Loading image on server from URL

[manuscript80]

I haven't been able to find any examples of loading an image from a URL while on the server. I've seen the recommendations to use a PUT HTTP request, but it seems really wasteful to download the file from a URL and then send it to Resumable.js. Here's what's I've been trying...

Meteor.http.get(media.url, {npmRequestOptions: { encoding: 'base64' }}, function(e, res) {
  var stream = MediaDocs.upsertStream(
      {
         _id: new Meteor.Collection.ObjectID(),
         filename: media.filename,
         contentType: 'image/' + media.extension
      },
      function (err, fileObj) {
         if (err) {
             console.log(err);
         } else {
             console.log("FileObj: ", fileObj);
         }
      }
   );
   stream.end(res.content);
});

My console output always shows a length of 765031 for fileObj, regardless of the actual content size. Should I be using a different encoding? Or is there a better way to save an image file from the server?

[vsivsi]

Hi, rather than use the Meteor HTTP calls, I'd just use the npm request package, which will return a readable stream that you can pipe to the result of fc.upsertStream()

So something more like the code below will enable you to efficiently handle retrieving large files and writing them into the fileCollection:

var request = require('request');

var stream = MediaDocs.upsertStream(
      {
         _id: new Meteor.Collection.ObjectID(),
         filename: media.filename,
         contentType: 'image/' + media.extension
      },
      function (err, fileObj) {
         if (err) {
             console.log(err);
         } else {
             console.log("FileObj: ", fileObj);
         }
      }
   );

stream.on('finish', function () { /* Perform any actions after write... */);

request(media.url)
   .on('error', function(err) { /* Handle error */ })
   .pipe(stream);

[manuscript80]

That did the trick, thank you. What a flexible package!

On Aug 15, 2016, at 7:11 PM, Vaughn Iverson notifications@github.com wrote:

Hi, rather than use the Meteor HTTP calls, I'd just use the npm request package https://www.npmjs.com/package/request#streaming, which will return a readable stream that you can pipe to the result of fc.upsertStream()

So something more like the code below will enable you to efficiently handle retrieving large files and writing them into the fileCollection:

var request = require('request');

var stream = MediaDocs.upsertStream(
{
_id: new Meteor.Collection.ObjectID(),
filename: media.filename,
contentType: 'image/' + media.extension
},
function (err, fileObj) {
if (err) {
console.log(err);
} else {
console.log("FileObj: ", fileObj);
}
}
);

stream.on('finish', function () { /* Perform any actions after write... */);

request(media.url)
.on('error', function(err) { /* Handle error */ })
.pipe(stream);
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub #140 (comment), or mute the thread https://github.com/notifications/unsubscribe-auth/AQPIW0kd8UEmfxfiXO9l2jJDwwf_bYK7ks5qgPILgaJpZM4JkCeT.

[vsivsi]

Great!