generic support library for signed-token-based auth schemes
Latest commit e270a02 Jan 8, 2014 @rfk rfk Prep version 0.3.1
Failed to load latest commit information.



This is generic support library for doing token-based authentication. You might use it to build a login system using bearer tokens, two-legged oauth, or MAC Access authentication.

Given a server-side master secret, you can serialize a dict of data into an opaque, unforgeable authentication token:

>>> token = tokenlib.make_token({"userid": 42}, secret="I_LIKE_UNICORNS")
>>> print token

Later, you can use the same secret to verify the token and extract the embedded data:

>>> data = tokenlib.parse_token(token, secret="I_LIKE_UNICORNS")
>>> print data
{u'userid': 42, u'expires': 1329875384.073159, u'salt': u'1c033f'}

Notice that the data includes an expiry time. If you try to parse an expired token, it will fail:

>>> # Use now=XXX to simulate a time in the future.
>>> tokenlib.parse_token(token, secret="I_LIKE_UNICORNS", now=9999999999)
Traceback (most recent call last):
ValueError: token has expired

Likewise, it will fail if the token was constructed with a non-matching secret key:

>>> tokenlib.parse_token(token, secret="I_HATE_UNICORNS")
Traceback (most recent call last):
ValueError: token has invalid signature

Each token also has an associated "token secret". This is a secret key that can be shared with the consumer of the token to enable authentication schemes such as MAC Access Authentication of Two-Legged OAuth:

>>> key = tokenlib.get_token_secret(token, secret="I_LIKE_UNICORNS")
>>> print key

For applications that are using the same settings over and over again, you will probably want to create a TokenManager object rather than using the module-level convenience functions:

>>> manager = tokenlib.TokenManager(secret="I_LIKE_UNICORNS")
>>> data = manager.parse_token(token)
>>> print data
{u'userid': 42, u'expires': 1329875384.073159, u'salt': u'1c033f'}

This will let you customize e.g. the token expiry timeout or hash module without repeating the settings in each call.