Parsing raw_body from AWS Chalice multipart/form-data http request

Question

I have the following simple AWS Chalice route:

@app.route('/submit', methods=['POST'],
           content_types=['multipart/form-data'])
def submit():
    request_info = app.current_request.raw_body

    return request_info

I then use a simple form with multipart data, including a docx file upload:

<form enctype="multipart/form-data" method="POST" action="http://localhost:8000/submit">
    <input name='foo' type="text">
    <br>
    <input name="bar" type="file">
    <br>
    <button type='submit'>
    Submit
    </button>
</form>

The raw_body attribute of the request is just the bytes of the http request, and I'm looking for a pre-existing Python library that will let me extract each form field and also write the bytes of the docx file to disk (in this case a tmp folder in AWS Lambda). Is there a library that will take raw_body as an argument and allow me to parse the individual fields so that I don't have to write such a parser myself? Trying to google for this is difficult, since most of the results that come back all have to do with using python to consume web APIs, which is not what I want.

score 0 · Accepted Answer · edited Jan 04 '19 at 08:25

Below is sample lambda code which will take a multipart/form-data and parse it and get the files and get the file type as well.

import magic
from io import BytesIO
import json
import cgi

def lambda_handler(event, context):
   content_type_obj = event['params']['header']['content-type']
   content_type, property_dict = cgi.parse_header(content_type_obj)
   form_data = cgi.parse_multipart(BytesIO(event['body-json'].decode('base64')), property_dict)
   form_file = form_data['file'][0]
   file_type = magic.from_buffer(form_file, mime=True)
   file_name = "new_file." + file_type.split('/')[-1] or "txt"
   # process your file 
   # file_type will give you mime type of the file like "image/png"
   print file_type

   return {'statusCode': 200,
           'body': json.dumps({"status": "success",
                               "message": "your request for uploading has been accepted."}),
           'headers': {
           'Content-Type': 'application/json',
          }}

for adding magic to lambda refer packaging

http://docs.aws.amazon.com/lambda/latest/dg/lambda-python-how-to-create-deployment-package.html

The cgi library works, even though it's temperamental. For anyone reading this that uses Chalice, my actual solution was to just base64 encode the files and decode in the route function definition. This increases data transfer by a percentage, but is way easier to troubleshoot. — Joe Stech, Oct 16 '17 at 04:40
@JoeStech you can use zappa with flask and the files uploaded will be available in request.files as we get normally in flask. Then its much easier to debug. Not mention you will have all the attributes of the file mimetype,content-length etc. — Anurag, Oct 24 '17 at 19:51

Parsing raw_body from AWS Chalice multipart/form-data http request

1 Answers1