Merge pull request #1 from helloint/feat/include-ffmpeg-dep

helloint · web-flow · commit 475c6cc28a0b · 2025-08-02T14:36:10.000+08:00
Feat/include ffmpeg dep
diff --git a/README.md b/README.md
@@ -1,12 +1,10 @@
 # 一键批量提取流媒体视频的中英字幕，并自动合并为双语字幕。
+[![NPM](https://nodei.co/npm/dual-subtitle.png?downloads=true)](https://www.npmjs.com/package/dual-subtitle)
 
 ## 依赖
-需要本地先安装好`FFmpeg`(包含`ffmpeg`, `ffprobe`这两个命令)，以及有`Node.js`环境(包括`npm`)
+需要本地有`Node.js`环境(包括`npm`)
 
 * 安装【[Node.js/npm环境](https://nodejs.org/zh-cn/)】
-* 安装【[FFmpeg](http://ffmpeg.org/download.html)】
-
-计划在下个版本里去掉FFmpeg，改为依赖npm的ffmpeg，简化使用。
 
 ## 使用
 
diff --git a/analyzeMedia.js b/analyzeMedia.js
@@ -1,37 +1,17 @@
-import { spawn } from 'child_process';
+import ffprobe from "ffprobe";
+import ffprobeStatic from "ffprobe-static";
 import {config} from "./config.js";
 
-const cmd = 'ffprobe';
-
-const getArgs = (file) => (['-i', file, '-hide_banner', '-select_streams', 's', '-show_entries', 'stream=index,codec_name:stream_tags=language,title', '-of', 'csv=p=0']);
-
 export const analyzeMedia = (file) => {
     return new Promise((resolve, reject) => {
-        const proc = spawn(cmd, getArgs(config.workdir + file));
-
-        let stdoutData = '';
-        let stderrData = '';
-
-        proc.stdout.on('data', function(data) {
-            stdoutData += data.toString();
-        });
-
-        proc.stderr.setEncoding("utf8");
-        proc.stderr.on('data', function(data) {
-            stderrData += data;
-        });
-
-        proc.on('close', function(code) {
-            // console.error('stderr:', stderrData);
-            if (code === 0) {
-                resolve(stdoutData);
-            } else {
-                reject(new Error(`Process exited with code ${code}: ${stderrData}`));
-            }
-        });
-
-        proc.on('error', function(err) {
-            reject(err);
-        });
+        ffprobe(config.workdir + file, { path: ffprobeStatic.path })
+            .then(function (info) {
+                // console.log(info);
+                resolve(info);
+            })
+            .catch(function (err) {
+                // console.error(err);
+                reject(err);
+            });
     });
 };
diff --git a/extractSub.js b/extractSub.js
@@ -1,40 +1,34 @@
-import { spawn } from 'child_process';
+import ffmpegInstaller from '@ffmpeg-installer/ffmpeg';
+import ffmpeg from 'fluent-ffmpeg';
 import {config} from "./config.js";
 import {removeExtension} from "./utils.js";
 
-const cmd = 'ffmpeg';
-
-const getArgs = (file, subIdx, main, secondary) => (['-y', '-i', file, '-map', `0:${subIdx[0]}`, '-c', 'copy', main, '-map', `0:${subIdx[1]}`, '-c', 'copy', secondary]);
+ffmpeg.setFfmpegPath(ffmpegInstaller.path);
 
 export const extractSub = (filename, subIdx) => {
     return new Promise((resolve, reject) => {
         const mainSrt = `${removeExtension(filename)}.chs.srt`;
         const secondarySrt = `${removeExtension(filename)}.eng.srt`;
-        const proc = spawn(cmd, getArgs(config.workdir + filename, subIdx, config.workdir + mainSrt, config.workdir + secondarySrt));
-
-        let stdoutData = '';
-        let stderrData = '';
-
-        proc.stdout.on('data', function(data) {
-            stdoutData += data.toString();
-        });
 
