See also [this thread](http://groups.google.com/group/django-developers/browse_thread/thread/ead13c9424b279a2). I'm an idiot; this snippet has now been updated to show how you do this without re-inventing the wheel.
PayPal's [https://www.paypal.com/ipn](IPN mechanism) is ridiculously easy to consume. You can tell PayPal to POST every single transaction on your account to a URL somewhere, then set up a handler at that URL which processes those transactions in some way. Security is ensured by POSTing the incoming data back to PayPal for confirmation that the transaction is legitimate. These classes are probably over-engineered, but they were a fun experiment in creating class-based generic views.
I needed this to work around a poorly configured redirecting service.
The date field in the newforms module includes American style mm/dd/yyyy , which anyone outside the US will recognise as being complete madness. This date field behaves sensibly.
A Django Form for creating a browser-based upload form that pushes files to Amazon S3. See http://developer.amazonwebservices.com/connect/entry.jspa?externalID=1434
Hook the show_url_patterns view function in to your URLconf to get a page which simply lists all of the named URL patterns in your system - useful for if your template developers need a quick reference as to what patterns they can use in the {% url %} tag.
I don't know how robust or secure this is, but it's working for me so far.
Every time I have to do this it takes me a solid half hour to figure it out, so I'm throwing it up here for future reference. I hate timezone calculations.
If your URLs need to contain numeric IDs a neat way of shortening them is to use base36.
Lots of people have asked me when the [django-openid](http://code.google.com/p/django-openid/) package will provide tools for running an OpenID provider (in addition to an OpenID consumer). The answer is "when it's done", but since it's such a common request I've decided to post some example code to help people who want to work it out for themselves. This is the openidserver.py file from [idproxy.net](http://idproxy.net/). It is **incomplete** - the urlconf, models, templates and some utility functions are not included. In other words, it's useless for anything other than providing a few hints as to how you can go about implementing a provider. Nonetheless, it's better than nothing. I hope this will prove a useful example for people trying to figure out how to best integrate the JanRain Python OpenID library with Django to build an OpenID provider.
Sometimes you just need to count things (or create unique-for-your-application IDs). This model class allows you to run as many persistent counters as you like. Basic usage looks like this: >>> Counter.next() 0 >>> Counter.next() 1L >>> Counter.next() 2L That uses the "default" counter. If you want to create and use a different counter, pass its name as a string as the parameter to the method: >>> Counter.next('hello') 0 >>> Counter.next('hey') 0 >>> Counter.next('hello') 1L >>> Counter.next('hey') 1L >>> Counter.next('hey') 2L You can also get the value as hex (if you want slightly shorter IDs, for use in URLs for example): >>> Counter.next_hex('some-counter-that-is-quite-high') 40e
I ended up not needing this (there's a good reason it's in JS and not Python, but most people would probably want to do this server-side instead) but I'm stashing it here in case I need it later. It uses jQuery for the .each() method - which is very easy to replace if you need it to work without that dependency. Usage: var src = generateChart({ "foo": 5, "bar": 3, "baz": 6 });
I needed to dynamically import a module based on a path to that file on disk, without it necessarily being on the Python Path.
I needed to sort a set of objects (a QuerySet for example) by an externally provided list of IDs - for example: >>> writers = Writer.objects.all() >>> sort_by_id_sequence(writers, [3, 1, 2]) [<Writer id: 3>, <Writer id: 1>, <Writer id: 2>]