I’ve been developing a new project that uses AWS Lambda, a service from Amazon Web Services that lets you run code without having to set up servers etc. AWS Lambda with Chalice, the AWS serverless architecture for Python, makes it easy to set up things like REST APIs.

Here, for example, is an application that simply echoes back a name that is sent to it using POST and x-www-form-urlencoded (i.e., typical form data).

from chalice import Chalice

app = Chalice(app_name='chalice_finsym')
app.debug = True

@app.route('/echoname',methods=["POST"],content_types=['application/x-www-form-urlencoded'],cors=True)
def echonm():
    rdata = parse_qs(app.current_request.raw_body.decode())
    name = rdata.get('nm')[0]
    resp = {}
    resp["name"] = name
    return resp

If you’re writing a REST API, you can allow different output formats to be chosen, e.g., JSON or XML. You might also be nice enough to allow different input content types - not just form data (using the x-www-form-urlencoded content-type), but maybe JSON as well. Chalice allows you to specify a list of content-types in the route specification, but what is not clear (or not commonly explaind on the web), is how to detect which content-type has been used on a particular request. You need to know this because it affects where you find the data in the request object.

So here is my solution.

@app.route('/echoname',methods=["POST"],content_types=['application/x-www-form-urlencoded','application/json'],cors=True)
def echonm():
    name = ''
    if ('content-type' in app.current_request.headers) :
        contenthdr = app.current_request.headers['content-type']
        if contenthdr == 'application/json' :
            rdata = app.current_request.json_body
            name = rdata.get('name')
        else :
            rdata = parse_qs(app.current_request.raw_body.decode())
            name = rdata.get('name')[0]
    else :
        # default content-type
        rdata = parse_qs(app.current_request.raw_body.decode())
        name = rdata.get('name')[0]
    resp = {}
    resp["name"] = name
    return resp

The key is to look for the content-type header in the headers dictionary of the request object, adn then check its value. Beware, though, that the header may not be there - in which case, this code defaults to assuming its x-www-form-urlencoded.