I would like to add python3 support

Reading through the code I don't see anyway to set additional keyword arguments to passed to the client; since the HTTPRequest api takes something like 28 different keyword arguments this omission makes it quite difficult to do things like disable ssl certificate validation or set basic http auth credentials.

There is no license or copyright statement in this repository. Please provide one, otherwise it can't be forked, distributed, worked on, etc.

def parse(self, body, headers):
    return parse_repo(json.loads(body))

"parse_repo" doesn't appear to be defined in the example or import from any modules. The example, as a result, is broken. Please include the function definition if it is necessary for the example code or remove the reference if it isn't needed.

I just stumpled upon this great package during my research for a rest api consuming client and it looks very promising.

Is it possible to use requests instead of tornado for the client?

Many rest API don't allow you to just pull down the entire contents of a large collection, they are paginated with offset and limit arguments. Can you please provide either built in utility to use such paginated sources or provide an example of how to write a parser that handles pagination.

I've just cloned the repo, created a virtualenv, activated it and installed requirements with pip install -r requiments.txt and ran python github.py

Tornado just throws the following exception:

ERROR:tornado.application:Exception in callback <functools.partial object at 0x7fb9e1809ba8>
Traceback (most recent call last):
  File "/home/gbrennon/Documents/Prog/Python/Tornado/finch_example/finch_repo/env/lib/python2.7/site-packages/tornado/ioloop.py", line 568, in _run_callback
    ret = callback()
  File "/home/gbrennon/Documents/Prog/Python/Tornado/finch_example/finch_repo/env/lib/python2.7/site-packages/tornado/stack_context.py", line 275, in null_wrapper
    return fn(*args, **kwargs)
  File "build/bdist.linux-x86_64/egg/finch/collection.py", line 57, in on_query
    self.on_error(partial(callback, None), response)
  File "build/bdist.linux-x86_64/egg/finch/collection.py", line 41, in on_error
    callback(errors.HTTPError(response.code))
  File "examples/github.py", line 54, in on_repos
    raise error
HTTPError: Forbidden

I would like to divide the current parse(body, headers) method, which parses the http response and returns a dict with the model or collection values, into a parse_response(body, headers) method that only parses the response and returns a raw dict, and a parse(raw) method that receives the raw dict from parse_response and returns a new dict with the values for the model or collection.

See the example below:

class User(Model):
    name = StringField()
    age = IntegerField()

    def parse_response(self, body, headers):
        return json.loads(body)

    def parse(self, raw):
        return {
            'name': raw['fullName'],
            'age': int(raw['age'])
        }

Also, that way we could call the Model.parse method for each item when retrieving an entire collection.