From 6efcc0dabe5cee388af684c63dab1ecf0bb2769a Mon Sep 17 00:00:00 2001
From: James Ketrenos <james_eikona@ketrenos.com>
Date: Tue, 26 Oct 2021 13:48:57 -0700
Subject: [PATCH] Added subtitle

Signed-off-by: James Ketrenos <james_eikona@ketrenos.com>
---
 subtitle-extract | 104 +++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 104 insertions(+)
 create mode 100755 subtitle-extract

diff --git a/subtitle-extract b/subtitle-extract
new file mode 100755
index 0000000..8eeda91
--- /dev/null
+++ b/subtitle-extract
@@ -0,0 +1,104 @@
+#!/usr/bin/nodejs
+const { spawn } = require('child_process'),
+  filename = process.argv[2],
+  child = spawn('ffmpeg', [
+    '-i', filename
+  ]),
+  streams = [],
+  basename = filename.replace(/\.[^.]+$/, "");
+
+let output = [];
+
+child.stdout.on('data', (data) => {
+});
+
+child.stderr.on('data', (data) => {
+  output += data;
+});
+
+/*
+    Stream #0:2(eng): Subtitle: ass (default)
+    Metadata:
+      title           : English
+      BPS-eng         : 130
+      DURATION-eng    : 00:47:44.040000000
+...
+    Stream #0:4(eng): Subtitle: hdmv_pgs_subtitle
+    Metadata:
+      BPS             : 3072
+      BPS-eng         : 3072
+      DURATION        : 00:21:51.602000000
+      DURATION-eng    : 00:21:51.602000000
+     
+*/
+
+let stream = undefined;
+
+child.on('close', code => {
+  output.toString().split('\n').forEach(line => {
+    const parts = line.match(/^\s*Stream\s+#([^ ]+) Subtitle:\s*([^ ]+).*$/);
+    if (parts && parts.length > 0) {
+      /* parts[1] = '0:2(eng):'
+         parts[2] = 'srt'
+      */
+      if (stream) {
+        streams.push(stream);
+      }
+      stream = {
+        extension: parts[2].replace(/hdmv_pgs_subtitle/, 'sup'), /* srt, sup */
+        number: parts[1].replace(/^.*([^(]+).*$/, '$1'),         /* 0:2 -- not used */
+        language: parts[1].replace(/^[^(]*\((.*)\).*$/, '$1')    /* eng */
+      };
+      return;
+    }
+    /* We tried parsing out 'title' for the langauge title, however
+     * that is not a consistent standard and title seems to be used 
+     * arbitrarily. Instead, we will rely on the 'language' we parse
+     * from the stream name, which is specified inISO 639-1:
+ 
+       https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes
+ 
+    */
+    if (stream) {
+      if (!stream.metadata) {
+        if (/^\s*Metadata:.*/.exec(line)) {    
+          stream.metadata = true;
+        } else {
+          streams.push(stream);
+          stream = undefined;
+        }
+        return;
+      }
+      const language = line.match(/^\s*title\s*:\s*([^\s]+)/);
+      if (language && language.length > 0) {
+        stream.language = language[1].toLowerCase();
+        streams.push(stream);
+        stream = undefined;
+      }
+    }
+  });
+  streams.forEach((stream, index) => {
+    const target = `${basename}.${index}_${stream.language}.${stream.extension}`,
+      ffmpeg = spawn('ffmpeg', [
+      '-i', filename,
+      '-c:s', 'copy',
+      '-map', `0:s:${index}`,
+      '-y',
+      target
+    ]);
+    console.log(`0:s:${index} => ${target}`);
+
+    ffmpeg.stdout.on('data', (data) => {
+      console.log(data.toString());
+    });
+
+    ffmpeg.stderr.on('data', (data) => {
+      console.error(data.toString());
+    });
+ 
+    ffmpeg.on('close', code => {
+      console.log(`Completed: ${target} ${code}`);
+    });
+  });
+  console.log(JSON.stringify(streams));
+});