics at the same time as X"* — read X's `updatedAt` from +`GET /api/notes/` and bracket it by a few minutes. + +Other flags: `--tolerance 6` (max duration difference, seconds), `--loose` +(accept matches whose artist doesn't line up — risky, see below). + +## Reading the output + +``` +eJBlOV6cM7Y LRCLIB synced | Israel Houghton – Holy You Are | 41 lines (was 38) +_n6dfB2Z-Ko UNSURE plain | Night Ranger – Still | 52 lines (was 44) + ↳ artist doesn't match "Hillsong Worship" / the video title — left alone (use --loose to accept) +tYM05iaVu3I no match | Jesus At The Centre | … | keeping 60 lines (auto-transcribed) +``` + +- **LRCLIB** — verified match, will be written on `--apply`. +- **UNSURE** — title and duration fit but the artist doesn't appear in the channel + name or the video title. **Left alone by default. Do not pass `--loose` to make + it go away** — check the song by hand instead; this guard is what stopped a + Hillsong song being overwritten with a Night Ranger one. +- **no match** — LRCLIB doesn't have it. Existing lyrics are kept. Fall back to + `web_lyrics.py --agy --ids --overwrite`, or fix it in the admin lyric editor. + +Replaced songs are tagged `from LRCLIB (synced)` / `(plain text)`, which is also +how you tell later what has already been fixed. + +## Restoring + +- Per song, in the UI: `/admin` → recent edits → **Restore** on the older revision. +- From the JSON backup: `PUT /api/notes//lyrics` with + `{"data": , "baseRev": >}`. + Use the *current* rev, not the backed-up one — `baseRev` is optimistic + concurrency, not a version to travel back to. + +## Gotchas + +- **LRCLIB rate-limits**: 503/429 on bursts. `http_json()` retries with backoff and + the loop sleeps 0.4 s between songs. A song that fails all retries is reported and + skipped — re-run it later rather than hammering. +- **Duration match is ±6 s** against `/api/streams` metadata. Live or extended cuts + legitimately miss; raise `--tolerance` deliberately, per song. +- **Title cleaning** strips "(Official Video)", "Lyrics", "[HD]" etc. + `clean_title`'s `NOISE` regex is used with `.sub()` and `.search()` — never give + it a `/g`-style shared match state; a stateful regex silently skipped every other + song once already. +- **Karaoke/minus-one** versions match the original recording's lyrics, which is + usually what you want, but the timing won't line up. Check before applying. + +Related skills: `lyrics-lookup` (songs with **no** lyrics), `deploy-prod`. diff --git a/.claude/skills/lyrics-lookup b/.claude/skills/lyrics-lookup new file mode 120000 index 0000000..8ee59e5 --- /dev/null +++ b/.claude/skills/lyrics-lookup @@ -0,0 +1 @@ +../../.agents/skills/lyrics-lookup \ No newline at end of file diff --git a/.claude/skills/lyrics-regenerate b/.claude/skills/lyrics-regenerate new file mode 120000 index 0000000..b468a72 --- /dev/null +++ b/.claude/skills/lyrics-regenerate @@ -0,0 +1 @@ +../../.agents/skills/lyrics-regenerate \ No newline at end of file diff --git a/.gitignore b/.gitignore index 718e347..83539cc 100755 --- a/.gitignore +++ b/.gitignore @@ -34,3 +34,6 @@ server/public # Old Tauri/Windows desktop build outputs — published artifacts, not source. # They are still in git history; keeping them out of the tip keeps deploy clones small. legacy/releases/ + +# python caches from scripts/lyrics +__pycache__/ diff --git a/frontend/admin.html b/frontend/admin.html index 0777a34..551eece 100644 --- a/frontend/admin.html +++ b/frontend/admin.html @@ -53,6 +53,28 @@ .hidden { display: none !important; } .src { font-family: var(--mono); font-size: 11px; color: var(--text-dim); } .src.api, .src.auto { color: var(--warn); } .src.restore { color: var(--ok); } + + /* --- lyric editor --- */ + #leWrap { margin-top: 12px; } + #leWave { width: 100%; height: 72px; display: block; border: 1px solid var(--line); border-radius: 9px; background: var(--bg-1); cursor: pointer; touch-action: none; } + .le-clock { font-family: var(--mono); font-size: 18px; min-width: 96px; text-align: center; } + .le-clock b { color: var(--accent); } + .le-lines { max-height: 56vh; overflow-y: auto; border: 1px solid var(--line); border-radius: 9px; background: var(--bg-1); } + .le-line { display: flex; gap: 6px; align-items: center; padding: 4px 6px; border-bottom: 1px solid var(--line); } + .le-line:last-child { border-bottom: 0; } + .le-line.now { background: color-mix(in srgb, var(--accent) 16%, transparent); } + .le-line.next { background: color-mix(in srgb, var(--accent) 7%, transparent); } + .le-line.untimed .le-t { color: var(--warn); } + .le-t { font-family: var(--mono); font-size: 12px; width: 74px; flex: none; background: var(--bg-2); color: var(--text); border: 1px solid var(--line); border-radius: 6px; padding: 5px 4px; text-align: center; } + .le-line input.le-text { flex: 1 1 auto; min-width: 80px; background: transparent; border: 0; color: var(--text); font-size: 14px; padding: 5px 4px; } + .le-line input.le-text:focus { outline: 1px solid var(--accent); border-radius: 6px; } + .le-line.section input.le-text { font-weight: 700; color: var(--text-2); } + .le-line.cue input.le-text { font-style: italic; color: var(--warn); } + .le-mini { padding: 3px 6px; font-size: 12px; line-height: 1.1; flex: none; } + .le-kind { background: var(--bg-2); color: var(--text-2); border: 1px solid var(--line); border-radius: 6px; font-size: 11px; padding: 4px; flex: none; } + #leRaw { width: 100%; min-height: 40vh; background: var(--bg-1); color: var(--text); border: 1px solid var(--line); border-radius: 9px; padding: 10px; font-family: var(--mono); font-size: 12px; } + .pill { font-family: var(--mono); font-size: 11px; border: 1px solid var(--line); border-radius: 999px; padding: 2px 8px; color: var(--text-dim); } + .pill.on { color: var(--accent); border-color: var(--accent); } @@ -101,6 +123,56 @@
+
+
+

Lyric editor

+ + no song loaded +
+

Play the song here and time every line against it: Set stamps the playhead onto a line, Tap mode walks down the lines as you tap (Space), and the waveform scrubs. Saving writes a new shared revision — the old one stays restorable below.

