CODEWITHSUNDEEP
node.jsmongodbgridfskoa2
mravich
mravich

Reputation: 33

Image file download or show in browser

I am making an image upload app with Node, Koa2, Mongo, GridFS, and React. Currently I have no problem uploading files with this code:

// HANDLE POST MULTIPART FORM DATA
app.use(async function(ctx, next) {
  if (
    !ctx.request.is('urlencoded', 'multipart') ||
    ('POST' != ctx.method && ctx.url === '/uploadFiles')
  ) return await next()

  console.log('POST REQUESTED AT: ', ctx.url)

  // HANDLE THIS MIME TYPE SOME OTHER WAY!!!!
  var mimeTyp = ''

  // WAIT UNTIL CONNECTION IS RESOLVED
  const conn = await connectionDB()

  // SETUP THE MAIN DATABASE
  const testDB = conn.db('gridfsTestDB')

  // SETUP FS.FILES CONNECTION
  const fsFiles = testDB.collection('fs.files')

  // SET THE BUCKET FOR GRIDFS
  const bucket = new mongodb.GridFSBucket(testDB)

  // DECLARE WHERE YOU WANT THE FILE TO BE UPLOADED AND YOU CAN SUPPLY ADITIONAL DATA HERE TO THE FILE
  const writeStream = bucket.openUploadStream()
  const readStream = bucket.openDownloadStream()
  console.log(readStream)

  // ASYNC-BUSBOY MIDDLEWARE HANDLES FORM SUBMITION AND FILE PIPEING
  const { fields } = await AsyncBusboy(ctx.req, {
    // THIS IS CALLED FOR EACH CHUNK OF FILE SENT THROUGH PIPE
    onFile: function(fieldname, file, filename, encoding, mimetype) {
      console.log('start pipe')
      console.log('MIMETYPE', mimetype)
      mimeTyp = mimetype
      file.pipe(writeStream)
      console.log('end pipe')
    }
  })

  // HERE WE WAIT FOR STREAM OF THE FILE TO FINISH AND THEN WE APPEND METADATA TO IT
  await onStreamFinish(writeStream)
})

But I have a problem reading the file from Mongo and showing it to the user. How would one do this with async approach? createReadStream I figure but had no luck implementing it. I am just looking for guidance on what to use and I will try it.

EDIT(thanks to Rich Churcher) : 

try {
    const conn = await connectionDB()
    const testDB = conn.db('gridfsTestDB')
    const fsFiles = testDB.collection('fs.files')
    const bucket = new mongodb.GridFSBucket(testDB)

    const metadataFile = await getfileToDownload(ctx.params.id, fsFiles)
    const readStream = bucket.openDownloadStream(ObjectId(ctx.params.id))
    console.log(metadataFile)
    ctx.set('Content-Type', metadataFile.contentType)
    ctx.set(
      'Content-Disposition',
      'inline; filename="' + normalizeString(metadataFile.metadata.DocumentName) + '"'
    )
    ctx.body = readStream
  } catch (err) {
    console.log(err)
    ctx.status = 503
    ctx.body = {
      message: 'Unable to preview the document. Error:  ' + err.message
    }
  }

Works just fine!

Upvotes: 3

Views: 2333

Answers (1)

Rich Churcher
Rich Churcher

Reputation: 7654

One of the Koa examples is stream-file, which is a good starting point. Usually it's just a case of setting ctx.body to the stream. Obviously you want to use GridFS as the source rather than the local file system.

I note gridfs-stream, which I think would make life easier for you. However, let's assume you don't want to add another dependency. Here's the bare minimum handler for a download:

const { GridFSBucket, MongoClient } = require('mongodb')

// ...

app.use(async ctx => {
  const connection = await MongoClient.connect('mongodb://localhost:27017')
  const db = connection.db('gfs-example')
  const bucket = new GridFSBucket(db, { bucketName: 'images' })
  ctx.type = 'image/jpg'
  ctx.body = bucket.openDownloadStreamByName('foo.jpg')
})

It looks like you already have your own database tooling set up, I'm just including a connection to make this answer more self-contained. Note that setting ctx.type isn't strictly necessary to get the file, but it helps indicate to the browser that it should display the image rather than treat it as a download.

Upvotes: 1

Related Questions