fix: optimize subtitle extraction with mkvextract batch mode, fix copyToWorkDir skip logic, target original library dir

LocalFlowPlugins/subtitles/extractSubtitlesAndOcr/1.0.0/index.js

@@ -8,11 +8,13 @@ const fs   = require('fs');
 /**
  * extractSubtitlesAndOcr
  * Extracts subtitle streams for ITA and ENG languages:
- *   - Text-based (subrip, srt, ass, ssa, webvtt): direct FFmpeg extraction
- *   - Bitmap-based (hdmv_pgs_subtitle, dvd_subtitle): FFmpeg .sup extract
- *     then SubtitleEdit CLI OCR -> .srt (with graceful fallback on failure)
+ *   - Text-based (subrip, srt, ass, ssa, webvtt):
+ *     Fast extraction via mkvextract (all tracks in one pass, seconds instead of minutes),
+ *     with single-pass FFmpeg (-vn -an) fallback.
+ *   - Bitmap-based (hdmv_pgs_subtitle, dvd_subtitle): .sup extract
+ *     then SubtitleEdit CLI OCR -> .srt (with graceful fallback on failure).
  *
- * Output files go next to the source file.
+ * Output files go to target directory (original library directory or next to source).
  * Naming: <baseName>.<lang>[.forced][.default][.sdh].srt
  */
 
@@ -22,8 +24,8 @@ const BITMAP_CODECS = new Set(['hdmv_pgs_subtitle', 'dvd_subtitle', 'pgssub', 'd
 const details = () => ({
   name: 'Extract Subtitles & OCR Fallback',
   description:
-    'Extracts ITA/ENG subtitle streams to SRT files next to the source. '
-    + 'Text subtitles are extracted directly via FFmpeg. '
+    'Extracts ITA/ENG subtitle streams to SRT files in the library directory. '
+    + 'Fast multi-track extraction via mkvextract / FFmpeg. '
     + 'Bitmap subtitles (PGS/VOBSUB) are OCR-d via SubtitleEdit CLI. '
     + 'OCR failures are logged as warnings without stopping the pipeline.',
   style: { borderColor: '#00BCD4' },
@@ -73,21 +75,25 @@ const buildSuffix = (lang, stream) => {
   if (stream.disposition && stream.disposition.forced)           parts.push('forced');
   if (stream.disposition && stream.disposition.default)          parts.push('default');
   if (stream.disposition && stream.disposition.hearing_impaired) parts.push('sdh');
-  // Also check title tag for SDH
   const title = ((stream.tags && stream.tags.title) || '').toLowerCase();
   if (!parts.includes('sdh') && title.includes('sdh'))           parts.push('sdh');
   return parts.join('.');
 };
 
-/**
- * Run FFmpeg via Tdarr's CLI helper.
- * Passes the full args object so cliUtils can access any field it needs
- * (inputFileObj, updateWorker, variables, etc.) without us having to predict them.
- */
-const runFFmpeg = async (spawnArgs, args) => {
+const getMkvExtract = (args) => {
+  let mkvExtract = 'mkvextract';
+  if (args.mkvpropeditPath) {
+    const dir       = path.dirname(args.mkvpropeditPath);
+    const candidate = path.join(dir, process.platform === 'win32' ? 'mkvextract.exe' : 'mkvextract');
+    if (fs.existsSync(candidate)) return candidate;
+  }
+  return mkvExtract;
+};
+
+const runCLI = async (cliBin, spawnArgs, args) => {
   const { CLI } = require('../../../../FlowHelpers/1.0.0/cliUtils');
   const cli = new CLI({
-    cli:          args.ffmpegPath,
+    cli:          cliBin,
     spawnArgs,
     WorkDir:      args.workDir,
     jobLog:       args.jobLog,
@@ -104,10 +110,13 @@ const plugin = async (args) => {
   const { jobLog, inputFileObj, variables, workDir } = args;
 
   const inputFile = inputFileObj._id;
-  const sourceDir = path.dirname(inputFile);
-  const baseName  = path.basename(inputFile, path.extname(inputFile));
+  // Always save output SRTs in original library directory if known (e.g. from copyToWorkDir)
+  const targetDir = variables.originalLibraryDir || path.dirname(inputFile);
+  const origFile  = variables.originalLibraryFile || inputFile;
+  const baseName  = path.basename(origFile, path.extname(origFile));
 
-  jobLog(`[extractSubtitlesAndOcr] Source: ${inputFile}\n`);
+  jobLog(`[extractSubtitlesAndOcr] Input: ${inputFile}\n`);
+  jobLog(`[extractSubtitlesAndOcr] Output Directory: ${targetDir}\n`);
 
   const subtitleEditPath = args.inputs.subtitleEditPath || 'SubtitleEdit';
   const targetLangs = (args.inputs.languages || 'ita,eng')
@@ -123,6 +132,9 @@ const plugin = async (args) => {
 
   jobLog(`[extractSubtitlesAndOcr] Found ${subtitleStreams.length} subtitle stream(s).\n`);
 