-        proc.stderr.setEncoding("utf8");
-        proc.stderr.on('data', function(data) {
-            stderrData += data;
-        });
-
-        proc.on('close', function(code) {
-            // console.error('stderr:', stderrData);
-            if (code === 0) {
+        ffmpeg(config.workdir + filename)
+            .output(config.workdir + mainSrt)
+            .outputOptions(['-map', `0:${subIdx[0]}`, '-c', 'copy'])
+            .output(config.workdir + secondarySrt)
+            .outputOptions(['-map', `0:${subIdx[1]}`, '-c', 'copy'])
+            .run()
+            .on('start', function (str) {
+                console.log('转换任务开始～', str);
+            })
+            .on('progress', function (progress) {
+                console.log(`进行中，完成${(progress.percent || 0)}%`);
+            })
+            .on('end', function (str) {
+                console.log('转换任务完成！');
                 resolve([mainSrt, secondarySrt]);
-            } else {
-                reject(new Error(`Process exited with code ${code}: ${stderrData}`));
-            }
-        });
-
-        proc.on('error', function(err) {
-            reject(err);
-        });
+            })
+            .on('error', function (err) {
+                console.log('转换任务出错：', err);
+                reject(err);
+            });
     });
 };
diff --git a/findIndex.js b/findIndex.js
@@ -1,42 +1,76 @@
-export const findIndex = (text) => {
-    const lines = text.split('\n');
-    const result = [];
-
-    for (const line of lines) {
-        const parts = line.split(',');
-        if (parts.length >= 3) { // 至少要有3部分：id,格式,语言代码
-            const id = parts[0];
-            const langCode = parts[2].toLowerCase();
-            const langName = parts[3] ? parts[3].toLowerCase() : '';
-
-            // 检查简体中文：代码为chi且名称包含"简体"或"simplified"
-            const isSimplifiedChinese = langCode === 'chi' && (
-                langName.includes('简体') ||
-                langName.includes('simplified') ||
-                langName === '简体' // 只有"简体"的情况
-            );
-
-            // 检查英语：代码为eng或名称包含"english"
-            const isEnglish = langCode === 'eng' ||
-                langName.includes('english') ||
-                langName === 'sdh' || // eng,SDH
-                langName === '';     // 只有eng代码的情况
-
-            if (isSimplifiedChinese || isEnglish) {
-                result.push(parseInt(id));
-            }
-        }
+export const findIndex = (info) => {
+    const subTitles = info.streams.filter((stream) => stream.codec_type === 'subtitle').map(stream => ({
+        index: stream.index,
+        code: stream.tags.language.toLowerCase(),
+        name: stream.tags.title ? stream.tags.title.toLowerCase() : ''
+    }));
+
+    const chsIdx = findChiSub(subTitles);
+    const engIdx = findEngSub(subTitles);
+
+    if (chsIdx === undefined) {
+        console.log('没有找到简体中文字幕');
+    } else {
+        console.log('找到简体中文字幕，索引为：', chsIdx);
+    }
+
+    if (engIdx === undefined) {
+        console.log('没有找到英语中文字幕');
+    } else {
+        console.log('找到英语字幕，索引为：', engIdx);
+    }
+
+    if (chsIdx === undefined || engIdx === undefined) {
+        throw new Error('字幕查找失败，中断执行');
+    }
+
+    return [chsIdx, engIdx];
+}
+
+/**
+ * 目前看到的数据可能有：
+ * chi,简体
+ * chi,Simplified Chinese
+ * 查找策略是：先找 'chi', 如果数量大于1，则进一步找 "简体"
+ */
+const findChiSub = (subTitles) => {
+    const chineseSubtitles = subTitles.filter(subTitle => subTitle.code === 'chi');
+
+    if (chineseSubtitles.length === 0) {
+        return undefined;
     }
 
-    // 按中文在前，英文在后排序
-    return result.sort((a, b) => {
-        const lineA = lines.find(l => l.startsWith(a + ','));
-        const lineB = lines.find(l => l.startsWith(b + ','));
-        const isChineseA = lineA.includes('chi');
-        const isChineseB = lineB.includes('chi');
-
-        if (isChineseA && !isChineseB) return -1;
-        if (!isChineseA && isChineseB) return 1;
-        return a - b;
-    });
+    if (chineseSubtitles.length === 1) {
+        return chineseSubtitles[0].index;
+    }
+
+    // If multiple Chinese subtitles, look for Simplified Chinese
+    const simplifiedChinese = chineseSubtitles.find(subTitle =>
+        subTitle.name.includes('简体') ||
+        subTitle.name.includes('simplified')
+    );
+
+    return simplifiedChinese?.index;
 }
+
+/**
+ * 目前看到的数据可能有：
+ * 9,subrip,eng
+ * 10,subrip,eng,SDH
+ * 23,subrip,eng,English[CC]
+ * 如果同时有SDH和非SDH版本，选非SDH版本。
+ */
+const findEngSub = (subTitles) => {
+    const englishSubs = subTitles.filter(sub => sub.code === 'eng');
+
+    if (englishSubs.length === 0) return undefined;
+    if (englishSubs.length === 1) return englishSubs[0].index;
+
+    // Filter out SDH subtitles if there are multiple English options
+    const nonSDHSubs = englishSubs.filter(sub =>
+        !sub.name.includes('sdh')
+    );
+
+    // Return first non-SDH sub if available, otherwise first English sub
+    return (nonSDHSubs[0] || englishSubs[0])?.index;
+};
diff --git a/index.js b/index.js
@@ -12,9 +12,9 @@ const main = async () => {
     const mediaFiles = fs.readdirSync(config.workdir).filter((file) => config.exts.includes(path.extname(file)));
 
     for (const file of mediaFiles) {
-        console.log(`processing: ${file}`);
-        const metadata = await analyzeMedia(file);
-        const subIndex = findIndex(metadata);
+        console.log(`正在处理：${file}`);
+        const mediaInfo = await analyzeMedia(file);
+        const subIndex = findIndex(mediaInfo);
         const srts = await extractSub(file, subIndex);
         mergeSrtFiles(srts[0], srts[1], `${removeExtension(file)}.${config.srtTag}.srt`);
         deleteFile(srts[0]);
diff --git a/mergeSub.js b/mergeSub.js
@@ -87,4 +87,5 @@ export const mergeSrtFiles = (mainSrtPath, secondarySrtPath, outputPath) => {
 
     // 写入输出文件
     fs.writeFileSync(`${config.workdir}${outputPath}`, outputContent.trim());
+    console.log(`生成双语字幕文件：${outputPath}`);
 }
diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json

Original file line number	Diff line number	Diff line change
`@@ -87,4 +87,5 @@ export const mergeSrtFiles = (mainSrtPath, secondarySrtPath, outputPath) => {`
`87`	`87`
`88`	`88`	`// 写入输出文件`
`89`	`89`	fs.writeFileSync(`${config.workdir}${outputPath}`, outputContent.trim());
	`90`	+ console.log(`生成双语字幕文件：${outputPath}`);
`90`	`91`	`}`