About

Restkit is an HTTP resource kit for Python. It allows you to easily access to HTTP resource and build objects around it. It's the base of couchdbkit a Python CouchDB framework.

Restkit is a full HTTP client using pure socket calls and its own HTTP parser. It's not based on httplib or urllib2.

Installation

Restkit requires Python 2.x superior to 2.5 and http-parser 0.5.3 or sup.

To install restkit using pip you must make sure you have a recent version of distribute installed:

$ curl -O http://python-distribute.org/distribute_setup.py
$ sudo python distribute_setup.py
$ easy_install pip

To install or upgrade to the latest released version of restkit:

$ pip install http-parser
$ pip install restkit

Note: if you get an error on MacOSX try to install with the following arguments:

$ env ARCHFLAGS="-arch i386 -arch x86_64" pip install http-parser

Usage

Perform HTTP call support with restkit.request.

Usage example, get friendpaste page:

from restkit import request
resp = request('http://friendpaste.com')
print resp.body_string()
print resp.status_int

Create a simple Twitter Search resource

Building a resource object is easy using restkit.Resource class. We use simplejson to handle deserialisation of data.

Here is the snippet:

from restkit import Resource

try:
    import simplejson as json
except ImportError:
    import json # py2.6 only

class TwitterSearch(Resource):

    def __init__(self,  pool_instance=None, **kwargs):
        search_url = "http://search.twitter.com"
        super(TwitterSearch, self).__init__(search_url, follow_redirect=True,
                                        max_follow_redirect=10,
                                        pool_instance=pool_instance,
                                        **kwargs)

    def search(self, query):
        return self.get('search.json', q=query)

    def request(self, *args, **kwargs):
        resp = super(TwitterSearch, self).request(*args, **kwargs)
        return json.loads(resp.body)

if __name__ == "__main__":
    s = TwitterSearch()
    print s.search("gunicorn")

Reuses connections

Reusing connections is good. Restkit can maintain for you the http connections and reuse them if the server allows it. To do that you can pass to any object a pool instance inheriting reskit.pool.PoolInterface. By default a threadsafe pool is used in any application:

from restkit import Resource, Manager

manager = TManager(max_conn=10)
res = Resource('http://friendpaste.com', manager=manager)

or if you use Gevent:

from gevent import monkey; monkey.patch_all()

from restkit import request
from restkit.globals import set_manager
from restkit.manager.mgevent import GeventManager

set_manager(GeventManager(timeout=300))

r = request('http://friendpaste.com')

Authentication

Restkit support for now basic authentication and OAuth. But any other authentication schema can easily be added using http filters.

Basic authentication

To use basic authentication in a Resource object you can do:

from restkit import Resource, BasicAuth

auth = BasicAuth("username", "password")
r = Resource("http://friendpaste.com", filters=[auth])

Or simply use an authentication url:

r = Resource("http://username:password@friendpaste.com")

OAuth

Restkit OAuth is based on simplegeo python-oauth2 module So you don't need other installation to use OAuth (you can also simply use restkit.oauth2 module in your applications).

The OAuth filter restkit.oauth2.filter.OAuthFilter allow you to associate a consumer per resource (path). Initalize Oauth filter with:

path, consumer, token, signaturemethod

token and method signature are optionnals. Consumer should be an instance of restkit.oauth2.Consumer, token an instance of restkit.oauth2.Token signature method an instance of oauth2.SignatureMethod (restkit.oauth2.Token is only needed for three-legged requests.

The filter is appleid if the path match. It allows you to maintain different authorization per path. A wildcard at the indicate to the filter to match all path behind.

Example the rule /some/resource/* will match /some/resource/other and /some/resource/other2, while the rule /some/resource will only match the path /some/resource.

Simple client example:

from restkit import OAuthFilter, request
import restkit.oauth2 as oauth

# Create your consumer with the proper key/secret.
consumer = oauth.Consumer(key="your-twitter-consumer-key",
  secret="your-twitter-consumer-secret")

# Request token URL for Twitter.
request_token_url = "http://twitter.com/oauth/request_token"

# Create our filter.
auth = oauth.OAuthFilter('*', consumer)

# The request.
resp = request(request_token_url, filters=[auth])
print resp.body_string()

If you want to add OAuth to your TwitterSearch resource:

# Create your consumer with the proper key/secret.
consumer = oauth.Consumer(key="your-twitter-consumer-key",
  secret="your-twitter-consumer-secret")

# Create our filter.
client = oauth.OAuthfilter('*', consumer)

s = TwitterSearch(filters=[client])