Node.js 谷歌应用引擎-上传到bucket后访问文件

我已经使用我的google app engine后端上传了一个文件到我的存储桶中，但现在我无法访问该文件以传入ffmpeg。我从try-catch中得到这个错误消息：“输入文件不存在”。我可以看到文件是上传的，因为我检查了存储桶下的开发人员控制台。我使用谷歌提供的样板代码，但添加了ffmpeg进行测试。我试图使用访问上传文件的路径，但这是不正确的，尽管我得到的是bucket.name值和blob.name值。我使用的是“flex”环境

const originalFilePath = `gs://${bucket.name}/${blob.name}`; 

以下是完整的代码：

const process = require('process'); // Required to mock environment variables
const express = require('express');
const helpers = require('./helpers/index');
const Multer = require('multer');
const bodyParser = require('body-parser');
const ffmpeg = require("ffmpeg"); //https://www.npmjs.com/package/ffmpeg
const {Storage} = require('@google-cloud/storage');

// Instantiate a storage client
const storage = new Storage();

const app = express();
app.set('view engine', 'pug');
app.use(bodyParser.json());

// Multer is required to process file uploads and make them available via
// req.files.
const multer = Multer({
storage: Multer.memoryStorage(),
 limits: {
  fileSize: 5 * 1024 * 1024, // no larger than 5mb, you can change as needed.
 },
});

// A bucket is a container for objects (files).
const bucket = storage.bucket(process.env.GCLOUD_STORAGE_BUCKET);

// Display a form for uploading files.
app.get('/', (req, res) => {
 res.render('form.pug');
});

// Process the file upload and upload to Google Cloud Storage.
app.post('/upload', multer.single('file'), (req, res, next) => {

if (!req.file) {
 res.status(400).send('No file uploaded.');
 return;
}

// Create a new blob in the bucket and upload the file data.
const blob = bucket.file(req.file.originalname);
const blobStream = blob.createWriteStream({
 resumable: false,
});

blobStream.on('error', err => {
 next(err);
});

blobStream.on('finish', () => {

const audioFile = helpers.replaceAllExceptNumbersAndLetters(new Date());

// this path is incorrect but I cannot find correct way to do it
const originalFilePath = `gs://${bucket.name}/${blob.name}`; 

const filePathOutput = `gs://${bucket.name}/${audioFile}.mp3`;

try {
 const process = new ffmpeg(originalFilePath);
 process.then(function (video) {
 // Callback mode
 video.fnExtractSoundToMP3(filePathOutput, (error, file) => {
 if (!error)
  res.send(`audio file: ${file}`);
 });
}, (error) => {
 res.send(`process error: ${error}`);

});
} catch (e) {
 res.send(`try catch error: ${JSON.stringify(e)} | bucket: ${JSON.stringify(bucket)} | 
 blob: : ${JSON.stringify(blob)}`);
}  


});

blobStream.end(req.file.buffer);

});

const PORT = process.env.PORT || 8080;
app.listen(PORT, () => {
 console.log(`App listening on port ${PORT}`);
 console.log('Press Ctrl+C to quit.');
});


module.exports = app;

---

我使用评论中的信息创建了这个社区wiki

显示所需的npm包

const functions = require('firebase-functions');
const { Storage } = require('@google-cloud/storage');
const path = require('path');
const os = require('os');
const fs = require('fs');
const ffmpeg = require('fluent-ffmpeg');
const ffmpeg_static = require('ffmpeg-static');

以及如何在bucket中正确构建文件的路径以将其传递到ffmpeg

// Get the file name.
  const fileName = path.basename(filePath);
  // Exit if the audio is already converted.
  if (fileName.endsWith('_output.flac')) {
    functions.logger.log('Already a converted audio.');
    return null;
  }

  // Download file from bucket.
  const bucket = gcs.bucket(fileBucket);
  const tempFilePath = path.join(os.tmpdir(), fileName);
  // We add a '_output.flac' suffix to target audio file name. That's where we'll upload the converted audio.
  const targetTempFileName = fileName.replace(/\.[^/.]+$/, '') + '_output.flac';
  const targetTempFilePath = path.join(os.tmpdir(), targetTempFileName);
  const targetStorageFilePath = path.join(path.dirname(filePath), targetTempFileName);

  await bucket.file(filePath).download({destination: tempFilePath});
  functions.logger.log('Audio downloaded locally to', tempFilePath);
  // Convert the audio to mono channel using FFMPEG.

  let command = ffmpeg(tempFilePath)
      .setFfmpegPath(ffmpeg_static)
      .audioChannels(1)
      .audioFrequency(16000)
      .format('flac')
      .output(targetTempFilePath);

  await promisifyCommand(command);
  functions.logger.log('Output audio created at', targetTempFilePath);

演示如何替换ffmpeg_static.path安装以及如何设置正确的路径

const ffmpegPath = require('@ffmpeg-installer/ffmpeg').path;
const ffmpeg = require('fluent-ffmpeg');

let command = ffmpeg(tempFilePath)
      .setFfmpegPath(ffmpegPath)
      .audioChannels(1)
      .audioFrequency(16000)
      .format('flac')
      .output(targetTempFilePath);

---

我认为dev env上没有

gs://..

。在本地环境中（至少对于图像），我通常会得到类似

http://localhost:8080/_ah/img/encoded_gs_file:....

。这里的关键是url有

编码的\u gs\u文件：

。如果这是生产，那么您应该使用url

https://storage.googleapis.com/${bucket.name}/${blob.name}

谢谢@NoCommandLine。我已经尝试过了，但是当我尝试将ffmpeg函数传递给它时，我遇到了一个错误。这是来自try-catch:{“code”：103，“msg”：“输入文件不存在”}的错误。你知道在以这种方式访问文件之前，需要在google开发者控制台中设置哪些特殊权限吗？经过更多检查，我现在相信这不是google API的问题，而是节点ffmpeg库的问题。它不接受带有路径的仅URL的本地文件。此repo向我展示了如何计算路径以及要使用的npm包：，此外，我还使用此帖子找到了ffmpeg静态的替代方案。现在一切都像一个符咒。