+
+ + + +
+
+ + +
+

Upload video / audio

Uploads are served by this server and appear in search and in the app's 📁 Library. Title, artist, duration, cover art and embedded lyrics are read from the file; a separate cover image and manual title/artist override them.

@@ -154,8 +226,9 @@ function videoIdFrom(input) { const s = String(input || '').trim(); - if (/^[A-Za-z0-9_-]{11}$/.test(s)) return s; - const m = s.match(/(?:v=|youtu\.be\/|shorts\/|embed\/|live\/)([A-Za-z0-9_-]{11})/); + if (/^[A-Za-z0-9_-]{11}$/.test(s) || /^upl_[a-f0-9]{12}$/.test(s)) return s; + const m = s.match(/(?:v=|youtu\.be\/|shorts\/|embed\/|live\/)([A-Za-z0-9_-]{11})/) + || s.match(/(upl_[a-f0-9]{12})/); return m ? m[1] : null; } @@ -176,6 +249,8 @@ loadTokens(); loadRevs(); loadUploads(); + const want = videoIdFrom(new URLSearchParams(location.search).get('v') || ''); + if (want) leOpen(want); } $('loginForm').addEventListener('submit', async (e) => { @@ -257,6 +332,7 @@ ${esc(r.source)} + ${r.kind === 'lyrics' ? `` : ''} `).join(''); @@ -279,6 +355,7 @@ $('revTable').addEventListener('click', async (e) => { const v = e.target.dataset.view; const r = e.target.dataset.restore; + if (e.target.dataset.edit) { leOpen(e.target.dataset.edit); return; } if (v) { const [id, kind, rev] = v.split('|'); const j = await api(`/api/notes/${id}/${kind}/revs/${rev}`); @@ -306,6 +383,7 @@ ${Math.round(u.duration)}s · ${(u.size / 1048576).toFixed(1)} MB ${esc(u.id)} Play + `).join(''); $('upTable').innerHTML = rows @@ -347,6 +425,7 @@ }); $('upTable').addEventListener('click', async (e) => { + if (e.target.dataset.lyrics) { leOpen(e.target.dataset.lyrics); return; } const id = e.target.dataset.del; if (!id || !confirm('Delete this upload? The file is removed from the server.')) return; const j = await api(`/api/admin/uploads/${encodeURIComponent(id)}`, { method: 'DELETE' }); @@ -354,6 +433,375 @@ loadUploads(); }); + /* ======================================================================== + * Lyric editor — the song plays here, so every line can be timed against + * what you actually hear. The rows are built ONCE per load and then only + * their classes/values change: rebuilding them would steal focus from the + * field being typed in and reset the scroll position. + * ====================================================================== */ + const LE = { + id: '', rev: 0, lines: [], tags: [], offset: 0, dur: 0, + audio: null, peaks: null, rows: [], tap: false, dirty: false, raw: false, + }; + + const clamp2 = (n) => Math.round(n * 100) / 100; + function fmtT(t) { + if (t == null) return ''; + const m = Math.floor(t / 60); + return `${m}:${(t - m * 60).toFixed(2).padStart(5, '0')}`; + } + function parseT(s) { + const v = String(s || '').trim(); + if (!v) return null; + const m = v.match(/^(\d+):(\d{1,2}(?:\.\d+)?)$/); + const n = m ? Number(m[1]) * 60 + Number(m[2]) : Number(v); + return Number.isFinite(n) && n >= 0 ? clamp2(n) : null; + } + + function leDirty(on) { + LE.dirty = on; + $('leDirty').textContent = on ? 'unsaved changes' : ''; + $('leDirty').className = on ? 'muted' : 'muted'; + } + function leSay(text, cls) { const m = $('leMsg'); m.className = 'msg' + (cls ? ' ' + cls : ''); m.textContent = text; } + + async function leLoad(id, keepAudio) { + leSay('Loading…'); + const [notes, streams] = await Promise.all([ + api(`/api/notes/${id}`), + fetch(`/api/streams?v=${encodeURIComponent(id)}`, { credentials: 'same-origin' }).then((r) => r.json()).catch(() => ({})), + ]); + const doc = (notes.lyrics && notes.lyrics.data) || { lines: [], tags: [], offset: 0 }; + LE.id = id; + LE.rev = (notes.lyrics && notes.lyrics.rev) || 0; + LE.lines = (doc.lines || []).map((l) => ({ t: l.t == null ? null : clamp2(l.t), text: String(l.text || ''), kind: l.kind || 'line' })); + LE.tags = doc.tags || []; + LE.offset = Number(doc.offset) || 0; + const meta = (streams.data && streams.data.meta) || {}; + LE.dur = Number(meta.duration) || 0; + $('leTitle').textContent = meta.title || id; + $('leMeta').textContent = ` ${meta.channel || ''} · ${LE.lines.length} lines · rev ${LE.rev}`; + $('leTags').value = LE.tags.join(', '); + $('leState').textContent = LE.lines.length ? `${LE.lines.filter((l) => l.t != null).length}/${LE.lines.length} timed` : 'no lyrics yet'; + $('leState').className = 'pill' + (LE.lines.length && LE.lines.every((l) => l.t != null) ? ' on' : ''); + $('leWrap').classList.remove('hidden'); + leRender(); + leDirty(false); + leSay(`Loaded ${id}`, 'ok'); + + if (!keepAudio) { + const url = (streams.data && streams.data.audioUrl) + || ((streams.data && streams.data.qualities || []).find((q) => q.hasAudio) || {}).url || ''; + if (LE.audio) { LE.audio.pause(); LE.audio.src = ''; } + LE.audio = null; + if (url) { + const a = new Audio(url); + a.preload = 'auto'; + a.addEventListener('timeupdate', leTick); + a.addEventListener('seeked', leTick); + a.addEventListener('loadedmetadata', () => { if (a.duration) { LE.dur = a.duration; $('leDur').textContent = fmtT(a.duration).slice(0, -3); } leDraw(); }); + a.addEventListener('play', () => { $('lePlay').textContent = '❚❚ Pause'; }); + a.addEventListener('pause', () => { $('lePlay').textContent = '▶︎ Play'; }); + a.addEventListener('error', () => leSay('No playable audio for this song on the server — you can still edit the text.', 'err')); + LE.audio = a; + } else { + leSay('This song is not cached on the server, so there is no audio to time against.', 'err'); + } + $('leDur').textContent = fmtT(LE.dur).slice(0, -3); + LE.peaks = null; + fetch(`/api/media/${id}/peaks`, { credentials: 'same-origin' }) + .then((r) => r.json()).then((j) => { if (j.ok) { LE.peaks = j.peaks; if (j.duration) LE.dur = j.duration; leDraw(); } }) + .catch(() => { /* peaks are a nicety */ }); + } + leDraw(); + } + + function leRow(line, i) { + const row = document.createElement('div'); + row.className = 'le-line' + (line.t == null ? ' untimed' : '') + (line.kind !== 'line' ? ' ' + line.kind : ''); + row.dataset.i = String(i); + row.innerHTML = ` + + + + + + + + + `; + row.querySelector('.le-kind').value = line.kind; + return row; + } + + function leRender() { + const box = $('leLines'); + box.innerHTML = ''; + LE.rows = LE.lines.map((l, i) => { const r = leRow(l, i); box.appendChild(r); return r; }); + if (!LE.lines.length) box.innerHTML = '
No lines yet — “+ Line”, paste them in ✎ Text, or pull them from LRCLIB.
'; + $('leMeta').textContent = $('leMeta').textContent.replace(/\d+ lines/, `${LE.lines.length} lines`); + } + + // Which line is sung at time t (doc offset included, as the app plays it). + function leIndexAt(t) { + let hit = -1; + for (let i = 0; i < LE.lines.length; i++) { + const lt = LE.lines[i].t; + if (lt != null && lt + LE.offset <= t) hit = i; else if (lt != null) break; + } + return hit; + } + + let leLastIdx = -2; + function leTick() { + const t = LE.audio ? LE.audio.currentTime : 0; + $('leCur').textContent = fmtT(t); + const i = LE.tap ? leTapIdx() : leIndexAt(t); + if (i !== leLastIdx) { + // Clear every row, not just the previous pair: in tap mode (and after a + // seek) the index jumps, and a stale highlight would be left behind. + for (const r of LE.rows) r.classList.remove('now', 'next'); + if (LE.rows[i]) LE.rows[i].classList.add('now'); + if (LE.rows[i + 1]) LE.rows[i + 1].classList.add('next'); + if ($('leFollow').checked && LE.rows[i]) LE.rows[i].scrollIntoView({ block: 'center', behavior: 'smooth' }); + leLastIdx = i; + } + leDraw(); + } + + function leDraw() { + const c = $('leWave'); + if (!c) return; + const w = c.clientWidth || 600, h = c.height; + const dpr = window.devicePixelRatio || 1; + if (c.width !== Math.round(w * dpr)) { c.width = Math.round(w * dpr); } + const g = c.getContext('2d'); + g.setTransform(dpr, 0, 0, dpr, 0, 0); + const css = getComputedStyle(document.body); + g.clearRect(0, 0, w, h); + const dur = LE.dur || 1; + const cur = LE.audio ? LE.audio.currentTime : 0; + if (LE.peaks && LE.peaks.length) { + const n = LE.peaks.length, bw = w / n; + for (let i = 0; i < n; i++) { + const bh = Math.max(1, (LE.peaks[i] / 100) * (h - 8)); + g.fillStyle = (i * dur) / n <= cur ? css.getPropertyValue('--accent') : css.getPropertyValue('--line'); + g.fillRect(i * bw, (h - bh) / 2, Math.max(1, bw - 0.5), bh); + } + } else { + g.fillStyle = css.getPropertyValue('--line'); + g.fillRect(0, h / 2 - 1, w, 2); + g.fillStyle = css.getPropertyValue('--accent'); + g.fillRect(0, h / 2 - 1, (cur / dur) * w, 2); + } + // line markers + g.fillStyle = css.getPropertyValue('--text-2'); + for (const l of LE.lines) { + if (l.t == null) continue; + const x = ((l.t + LE.offset) / dur) * w; + g.fillRect(x, 0, 1, 6); + g.fillRect(x, h - 6, 1, 6); + } + g.fillStyle = css.getPropertyValue('--text'); + g.fillRect((cur / dur) * w - 1, 0, 2, h); + } + + function leSeek(t) { + if (!LE.audio) return; + LE.audio.currentTime = Math.max(0, Math.min(LE.dur || 1e9, t)); + leTick(); + } + + // Tap mode: the next line without a time (or the one after the playhead). + function leTapIdx() { + const i = LE.lines.findIndex((l) => l.t == null); + if (i >= 0) return i; + return Math.min(LE.lines.length - 1, leIndexAt(LE.audio ? LE.audio.currentTime : 0) + 1); + } + function leStamp(i) { + if (!LE.audio || i < 0 || i >= LE.lines.length) return; + LE.lines[i].t = clamp2(LE.audio.currentTime); + const row = LE.rows[i]; + if (row) { row.querySelector('.le-t').value = fmtT(LE.lines[i].t); row.classList.remove('untimed'); } + leDirty(true); + leLastIdx = -2; + leTick(); + } + + $('leLoadForm').addEventListener('submit', (e) => { + e.preventDefault(); + const id = videoIdFrom($('leId').value); + if (!id) { leSay('Not a YouTube URL, video id or upl_ id', 'err'); return; } + if (LE.dirty && !confirm('Discard the unsaved changes to the song you have open?')) return; + leLoad(id); + }); + + $('leFromWeb').addEventListener('click', async () => { + const id = LE.id || videoIdFrom($('leId').value); + if (!id) { leSay('Load a song first', 'err'); return; } + leSay('Asking LRCLIB…'); + const j = await api(`/api/notes/${id}/lyrics/web`, { method: 'POST', body: { overwrite: true } }); + if (!j.ok) { leSay(j.error || 'LRCLIB has nothing for this song', 'err'); return; } + await leLoad(id, true); + leSay(`LRCLIB: ${j.lines} ${j.synced ? 'synced' : 'plain'} lines saved as rev ${j.rev} — the previous version is restorable below.`, 'ok'); + loadRevs(); + }); + + $('leLines').addEventListener('click', (e) => { + const btn = e.target.closest('button[data-act]'); + if (!btn) return; + const i = Number(btn.closest('.le-line').dataset.i); + const act = btn.dataset.act; + if (act === 'set') leStamp(i); + else if (act === 'seek') { leSeek((LE.lines[i].t || 0) + LE.offset); if (LE.audio && LE.audio.paused) LE.audio.play().catch(() => {}); } + else if (act === 'nudge') { + if (LE.lines[i].t == null) return; + LE.lines[i].t = clamp2(Math.max(0, LE.lines[i].t + Number(btn.dataset.d))); + btn.closest('.le-line').querySelector('.le-t').value = fmtT(LE.lines[i].t); + leDirty(true); leDraw(); + } else if (act === 'add') { + LE.lines.splice(i + 1, 0, { t: null, text: '', kind: 'line' }); + leRender(); leDirty(true); + const next = LE.rows[i + 1]; if (next) next.querySelector('.le-text').focus(); + } else if (act === 'del') { + LE.lines.splice(i, 1); + leRender(); leDirty(true); leDraw(); + } + }); + + $('leLines').addEventListener('input', (e) => { + const row = e.target.closest('.le-line'); + if (!row) return; + const i = Number(row.dataset.i); + if (e.target.classList.contains('le-text')) { LE.lines[i].text = e.target.value; leDirty(true); } + }); + $('leLines').addEventListener('change', (e) => { + const row = e.target.closest('.le-line'); + if (!row) return; + const i = Number(row.dataset.i); + if (e.target.classList.contains('le-t')) { + LE.lines[i].t = parseT(e.target.value); + e.target.value = fmtT(LE.lines[i].t); + row.classList.toggle('untimed', LE.lines[i].t == null); + leDirty(true); leDraw(); + } else if (e.target.classList.contains('le-kind')) { + LE.lines[i].kind = e.target.value; + row.classList.remove('section', 'cue'); + if (e.target.value !== 'line') row.classList.add(e.target.value); + leDirty(true); + } + }); + // Enter in a line = insert the next one below (writing lyrics by ear). + $('leLines').addEventListener('keydown', (e) => { + if (e.key !== 'Enter' || !e.target.classList.contains('le-text')) return; + e.preventDefault(); + const i = Number(e.target.closest('.le-line').dataset.i); + LE.lines.splice(i + 1, 0, { t: null, text: '', kind: LE.lines[i].kind === 'section' ? 'line' : LE.lines[i].kind }); + leRender(); leDirty(true); + if (LE.rows[i + 1]) LE.rows[i + 1].querySelector('.le-text').focus(); + }); + + $('lePlay').addEventListener('click', () => { + if (!LE.audio) return; + if (LE.audio.paused) LE.audio.play().catch((err) => leSay('Playback failed: ' + err.message, 'err')); + else LE.audio.pause(); + }); + $('leBack').addEventListener('click', () => leSeek((LE.audio ? LE.audio.currentTime : 0) - 5)); + $('leFwd').addEventListener('click', () => leSeek((LE.audio ? LE.audio.currentTime : 0) + 5)); + $('leRate').addEventListener('change', () => { if (LE.audio) LE.audio.playbackRate = Number($('leRate').value); }); + $('leTap').addEventListener('click', () => { + LE.tap = !LE.tap; + $('leTap').classList.toggle('primary', LE.tap); + leSay(LE.tap ? 