Social networking middleware for Facebook, Twitter, LinkedIn and Google (OpenID + Buzz)
Project description
About the middleware
ao.social is a social networking middleware that aims to provide a generic interface for multiple social networking web services (Facebook, Twitter, Google (OpenID login) and LinkedIn are currently implemented). It provides a standard WSGI middleware that adds the ao.social.user environment variable to the WSGI environment.
Note that you also need to add Beaker to your WSGI pipline, otherwise the middleware won’t be able to remember the users. It’s up to the developer to configure Beaker. The recommended setup is using secure cookies to store session data; however, other methods should work just as fine.
To use the ao.social middleware with Django, you need to add it to the WSGI pipeline just as you would with any other framework. ao.social does not provide a Django middleware. Instead, it won’t even call Django if it is not required (i.e. on the login handler pages). The recommended way of plumbing Django into the WSGI pipeline along with Beaker and ao.social is using the twod.wsgi package.
Configurung the middleware
The social middleware is a generic, stand-alone component that can be used for common social networking interactions, regardless of the web framework being used. However, it needs some configurations. To make things easier, you can store this configuration in an external settings file. For example, you can create a YAML config file like this:
login_path: /login/%s/ user_class: foomodule.models.User facebook: key: your-facebook-api-key secret: your-facebook-api-secret twitter: key: your-twitter-consumer-key secret: your-twitter-consumer-secret google: realm: http://www.example.com/ secret: your-google-api-secret callback: http://www.example.com/login/google/ linkedin: key: your-linkedin-consumer-key secret: your-linkedin-consumer-secret
Note that this is a minimal configuration. You can override some of the default middleware behavior by using extra parameters in the configuration. For more information, take a look at the documented tests and the source code.
If you do not wish to use one of the services, simply leave out that section from the config file. That way the client machinery won’t be loaded at all.
To load the config file from the file, use the PyYAML module:
>>> import yaml >>> with open('auth.yaml', 'r') as file: ... conf = yaml.load(confstr)
For the google login to work, the callback must be the login path for google:
>>> conf['login_path'] % 'google' '/login/google/' >>> conf['google']['callback'] 'http://www.example.com/login/google/'
Say your downstream WSGI application is wsgi_app, you can initialize the middleware like this:
>>> from ao.social import middleware >>> app = middleware.AuthMiddleware(wsgi_app, conf)
About the User object
The ao.social middleware will put the user object in the WSGI enviromnent. You chould supply the user class when instantiating the middleware, and the user class should be available to the middleware at that time. The base user class is available as ao.social.UserBase, but you should subclass it with the ORM model of your choice to make it persistent. A basic interface is provided, but you need to implement some additional methods. Take a look at the source code and the test for a list of methods that you need to implement. (Those are the ones that raise NotImplementedError.)
To log in a user with, say, twitter, redirect to config['login_path'] % 'twitter'. Same rule applies for LinkedIn and Google. Facebook is a little different, you need to use the XFBML tags to log in, and upon a successful login, redirect to config['login_path'] % 'facebook' so that the user’s key gets added to the Beaker session and the credentials (and tokens) get stored in the user model.
For facebook, it is enough to ping the login path (i.e. make a simple AJAX call).
To post to the user’s profile, simply use user.post(message), where message is a string or unicode that contains the message to be posted. This works for Facebook (status updates), Twitter (tweets) and LinkedIn (status updates). Google Buzz is not implemented yet, as the Buzz folks haven’t added OAuth support at the time of writing this module. It will probably be supported in the future.
Django template context processors
Add the ao.social.user template context processor to your django configuration and you’ll have the user variable available in all your templates.
TODO
Better test coverage
Add support for Google Buzz
Changelog
1.0.0 (2010-04-14)
First public release
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.