+  const textTasks = [];
+  const bitmapTasks = [];
+
   for (const stream of subtitleStreams) {
     const lang = normLang((stream.tags && stream.tags.language) || '');
     if (!targetLangs.includes(lang)) {
@@ -132,77 +144,106 @@ const plugin = async (args) => {
 
     const codec  = (stream.codec_name || '').toLowerCase();
     const suffix = buildSuffix(lang, stream);
-    const outSrt = path.join(sourceDir, `${baseName}.${suffix}.srt`);
+    const outSrt = path.join(targetDir, `${baseName}.${suffix}.srt`);
 
-    jobLog(`[extractSubtitlesAndOcr] Stream #${stream.index} lang=${lang} codec=${codec} -> ${path.basename(outSrt)}\n`);
-
-    // ── Text-based subtitles ──────────────────────────────────────────────
     if (TEXT_CODECS.has(codec)) {
-      const { cliExitCode } = await runFFmpeg(
-        ['-y', '-i', inputFile, '-map', `0:${stream.index}`, '-c:s', 'srt', outSrt],
+      textTasks.push({ stream, lang, codec, outSrt });
+    } else if (BITMAP_CODECS.has(codec)) {
+      bitmapTasks.push({ stream, lang, codec, outSrt });
+    } else {
+      jobLog(`[extractSubtitlesAndOcr] Stream #${stream.index} codec "${codec}" not supported — skipping.\n`);
+    }
+  }
+
+  // ── 1. Batch extract text-based subtitles (mkvextract or single-pass FFmpeg)
+  if (textTasks.length > 0) {
+    jobLog(`[extractSubtitlesAndOcr] Extracting ${textTasks.length} text subtitle track(s)...\n`);
+    textTasks.forEach((t) => {
+      jobLog(`  -> Stream #${t.stream.index} (${t.lang}, ${t.codec}) to ${path.basename(t.outSrt)}\n`);
+    });
+
+    let extracted = false;
+    const isMkv = inputFile.toLowerCase().endsWith('.mkv') || (inputFileObj.container === 'mkv');
+    const mkvExtract = getMkvExtract(args);
+
+    // Try mkvextract first (instantaneous on MKV)
+    if (isMkv) {
+      try {
+        const mkvArgs = ['tracks', inputFile];
+        textTasks.forEach((t) => mkvArgs.push(`${t.stream.index}:${t.outSrt}`));
+        const { cliExitCode } = await runCLI(mkvExtract, mkvArgs, args);
+        if (cliExitCode === 0) {
+          extracted = true;
+          jobLog('[extractSubtitlesAndOcr] mkvextract completed successfully.\n');
+        } else {
+          jobLog(`[extractSubtitlesAndOcr] mkvextract exited with code ${cliExitCode}, falling back to FFmpeg.\n`);
+        }
+      } catch (err) {
+        jobLog(`[extractSubtitlesAndOcr] mkvextract error (${err.message}), falling back to FFmpeg.\n`);
+      }
+    }
+
+    // Fallback: single-pass FFmpeg (-vn -an for fast demuxing without video/audio decoding)
+    if (!extracted) {
+      const ffmpegArgs = ['-y', '-i', inputFile, '-vn', '-an'];
+      textTasks.forEach((t, i) => {
+        ffmpegArgs.push('-map', `0:${t.stream.index}`, `-c:s:${i}`, 'srt', t.outSrt);
+      });
+      const { cliExitCode } = await runCLI(args.ffmpegPath, ffmpegArgs, args);
+      if (cliExitCode === 0) {
+        jobLog('[extractSubtitlesAndOcr] FFmpeg text extraction completed successfully.\n');
+      } else {
+        jobLog(`[extractSubtitlesAndOcr] WARNING: FFmpeg extraction failed with exit code ${cliExitCode}.\n`);
+      }
+    }
+  }
+
+  // ── 2. Process bitmap subtitles (PGS / VOBSUB) with SubtitleEdit OCR ─────
+  for (const t of bitmapTasks) {
+    const tmpSup = path.join(workDir, `${baseName}_s${t.stream.index}.sup`);
+    jobLog(`[extractSubtitlesAndOcr] Extracting bitmap stream #${t.stream.index} (${t.codec}) for OCR...\n`);
+
+    // Extract to .sup using mkvextract if mkv, else ffmpeg
+    let extractedSup = false;
+    const isMkv = inputFile.toLowerCase().endsWith('.mkv') || (inputFileObj.container === 'mkv');
+    if (isMkv) {
+      const mkvExtract = getMkvExtract(args);
+      const { cliExitCode } = await runCLI(mkvExtract, ['tracks', inputFile, `${t.stream.index}:${tmpSup}`], args);
+      if (cliExitCode === 0) extractedSup = true;
+    }
+    if (!extractedSup) {
+      const { cliExitCode } = await runCLI(
+        args.ffmpegPath,
+        ['-y', '-i', inputFile, '-vn', '-an', '-map', `0:${t.stream.index}`, '-c:s', 'copy', tmpSup],
         args,
       );
       if (cliExitCode !== 0) {
-        jobLog(`[extractSubtitlesAndOcr] WARNING: FFmpeg failed for stream #${stream.index} (exit ${cliExitCode}).\n`);
-      } else {
-        jobLog(`[extractSubtitlesAndOcr] Extracted: ${path.basename(outSrt)}\n`);
-      }
-      continue;
-    }
-
-    // ── Bitmap-based subtitles (PGS / VOBSUB) ────────────────────────────
-    if (BITMAP_CODECS.has(codec)) {
-      const tmpSup = path.join(workDir, `${baseName}_s${stream.index}.sup`);
-
-      // Step 1: Extract to .sup
-      const { cliExitCode: extractExit } = await runFFmpeg(
-        ['-y', '-i', inputFile, '-map', `0:${stream.index}`, '-c:s', 'copy', tmpSup],
-        args,
-      );
-
-      if (extractExit !== 0) {
-        jobLog(`[extractSubtitlesAndOcr] WARNING: .sup extraction failed for stream #${stream.index}.\n`);
+        jobLog(`[extractSubtitlesAndOcr] WARNING: .sup extraction failed for stream #${t.stream.index}.\n`);
         continue;
       }
-
-      // Step 2: OCR with SubtitleEdit
-      try {
-        const { CLI } = require('../../../../FlowHelpers/1.0.0/cliUtils');
-        const seCli = new CLI({
-          cli: subtitleEditPath,
-          spawnArgs: [
-            '/convert', tmpSup, 'subrip',
-            '/track-number:0',
-            `/outputfolder:${sourceDir}`,
-          ],
-          WorkDir:      workDir,
-          jobLog,
-          inputFileObj,
-          updateWorker: args.updateWorker,
-          args,
-        });
-        const { cliExitCode: seExit } = await seCli.runCli();
-
-        if (seExit !== 0) {
-          jobLog(`[extractSubtitlesAndOcr] WARNING: SubtitleEdit OCR failed (exit ${seExit}) for stream #${stream.index}.\n`);
-        } else {
-          // SubtitleEdit names the output after the .sup filename
-          const seOut = path.join(sourceDir, path.basename(tmpSup, '.sup') + '.srt');
-          if (fs.existsSync(seOut)) {
-            fs.renameSync(seOut, outSrt);
-            jobLog(`[extractSubtitlesAndOcr] OCR complete: ${path.basename(outSrt)}\n`);
-          }
-        }
-      } catch (err) {
-        jobLog(`[extractSubtitlesAndOcr] WARNING: SubtitleEdit not available or crashed: ${err.message}\n`);
-      } finally {
-        // Clean up temp .sup
-        try { fs.unlinkSync(tmpSup); } catch (_) { /* ignore */ }
-      }
-      continue;
     }
 