'Tap mode: press Space (or S) on the beat to stamp the highlighted line; it then moves to the next one.' : '', LE.tap ? 'ok' : ''); + leLastIdx = -2; leTick(); + }); + $('leWave').addEventListener('pointerdown', (e) => { + const r = e.currentTarget.getBoundingClientRect(); + leSeek(((e.clientX - r.left) / r.width) * (LE.dur || 0)); + }); + window.addEventListener('resize', leDraw); + + document.addEventListener('keydown', (e) => { + if ($('leWrap').classList.contains('hidden')) return; + const typing = /^(INPUT|TEXTAREA|SELECT)$/.test(e.target.tagName); + if (e.code === 'Space' && LE.tap && !typing) { e.preventDefault(); leStamp(leTapIdx()); return; } + if ((e.key === 's' || e.key === 'S') && !typing && !e.metaKey && !e.ctrlKey) { e.preventDefault(); leStamp(LE.tap ? leTapIdx() : leIndexAt(LE.audio ? LE.audio.currentTime : 0)); return; } + if (e.key === 'k' && !typing) { e.preventDefault(); $('lePlay').click(); } + }); + + $('leShiftBtn').addEventListener('click', () => { + const d = Number($('leShift').value); + if (!Number.isFinite(d) || !d) return; + for (const l of LE.lines) if (l.t != null) l.t = clamp2(Math.max(0, l.t + d)); + leRender(); leDirty(true); leDraw(); + leSay(`Shifted every timed line by ${d > 0 ? '+' : ''}${d}s`, 'ok'); + }); + $('leAddBtn').addEventListener('click', () => { + LE.lines.push({ t: LE.audio ? clamp2(LE.audio.currentTime) : null, text: '', kind: 'line' }); + leRender(); leDirty(true); + const last = LE.rows[LE.rows.length - 1]; if (last) last.querySelector('.le-text').focus(); + }); + $('leTags').addEventListener('input', () => { LE.tags = $('leTags').value.split(',').map((s) => s.trim()).filter(Boolean); leDirty(true); }); + + // ✎ Text — the same LRC superset the app's own editor uses. + $('leRawBtn').addEventListener('click', () => { + LE.raw = !LE.raw; + $('leRaw').classList.toggle('hidden', !LE.raw); + $('leLines').classList.toggle('hidden', LE.raw); + $('leRawBtn').textContent = LE.raw ? '≡ Lines' : '✎ Text'; + if (LE.raw) { + $('leRaw').value = LE.lines.map((l) => (l.t == null ? '' : `[${fmtT(l.t)}] `) + (l.kind === 'section' ? '# ' : l.kind === 'cue' ? '! ' : '') + l.text).join('\n'); + } else { + LE.lines = $('leRaw').value.split('\n').map((raw) => { + let rest = raw.trim(), t = null; + const m = rest.match(/^\[(\d{1,3}):(\d{1,2}(?:[.:]\d{1,3})?)\]\s*/); + if (m) { t = clamp2(Number(m[1]) * 60 + Number(m[2].replace(':', '.'))); rest = rest.slice(m[0].length); } + let kind = 'line'; + if (rest.startsWith('# ')) { kind = 'section'; rest = rest.slice(2); } + else if (rest.startsWith('! ')) { kind = 'cue'; rest = rest.slice(2); } + return { t, text: rest, kind }; + }).filter((l) => l.text); + leRender(); leDirty(true); leDraw(); + } + }); + + $('leReload').addEventListener('click', () => { if (LE.id && confirm('Throw away the unsaved changes and reload from the server?')) leLoad(LE.id, true); }); + + $('leSave').addEventListener('click', async () => { + if (!LE.id) return; + const lines = LE.lines.filter((l) => l.text.trim()).map((l) => ({ t: l.t, text: l.text.trim(), kind: l.kind })); + if (!lines.length) { leSay('Nothing to save', 'err'); return; } + const data = { lines, tags: LE.tags, offset: LE.offset }; + leSay('Saving…'); + let j = await api(`/api/notes/${LE.id}/lyrics`, { method: 'PUT', body: { data, baseRev: LE.rev } }); + if (!j.ok && j.current) { + if (!confirm(`Someone else saved this song while you were editing (their rev ${j.current.rev}).\n\nOK = keep YOUR version, Cancel = load theirs.`)) { + await leLoad(LE.id, true); + return; + } + j = await api(`/api/notes/${LE.id}/lyrics`, { method: 'PUT', body: { data, force: true } }); + } + if (!j.ok) { leSay(j.error || 'save failed', 'err'); return; } + LE.rev = j.rev; + leDirty(false); + const timed = lines.filter((l) => l.t != null).length; + $('leState').textContent = `${timed}/${lines.length} timed`; + $('leState').className = 'pill' + (timed === lines.length ? ' on' : ''); + leSay(`Saved ${lines.length} lines as rev ${j.rev}.