Hawk is an HTTP authentication scheme using a message authentication code (MAC) algorithm to provide partial HTTP request cryptographic verification.
PyHawk is great for consuming or providing webservices from Python.
PyHawk's goal is to track as closely to the original NodeJS' hawk code, because hawk is a primarily an authentication scheme documented by the implementaiton (as opposed to a standard).
If you find this module un-pythonic, also consider:
- mohawk Pythonic Hawk library
- hawkauthlib
If you had code that consumed a HAWK authenticated webservice, you could do something like the following
import hawk import requests # Hawk is secured with a shared secret credentials = db.lookup_secrets(some_id) # Prepare your request headers header = hawk.client.header(url, 'GET', { 'credentials': credentials, 'ext': 'Yo Yo'}) # Which goes into Authorization field of HTTP headers headers = [('Authorization', header['field'])] res = requests.get(url, data=params, headers=headers) response = { 'headers': res.headers } # We can verify we're talking to our trusted server verified = hawk.client.authenticate(response, credentials, header['artifacts'], {'payload': res.text}) if verified: print res.text else: print "Something fishy going on."
See sample_client.py for details.
If you provide a webservice and want to do authentication via HAWK, do something like the following:
import hawk # A callback function for looking up credentials def lookup_hawk_credentials(id): # Some collection of secrets return db.lookup(id) # req is a Request object from your webserver framework if 'Hawk ' in req.headers['Authorization']: return check_auth_via_hawk(req) else: return failure(req, res) def check_auth_via_hawk(req): server = hawk.Server(req, lookup_hawk_credentials) # This will raise a hawk.util.HawkException if it fails artifacts = server.authenticate() # Sign our response, so clients can trust us auth = server.header(artifacts, { 'payload': payload, 'contentType': 'text/plain' }) headers = [('Content-Type', 'text/plain'), ('Server-Authorization', auth)] start_response(status, headers) return payload
See sample_server.py for details.
PyHawk uses python logging to emit information about why authorization is
failing and so on. You can configure these logger channels with INFO
,
DEBUG
, etc, to get some helpful output.
- hawk
- All hawk logging, including everything below.
- hawk.client
- All hawk client related messages, including header construction.
- hawk.server
- All hawk server related messages, including authorization.
- hawk.hcrypto
- All hawk crypto related messages, including bewit handling.
- hawk.util
- All shared hawk code such as header normalization.
This is under development, ready for adventurous users. There doesn't appear to be a Python library for HAWK. Let me know if there is already a robust library.
Optionally use env as a virtualenv
virtualenv env source env/bin/activate
Locally install source:
python setup.py develop
Unit tests are in hawk/tests.
python hawk/tests/test_*.py
Additionally, one can test compatibility:
The compatibility/nodejs directory has a server.js and a client.js (Node code) which are from HAWK's usage.js.
To test the server, do the following:
- python sample_server.py
- cd compatibility/nodejs/
- node client.js
Output should be
Authenticated Request is 200 (OK) Response validates (OK) Unauthenticated request should 401 - (OK)
Note: the port numbers in test_pyhawk.py and client.js must match.
To test the client, do the following:
- cd compatibility/nodejs/
- node server.js
- cd ../..
- python sample_client.py
Output should be
Response validates (OK)
Edit setup.py and bump the version number.
python setup.py sdist upload
You should see your updates at https://pypi.python.org/pypi?%3Aaction=pkg_edit&name=PyHawk