Add searchable lyrics index

server/db.js

@@ -20,6 +20,7 @@ const DB_PATH = process.env.DB_PATH || join(process.cwd(), 'data', 'ytplayer.db'
 export const db = createClient({
   url: 'file:' + DB_PATH,
 });
+let lyricsFtsAvailable = false;
 
 // ---- Schema ----------------------------------------------------------------
 
@@ -279,6 +280,7 @@ export async function initDb() {
     CREATE INDEX IF NOT EXISTS idx_audit_actor_action ON audit_log (actor, action, ts DESC);
   `);
   await migrateVideoMeta();
+  await initLyricsSearch();
   try { await db.execute('ALTER TABLE media_cache ADD COLUMN pinned INTEGER NOT NULL DEFAULT 0'); } catch { /* already migrated */ }
   // Uploads shared from devices (PWA share target): who sent it, and whether it
   // is listed in everyone's search (admin uploads) or reachable only by link.
@@ -307,6 +309,75 @@ export async function initDb() {
   }
 }
 
+function lyricsSearchContent(value) {
+  try {
+    const data = typeof value === 'string' ? JSON.parse(value) : value;
+    return [...(data?.lines || []).flatMap(line => [line?.text, line?.secondary, line?.phonetic]), ...(data?.tags || [])]
+      .filter(Boolean).join(' ').slice(0, 500_000);
+  } catch { return ''; }
+}
+
+async function initLyricsSearch() {
+  await db.execute('CREATE TABLE IF NOT EXISTS lyrics_search_state (id INTEGER PRIMARY KEY CHECK(id=1), indexed_at INTEGER NOT NULL)');
+  try {
+    await db.execute("CREATE VIRTUAL TABLE IF NOT EXISTS lyrics_search_fts USING fts5(video_id UNINDEXED, content, tokenize='unicode61 remove_diacritics 2')");
+    lyricsFtsAvailable = true;
+  } catch {
+    await db.execute('CREATE TABLE IF NOT EXISTS lyrics_search_fts (video_id TEXT PRIMARY KEY, content TEXT NOT NULL)');
+    await db.execute('CREATE INDEX IF NOT EXISTS idx_lyrics_search_id ON lyrics_search_fts(video_id)');
+    lyricsFtsAvailable = false;
+  }
+  const state = await db.execute('SELECT id FROM lyrics_search_state WHERE id=1');
+  if (state.rows.length) return;
+  await db.execute('DELETE FROM lyrics_search_fts');
+  const rows = await db.execute("SELECT video_id,data FROM video_notes WHERE kind='lyrics'");
+  for (let i = 0; i < rows.rows.length; i += 100) {
+    const batch = rows.rows.slice(i, i + 100).map(row => ({
+      sql: lyricsFtsAvailable ? 'INSERT INTO lyrics_search_fts(video_id,content) VALUES(?,?)' : 'INSERT OR REPLACE INTO lyrics_search_fts(video_id,content) VALUES(?,?)',
+      args: [row.video_id, lyricsSearchContent(row.data)],
+    }));
+    if (batch.length) await db.batch(batch, 'write');
+  }
+  await db.execute('INSERT INTO lyrics_search_state(id,indexed_at) VALUES(1,unixepoch())');
+}
+
+export async function searchNoteLyrics(query, { limit = 30, offset = 0, hasChapters = false, withoutChapters = false } = {}) {
+  const words = String(query || '').normalize('NFD').replace(/[\u0300-\u036f]/g, '').match(/[\p{L}\p{N}_-]+/gu) || [];
+  if (!words.length) return [];
+  const boundedLimit = Math.max(1, Math.min(100, Number(limit) || 30));
+  const boundedOffset = Math.max(0, Math.min(1_000_000, Number(offset) || 0));
+  const chapterFilter = hasChapters ? "AND EXISTS (SELECT 1 FROM video_notes chapter WHERE chapter.video_id=lyrics_search_fts.video_id AND chapter.kind='chapters')" : withoutChapters ? "AND NOT EXISTS (SELECT 1 FROM video_notes chapter WHERE chapter.video_id=lyrics_search_fts.video_id AND chapter.kind='chapters')" : '';