`, 'ok'); + loadRevs(); + }); + + window.addEventListener('beforeunload', (e) => { if (LE.dirty) { e.preventDefault(); e.returnValue = ''; } }); + + // "Edit" from the revisions feed and ?v= in the URL open a song here. + function leOpen(id) { + $('leId').value = id; + leLoad(id); + $('leWrap').scrollIntoView({ behavior: 'smooth', block: 'start' }); + } + function renderDocs() { const o = location.origin; $('apiDocs').textContent = `# Read (public) @@ -388,6 +836,8 @@ curl ${o}/api/uploads?q=worship # public list / search curl -X POST ${o}/api/notes/VIDEO_ID/lyrics/web \\ -H "Authorization: Bearer $YTP_TOKEN" -H "Content-Type: application/json" -d '{"overwrite": false}' +# Lyric editor deep link (admin): /admin?v=VIDEO_ID opens that song in the editor + # History (public) curl ${o}/api/notes/VIDEO_ID/lyrics/revs curl ${o}/api/notes/VIDEO_ID/lyrics/revs/3 diff --git a/frontend/styles.css b/frontend/styles.css index 73535bc..db4d21a 100755 --- a/frontend/styles.css +++ b/frontend/styles.css @@ -1530,6 +1530,20 @@ input[type="range"]::-webkit-slider-thumb:hover { transform: scale(1.25); } @media (max-width: 860px) { .app { grid-template-columns: 1fr; } + /* Keep the five playback buttons on one line at the top of the control + deck; everything else (📍, A-B, volume, transition, EQ, PiP, speed…) + flows into a tool row underneath. A control with no explicit order + defaults to 0 and would otherwise jump ahead of the transport group. */ + .btn-row { justify-content: center; } + .btn-row::before { content: ""; width: 100%; order: 6; height: 0; } + #shuffleBtn { order: 1; } + #prevBtn { order: 2; } + #playBtn { order: 3; } + #nextBtn { order: 4; } + #repeatBtn { order: 5; } + .btn-row > *:not(#shuffleBtn):not(#prevBtn):not(#playBtn):not(#nextBtn):not(#repeatBtn) { order: 7; } + .btn-row .spacer { display: none; } + .sidebar-toggle { display: inline-flex; } .sidebar { @@ -2051,6 +2065,19 @@ input[type="range"]::-webkit-slider-thumb:hover { transform: scale(1.25); } gap: 7px; } + /* The five playback buttons stay together on the first line: every OTHER + control is pushed into the tool row below. Without an explicit order a + button defaults to 0 and jumps ahead of the transport group. */ + .btn-row::before { content: ""; width: 100%; order: 6; height: 0; } + #shuffleBtn { order: 1; } + #prevBtn { order: 2; } + #playBtn { order: 3; } + #nextBtn { order: 4; } + #repeatBtn { order: 5; } + .btn-row > *:not(#shuffleBtn):not(#prevBtn):not(#playBtn):not(#nextBtn):not(#repeatBtn) { order: 7; } + .btn-row { justify-content: center; } + .btn-row .spacer { display: none; } + /* Now-playing meta: stack vertically */ .now-meta { flex-direction: column; @@ -2264,7 +2291,8 @@ input[type="range"]::-webkit-slider-thumb:hover { transform: scale(1.25); } /* Row B — secondary tools, uniform 44px targets (sizes only: colors and .active/.sleep-active states keep their class-based styling) */ - #abABtn, #abBBtn, #abClearBtn, #muteBtn, #loopBtn, #sleepTimerBtn, #fsBtn { + #abABtn, #abBBtn, #abClearBtn, #muteBtn, #loopBtn, #sleepTimerBtn, #fsBtn, + #rememberPosBtn, #transitionBtn, #eqBtn, #pipBtn { order: 7; min-width: 44px; height: 44px; diff --git a/scripts/lyrics/lrclib_regen.py b/scripts/lyrics/lrclib_regen.py new file mode 100644 index 0000000..73f048c --- /dev/null +++ b/scripts/lyrics/lrclib_regen.py @@ -0,0 +1,242 @@ +#!/usr/bin/env python3 +"""lrclib_regen.py — back up and regenerate lyrics on worship.hesed.sbs from LRCLIB. + +Machine transcripts (Whisper/Scribe) mishear words and drift; published lyrics +on LRCLIB are usually correct and often SYNCED. This tool: + + 1. writes a JSON backup of the CURRENT lyrics of every song it will touch + (the server also keeps each previous version as a revision — nothing is + ever lost, this file is just an offline copy), + 2. looks each song up on LRCLIB by cleaned title + artist + duration, + 3. replaces the lyrics only when a confident match is found. + +Song metadata (title / artist / duration) comes from the public +/api/streams endpoint, so this works without admin access; writing needs +YTP_ADMIN_PASSWORD or YTP_TOKEN. + + YTP_ADMIN_PASSWORD=… python3 scripts/lyrics/lrclib_regen.py --ids A,B --dry-run + YTP_ADMIN_PASSWORD=… python3 scripts/lyrics/lrclib_regen.py --since '2026-09-19 01:45' --until '2026-09-19 02:00' + YTP_ADMIN_PASSWORD=… python3 scripts/lyrics/lrclib_regen.py --tagged auto-transcribed --apply +""" +import argparse +import datetime as dt +import json +import os +import re +import sys +import time +import urllib.parse +import urllib.request + +sys.path.insert(0, os.path.dirname(os.path.abspath(__file__))) +from auto_lyrics import Api # noqa: E402 + +LRCLIB = 'https://lrclib.net/api' +UA = {'User-Agent': 'ytplayer-lyrics-regen (https://worship.hesed.sbs)'} +NOISE = re.compile(r'\b(official|video|audio|lyrics?