ansaurus

Question

Answer 1

+2 A:

Not sure what you mean by "without using tools" -- Python is "a tool", right?

With just Python and its standard library (2.6 or better), add at the top of your module

import json

and change the return statement to

return json.dumps(message)

Alex Martelli 2010-09-04 05:20:06

In Python 2.5, you can use the [simplejson](http://pypi.python.org/pypi/simplejson/) package.

detly 2010-09-04 05:24:01

Actually, what I meant by saying "without tools" was without using decorators. I'm still getting used to python's idiosyncrasies. Yes, I have seen simplejson used, but I want to be able to do it using the base libraries first.

Sean Ochoa 2010-09-04 06:29:23

@Sean, `simplejson` is just the older version of `json`, from before the latter was incorporated into the Python standard library. IOW, `json` is not any more "base" than `simplejson`, it's just a "packaging" decision by the PSF (specifically by the core developers of Python itself and the standard library that goes with it).

Alex Martelli 2010-09-04 14:34:32

@Alex: "Tool" is CherryPy jargon for "plugin". There's a whole subsystem for using them and making your own: http://docs.cherrypy.org/dev/intro/concepts/tools.html

fumanchu 2010-09-06 01:16:19

@fumanchu, tx for the translation!-)

Alex Martelli 2010-09-06 01:50:46

Answer 2

+1 A:

Note that in CherryPy 3.2 (almost done!) there will be a pair of JSON tools to make the above even easier:

@cherrypy.expose
@tools.json_out()
def monkey(self, **kwargs):
    return {"message": "Hello World!"}

json_out encodes the output and sets the header for you.

fumanchu 2010-09-04 16:58:54

Thanks for this. In reading Alex's post above and doing some testing on my own, I realize that your suggestion is more succinct, explicit, and readable than using JSON.dumps().

Sean Ochoa 2010-09-05 22:37:19

ansaurus

tags:

views:

answers:

CherryPy How to respond with JSON?

related questions