+  if (lyricsFtsAvailable) {
+    const match = words.map(word => `"${word.replaceAll('"', '""')}"*`).join(' AND ');
+    const result = await db.execute({ sql: `SELECT video_id, snippet(lyrics_search_fts,1,char(1),char(2),' … ',18) AS snippet
+      FROM lyrics_search_fts WHERE lyrics_search_fts MATCH ? ${chapterFilter} ORDER BY bm25(lyrics_search_fts) LIMIT ? OFFSET ?`, args: [match, boundedLimit, boundedOffset] });
+    return result.rows.map(row => ({ videoId: row.video_id, snippet: row.snippet }));
+  }
+  const filters = words.map(() => 'content LIKE ?').join(' AND ');
+  const result = await db.execute({ sql: `SELECT video_id,substr(content,1,220) AS snippet FROM lyrics_search_fts WHERE ${filters} ORDER BY video_id LIMIT ? OFFSET ?`, args: [...words.map(word => `%${word}%`), boundedLimit, boundedOffset] });
+  return result.rows.map(row => ({ videoId: row.video_id, snippet: row.snippet }));
+}
+
+export async function countNoteLyrics(query, { hasChapters = false, withoutChapters = false } = {}) {
+  const words = String(query || '').normalize('NFD').replace(/[\u0300-\u036f]/g, '').match(/[\p{L}\p{N}_-]+/gu) || [];
+  if (!words.length) return 0;
+  const chapterFilter = hasChapters ? "AND EXISTS (SELECT 1 FROM video_notes chapter WHERE chapter.video_id=lyrics_search_fts.video_id AND chapter.kind='chapters')" : withoutChapters ? "AND NOT EXISTS (SELECT 1 FROM video_notes chapter WHERE chapter.video_id=lyrics_search_fts.video_id AND chapter.kind='chapters')" : '';
+  if (lyricsFtsAvailable) {
+    const match = words.map(word => `"${word.replaceAll('"', '""')}"*`).join(' AND ');
+    return Number((await db.execute({ sql: `SELECT COUNT(*) AS n FROM lyrics_search_fts WHERE lyrics_search_fts MATCH ? ${chapterFilter}`, args: [match] })).rows[0]?.n) || 0;
+  }
+  const filters = words.map(() => 'content LIKE ?').join(' AND ');
+  const result = await db.execute({ sql: `SELECT COUNT(*) AS n FROM lyrics_search_fts WHERE ${filters} ${chapterFilter}`, args: words.map(word => `%${word}%`) });
+  return Number(result.rows[0]?.n) || 0;
+}
+
+export async function rebuildLyricsSearch() {
+  await db.execute('DELETE FROM lyrics_search_fts');
+  await db.execute('DELETE FROM lyrics_search_state');
+  await initLyricsSearch();
+  return Number((await db.execute("SELECT COUNT(*) AS n FROM video_notes WHERE kind='lyrics'")).rows[0]?.n) || 0;
+}
+
 const YOUTUBE_ID = /^[\w-]{11}$/;
 const folded = value => String(value || '').normalize('NFD').replace(/[\u0300-\u036f]/g, '').toLowerCase();
 
@@ -395,6 +466,11 @@ export async function saveNote({ videoId, kind, data, baseRev, source, updatedBy
             VALUES (?, ?, ?, ?, ?, ?, unixepoch())`,
       args: [videoId, kind, rev, json, source, updatedBy || null],
     });
+    if (kind === 'lyrics') {
+      await tx.execute({ sql: 'DELETE FROM lyrics_search_fts WHERE video_id=?', args: [videoId] });
+      const content = lyricsSearchContent(data);
+      if (content) await tx.execute({ sql: 'INSERT INTO lyrics_search_fts(video_id,content) VALUES(?,?)', args: [videoId, content] });
+    }
     await tx.commit();
     return { ok: true, rev };
   } catch (err) {