tmfksoft
commented
7 years ago NOTICE REGARDING CHANGING THE RETURNED DATA
I've also updated the README to reflect the changes. I ended up changing the format of the files to suit following spec where "name" is used to refer to files uploaded via a HTTP Form.
The data returned is no longer an array of files and is instead an object. The object keys will be the "name" of each file. Thus, a file uploaded with the name being "avatar" you would access the file as form.avatar In the event "name" is not supplied it will default to null. If multiple files are supplied with the same name, such as during a multiple file upload an array of files will be supplied instead.
Here's an example of what to expect:
{
avatar: {
type: 'image/png',
filename: 'avatar.png',
name: 'avatar',
data: <Buffer 00 00 00>
}
}If you recieve multiple avatars:
{
avatar: [
{
type: 'image/png',
filename: 'avatar.png',
name: 'avatar',
data: <Buffer 00 00 00>
},
{
type: 'image/png',
filename: 'avatar.png',
name: 'avatar',
data: <Buffer 00 00 00>
},
]
}
christiansalazar
This PR rewrites the process function. In the PR I've added support for lacking Content-Disposition as per a multipart/related request where it isn't mandatory.
Added support for lacking filename which also isn't mandatory. Feel free to use any of the code in the PR.
It should be pretty easy to expand on my code to add support for different headers. The code was tested with Amazon AWS V1 Speech Recognition.
An example payload from Amazon AWS V1: http://haste.thomas-edwards.me/usitucezad.txt
Note the lack of Content-Disposition.
One last addition, A requests entire headers object and now be supplied to getBoundary and it'll try and extract content-type. Although will work fine as it did originally.