-    jobLog(`[extractSubtitlesAndOcr] Stream #${stream.index} codec "${codec}" not supported — skipping.\n`);
+    // Run OCR with SubtitleEdit CLI
+    try {
+      const seCli = await runCLI(
+        subtitleEditPath,
+        ['/convert', tmpSup, 'subrip', '/track-number:0', `/outputfolder:${targetDir}`],
+        args,
+      );
+      if (seCli.cliExitCode !== 0) {
+        jobLog(`[extractSubtitlesAndOcr] WARNING: SubtitleEdit OCR failed (exit ${seCli.cliExitCode}) for stream #${t.stream.index}.\n`);
+      } else {
+        const seOut = path.join(targetDir, path.basename(tmpSup, '.sup') + '.srt');
+        if (fs.existsSync(seOut)) {
+          fs.renameSync(seOut, t.outSrt);
+          jobLog(`[extractSubtitlesAndOcr] OCR complete: ${path.basename(t.outSrt)}\n`);
+        }
+      }
+    } catch (err) {
+      jobLog(`[extractSubtitlesAndOcr] WARNING: SubtitleEdit not available: ${err.message}\n`);
+    } finally {
+      try { if (fs.existsSync(tmpSup)) fs.unlinkSync(tmpSup); } catch (_) { /* ignore */ }
+    }
   }
 
   jobLog('[extractSubtitlesAndOcr] Done.\n');