|lyric|hd|hq|4k|live|mv|karaoke|minus[\s-]?one|instrumental|cover|remaster(ed)?|visualizer|performance|version)\b', re.I) + + +def clean_title(raw): + t = str(raw or '').split('|')[0] + t = re.sub(r'\([^)]*\)|\[[^\]]*\]', lambda m: ' ' if NOISE.search(m.group()) else m.group(), t) + t = NOISE.sub(' ', t) + t = re.sub(r'\(\s*\)|\[\s*\]', ' ', t) + return re.sub(r'\s{2,}', ' ', t).strip(' -–—,') + + +def clean_artist(raw): + a = re.sub(r'\s*-\s*Topic$', '', str(raw or ''), flags=re.I) + a = re.sub(r'VEVO$', '', a, flags=re.I) + return re.sub(r'\s{2,}', ' ', NOISE.sub(' ', a)).strip() + + +def parse_lrc(text): + out = [] + for raw in str(text or '').replace('\r', '').split('\n'): + rest = raw.strip() + if not rest or re.fullmatch(r'\[[a-z]+:[^\]]*\]', rest, re.I): + continue + stamps = [] + while True: + m = re.match(r'^\[(\d{1,3}):(\d{1,2})(?:[.:](\d{1,3}))?\]', rest) + if not m: + break + stamps.append(int(m.group(1)) * 60 + int(m.group(2)) + (float('0.' + m.group(3)) if m.group(3) else 0)) + rest = rest[m.end():].strip() + if not rest: + continue + if not stamps: + out.append({'t': None, 'text': rest[:300], 'kind': 'line'}) + for t in stamps: + out.append({'t': round(t, 2), 'text': rest[:300], 'kind': 'line'}) + if out and all(l['t'] is not None for l in out): + out.sort(key=lambda l: l['t']) + return out + + +def norm(s): + return re.sub(r'\s+', ' ', re.sub(r'[^a-z0-9 ]+', ' ', str(s or '').lower())).strip() + + +def http_json(url, tries=4): + """LRCLIB is free and rate-limits bursts with 503/429 — back off and retry.""" + for attempt in range(tries): + try: + with urllib.request.urlopen(urllib.request.Request(url, headers=UA), timeout=30) as r: + return json.loads(r.read() or b'null') + except urllib.error.HTTPError as e: + if e.code == 404: + return None + if e.code in (429, 503) and attempt < tries - 1: + time.sleep(2 * (attempt + 1)) + continue + raise + except Exception: + if attempt < tries - 1: + time.sleep(2 * (attempt + 1)) + continue + raise + return None + + +def artist_ok(lrc_artist, channel, video_title): + """True when the LRCLIB artist plausibly matches the video. + + Lyric-video channels ("Christian Lyrics", a person's name) carry no artist, + so the artist is also looked for in the video title. Songs whose title is + shared across genres ("Still") otherwise match the wrong recording.""" + a = set(norm(lrc_artist).split()) - {'the', 'and', 'of', 'band', 'music', 'worship', 'ministries'} + if not a: + return False + hay = set(norm(channel).split()) | set(norm(video_title).split()) + return bool(a & hay) + + +def lrclib_lookup(title, artist, duration, tolerance=6): + """Best LRCLIB entry for a song, or None. Prefers synced lyrics.""" + q = {'track_name': title, 'artist_name': artist or ''} + if duration: + q['duration'] = str(int(round(duration))) + hit = http_json(f'{LRCLIB}/get?' + urllib.parse.urlencode(q)) + if not hit: + results = http_json(f'{LRCLIB}/search?' + urllib.parse.urlencode({'q': f'{title} {artist}'.strip()})) or [] + want = norm(title) + scored = [] + for x in results: + if x.get('instrumental') or not (x.get('syncedLyrics') or x.get('plainLyrics')): + continue + t = norm(x.get('trackName')) + dd = abs((x.get('duration') or 0) - duration) if duration else 99 + title_hit = 2 if t == want else 1 if (want in t or t in want) else 0 + if not title_hit or (duration and dd > tolerance): + continue + scored.append((title_hit * 10 + (3 if x.get('syncedLyrics') else 0) - min(9, dd), x)) + if not scored: + return None + hit = max(scored, key=lambda p: p[0])[1] + lines = parse_lrc(hit.get('syncedLyrics') or '') + synced = bool(lines) + if not lines: + lines = parse_lrc(hit.get('plainLyrics') or '') + return {'lines': lines, 'synced': synced, 'hit': hit} if lines else None + + +def describe(hit, cur_lines): + h = hit['hit'] + return f'{"synced" if hit["synced"] else "plain "} | {h.get("artistName", "")[:22]:22} – {h.get("trackName", "")[:26]:26} | {len(hit["lines"])} lines (was {cur_lines})' + + +def main(): + ap = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter) + ap.add_argument('--base', default=os.environ.get('YTP_BASE', 