django-cache-purge-hooks

django-cache-purge-hooks is a reusable Django app to handle cache invalidation.

This app basically provides a mechanism to easily hook into the model callbacks to invalidate your front-end cache when needed.

Currently, only a varnish backend is implemented, but the design was created with other possible backends in mind.

Usage

Add 'cache_purge_hooks' to your INSTALLED_APPS.

#our pretend model
class Post(models.Model):
  title = models.CharField(max_length=200)
  slug = models.SlugField(unique=True,max_length=200)
  body = models.TextField(blank=True,null=True)

from cache_purge_hooks import cache_purge_hook
cache_purge_hook(Post)

As shown above, cache_purge_hooks module contains a function cache_purge_hook. Pass in as the first argument a model. The cache mechanism piggy-backs on the model instance's get_absolute_url() method and clears the backend cache for that particular url.

If any instance has multiple related urls, define a custom get_absolute_urls() method that returns a list of urls to invalidate:

class Category(models.Model):
  name = models.CharField(max_length=200)
  slug = models.SlugField(unique=True,max_length=200)
  title = models.CharField(max_length=127,blank=True)

  def get_absolute_url(self):
    return reverse("category", kwargs={
      "category": self.slug
    })

  def get_absolute_urls(self):
    gau = self.get_absolute_url()
    return [gau, reverse('blog_home'),]

In the above model, we have a blog post Category model. It's get_absolute_url() method is tied to a particular named route defined in urls.py, but any change to the category must also be reflected in the blog home page (say for example, the home page contains a tag cloud of all categories). That page must be invalidated when this information is changed.

Notice you must also explicitly grab the get_absolute_url() value. It will not get called if get_absolute_urls() exists.

If that isn't to your liking, you can also pass as a 2nd argument to purge_related_blog a function that takes an instance and returns a list of urls.

For example:

from facebook_comments.models import FacebookCommentCache
def purge_related_blog(instance):
  pr = urlparse.urlparse(instance.url)
  return [pr.path, ]

cache_purge_hook(FacebookCommentCache, purge_related_blog)

This is also useful because in the case above, the site is utilizing another reusable app (facebook_cached_comments). This is much better than hacking up a third party code to provide a get_absolue_urls().

Requirements:

• Python
• Django
• python-varnish