node.js - 从 Fluent-Ffmpeg Api 调用 Ffmpeg 二进制文件时 Node 出现 ENOENT 错误

Question

背景

我正在 Node 中连接 firebase 函数。目的是将入站音频剪辑解析为设定长度。使用 ffmpeg 和 Fluent-ffmpeg。

问题

当该函数在 firebase 中触发时，当 Fluent-Ffmpeg 尝试访问 Ffmpeg 二进制文件时，我收到 ENOENT 错误

Firebase 调试输出

Error: { Error: spawn ./Cloud/functions/node_modules/ffmpeg-binaries/bin/ffmpeg ENOENT at exports._errnoException (util.js:1018:11) at Process.ChildProcess._handle.onexit (internal/child_process.js:193:32) at onErrorNT (internal/child_process.js:367:16) at _combinedTickCallback (internal/process/next_tick.js:80:11) at process._tickDomainCallback (internal/process/next_tick.js:128:9) code: 'ENOENT', errno: 'ENOENT', syscall: 'spawn ./Cloud/functions/node_modules/ffmpeg-binaries/bin/ffmpeg', path: './Cloud/functions/node_modules/ffmpeg-binaries/bin/ffmpeg',
spawnargs: [ '-formats' ] }

预期结果

入站文件被下载到临时目录，进行裁剪，然后作为裁剪后的文件重新上传到 Firebase 存储。

环境

• Mac 客户端/Firebase 存储
• Node v8.1.0
• ffmpeg v3.2.2
• 流利的ffmpeg v2.1.2

代码[已更新以反射(reflect) Svenskunganka 的更改。现在有效]

const ffmpeg = require('fluent-ffmpeg');
const PREVIEW_PREFIX = 'preview_';

exports.generatePreviewClip = functions.storage.object('audioFiles').onChange(event => {

      //console.log('Times this function has run: ', run++);

      const object = event.data; // The Storage object.
      const fileBucket = object.bucket; // The Storage bucket that contains the file.
      const filePath = object.name; // File path in the bucket.
      const contentType = object.contentType; // File content type.
      const resourceState = object.resourceState; // The resourceState is 'exists' or 'not_exists' (for file/folder deletions).
      const metageneration = object.metageneration; // Number of times metadata has been generated. New objects have a value of 1.

      // Exit if this is triggered on a file that is not an audio file.
      if (!contentType.startsWith('audio/')) {
        console.log('This is not an audio file.');
        console.log('This is the file:', filePath);
        return;
      }

      // Get the file name.
      const fileName = path.basename(filePath);
      console.log('Working with filename', fileName);
      // Exit if the file is already an audio clip.
      if (fileName.startsWith(PREVIEW_PREFIX)) {
        console.log('Already a preview clip.');
        return;
      }

      // Exit if this is a move or deletion event.
      if (event.data.resourceState === 'not_exists') {
        console.log('This is a deletion event.');
        return;
      }

      // Exit if file exists but is not new and is only being triggered
      // because of a metadata change.
      if (resourceState === 'exists' && metageneration > 1) {
        console.log('This is a metadata change event.');
        return;
      }

      // Download file from bucket.

      const bucket = gcs.bucket(fileBucket);
      const tempFilePath = path.join(os.tmpdir(), fileName);
      return bucket.file(filePath).download({
        destination: tempFilePath
      }).then(() => {

        console.log('Audio file downloaded locally to temp directory', tempFilePath);

    var ffmpegPath = require("ffmpeg-binaries").ffmpegPath();
    var ffprobePath = require("ffmpeg-binaries").ffprobePath();

    // Generate a croped file using ffmpeg.
    var command = new ffmpeg(tempFilePath);
        command.setFfmpegPath(ffmpegPath);
        command.setFfprobePath(ffprobePath);

        command
              .setStartTime('00:00:03')
              .setDuration('10')
              .output(tempFilePath)
              .on('end', function() {
                    console.log('Audio Crop Done Successfully');
               })
               .on('error', function(err)
               {
                  console.log('Error:', err);
               }).run();

              }).then(() => {
        console.log('Preview file created at', tempFilePath);
        // We add a 'preview_' prefix to the audio file name. that's how it will appear in firebase.
        const previewFileName = PREVIEW_PREFIX + fileName;
        console.log('previewFileName is', previewFileName)
        const previewFilePath = path.join(path.dirname(filePath), previewFileName);
        console.log('previewFilePath is', previewFilePath);
        // Uploading the preview file.
        return bucket.upload(tempFilePath, {destination: previewFilePath});
      // Once the file has been uploaded delete the local file to free up disk space.
      }).then(() => fs.unlinkSync(tempFilePath));

      // [END audio file generation]

    });

我的 ffmpeg-binaries/bin 目录的内容和结构

-rwxrwxrwx  1 sherpa  staff    24M Dec 10  2016 ffmpeg
-rwxr--r--  1 sherpa  staff    35M Jan 12  2017 ffmpeg.exe
-rwxr--r--  1 sherpa  staff    35M Jan 12  2017 ffplay.exe
-rwxrwxrwx  1 sherpa  staff    24M Dec 10  2016 ffprobe
-rwxr--r--  1 sherpa  staff    35M Jan 12  2017 ffprobe.exe
-rwxrwxrwx  1 sherpa  staff    22M Dec 10  2016 ffserver

我尝试过的事情

• 我可以从命令行执行 ffmpeg
• sudo chmod -R u+x ffmpeg-binaries/
• 在全局路径中设置 ffmpeg
• 在 setFfmpegPath 中使用 ffmpeg.exe 二进制文件，得到相同的结果
  • 错误:{错误:spawn ./Cloud/functions/node_modules/ffmpeg-binaries/bin/ffmpeg.exe ENOENT
• 使用了许多不同的 setFfmpegPath 路径结构，例如:
  • ./Cloud/functions/node_modules/ffmpeg-binaries/bin/ffmpeg
  • node_modules/ffmpeg-binaries/bin/ffmpeg
  • ./Cloud/functions/node_modules/ffmpeg-binaries/bin/

感谢您的任何建议。

Answer 1

我们在问题的评论中解决了该问题，但我会为可能遇到相同问题的任何 future 用户发布答案。问题是提供给 setFfmpegPath() 方法的路径是相对的，而应该是绝对的。 ffmpeg-binaries 模块导出几个辅助函数，您可以调用它来获取其二进制文件的路径:

var ffmpeg = require("fluent-ffmpeg")
var ffmpegPath = require("ffmpeg-binaries")

ffmpeg
  .setFfmpegPath(ffmpegPath)
  ...

确保您已使用 npm i -S ffmpeg-binaries 安装了 ffmpeg-binaries。

2018 年 11 月 7 日更新:

ffmpeg-binaries 包在 4.0.0 版本中发布了一项新的重大更改，删除了它导出的所有函数，而仅导出指向 ffmpeg 所在目录的字符串。这已在提交 009e4d5 中更改.
我已更新答案以反射(reflect)这些更改。