'https://worship.hesed.sbs')) + ap.add_argument('--ids', help='comma-separated video ids') + ap.add_argument('--since', help='only songs whose lyrics were last saved after this local time (YYYY-MM-DD HH:MM)') + ap.add_argument('--until', help='…and before this one') + ap.add_argument('--tagged', help='only songs whose lyrics carry this tag, e.g. auto-transcribed') + ap.add_argument('--backup-dir', default=os.environ.get('YTP_BACKUP_DIR', ''), help='where to write the backup JSON (default: Windows Documents on WSL, else ~/)') + ap.add_argument('--apply', action='store_true', help='actually write the new lyrics (default: dry run)') + ap.add_argument('--tolerance', type=float, default=6, help='max duration difference in seconds') + ap.add_argument('--loose', action='store_true', help="accept matches whose artist doesn't line up with the video (risky)") + args = ap.parse_args() + + api = Api(args.base, os.environ.get('YTP_TOKEN'), os.environ.get('YTP_ADMIN_PASSWORD')) + + # Which songs? Explicit ids, or every song the server has lyrics for, + # filtered by when they were saved and/or their tag. + ids = [x.strip() for x in (args.ids or '').split(',') if x.strip()] + if not ids: + st, r = api.call('GET', '/api/admin/media') + if st != 200: + sys.exit(f'need --ids (listing needs the newer server build: {r.get("error")})') + ids = [m['id'] for m in r['media'] if m.get('lyricsLines')] + to_ts = lambda s: dt.datetime.strptime(s, '%Y-%m-%d %H:%M').timestamp() if s else None + since, until = to_ts(args.since), to_ts(args.until) + + picked, backup = [], {} + for vid in ids: + st, n = api.call('GET', f'/api/notes/{vid}') + cur = (n or {}).get('lyrics') if st == 200 else None + if not cur: + continue + when = cur['updatedAt'] + if (since and when < since) or (until and when > until): + continue + if args.tagged and not any(args.tagged.lower() in t.lower() for t in cur['data'].get('tags', [])): + continue + st, sres = api.call('GET', f'/api/streams?v={vid}') + meta = ((sres or {}).get('data') or {}).get('meta') or {} + picked.append({'id': vid, 'cur': cur, 'meta': meta}) + backup[vid] = {'savedAt': when, 'rev': cur['rev'], 'title': meta.get('title', ''), 'data': cur['data']} + + if not picked: + sys.exit('nothing matched') + + # 1) Backup first — always, even on a dry run. + out_dir = args.backup_dir or ('/mnt/c/Users/josh/Documents' if os.path.isdir('/mnt/c/Users/josh/Documents') else os.path.expanduser('~')) + stamp = dt.datetime.now().strftime('%Y%m%d-%H%M%S') + path = os.path.join(out_dir, f'ytplayer-lyrics-backup-{stamp}.json') + with open(path, 'w', encoding='utf-8') as f: + json.dump({'exportedAt': dt.datetime.now().isoformat(), 'base': args.base, 'songs': backup}, f, ensure_ascii=False, indent=1) + print(f'backup: {len(backup)} songs → {path}\n') + + # 2) Look each one up and (optionally) replace. + changed = skipped = 0 + for p in picked: + vid, meta, cur = p['id'], p['meta'], p['cur'] + title, artist = clean_title(meta.get('title')), clean_artist(meta.get('channel')) + dur = float(meta.get('duration') or 0) + try: + hit = lrclib_lookup(title, artist, dur, args.tolerance) + except Exception as e: # network hiccup — keep going + print(f'{vid} lookup failed: {e}') + continue + time.sleep(0.4) # be polite to a free service + if not hit: + skipped += 1 + print(f'{vid} no match | {title[:42]:42} | {artist[:20]:20} | keeping {len(cur["data"]["lines"])} lines ({",".join(cur["data"].get("tags") or [])[:24]})') + continue + h = hit['hit'] + sure = artist_ok(h.get('artistName'), meta.get('channel'), meta.get('title')) + mark = 'LRCLIB' if sure else 'UNSURE' + print(f'{vid} {mark} {describe(hit, len(cur["data"]["lines"]))}') + if not sure and not args.loose: + skipped += 1 + print(f' ↳ artist doesn\'t match "{meta.get("channel", "")}" / the video title — left alone (use --loose to accept)') + continue + if not args.apply: + continue + doc = { + 'lines': hit['lines'], + 'tags': [f'from LRCLIB ({"synced" if hit["synced"] else "plain text"})'], + 'offset': 0, + } + st, rr = api.call('PUT', f'/api/notes/{vid}/lyrics', {'data': doc, 'baseRev': cur['rev']}) + if st == 200: + changed += 1 + else: + print(f' write failed {st}: {rr.get("error")}') + print(f'\n{changed} replaced, {skipped} left alone' + ('' if args.apply else ' (dry run — add --apply)')) + + +if __name__ == '__main__': + main()