django-jackfrost 0.4.0

Convert your Django views into a collection of static HTML files. Or, put another way, a Django based static-site-generator with few opinions.

This is my attempt at a different type of static site renderer, instead leveraging the availability of Django's staticfiles functionality to leave specifics to someone else :) The theory is thus that you could choose a third party storage from, say, django-storages and plug it into jackfrost and have things Just Work.

I don't actually know if that's true though.

Release	Status
stable (0.4.0)
master

Alternatives

• There's django-medusa, though it doesn't appear to be active anymore.
• also django-bakery, which takes a different approach whereby one must extend specific views or models.

Differences

• Unlike django-medusa there is no autodiscovery, and no requirement that renderers go in a specific place.
• Unlike django-bakery, existing views and models ought to be usable unchanged, because of the renderer approach I've taken, which is more similar to django-medusa or Django RSS Feeds.

Dependencies

• Django
• pytest

Installing

pip installing the latest release via PyPI:

pip install django-jackfrost==0.4.0

If you want to get the latest, unstable version, you can use something like this (again with pip) I think:

pip install git+https://github.com/kezabelle/django-jackfrost.git#egg=django-jackfrost

Put jackfrost into your INSTALLED_APPS:

INSTALLED_APPS = (
    'django.contrib.auth',
    # ...
    'jackfrost',
    # ...
)

which will enable the management command:

python manage.py collectstaticsite --processes=N

Configuration & usage

Set JACKFROST_STORAGE to whatever storage backend you'd like to use, in your project's settings. By default, a subclass of django.contrib.staticfiles.storage.StaticFilesStorage which puts output into a __jackfrost directory will be used.

If your storage backend needs any arguments that can't be gleaned from individual settings, you can set JACKFROST_STORAGE_KWARGS to a dictionary of arguments to be used when instantiating the JACKFROST_STORAGE

Selecting renderers

Add a JACKFROST_RENDERERS setting, which should be a list or tuple of dotted paths to python classes or functions, much like MIDDLEWARE_CLASSES, TEMPLATE_CONTEXT_PROCESSORS etc:

JACKFROST_RENDERERS = (
    'myapp.renderers.MyModelRenderer',
    'my_other_app.utils.SomeOtherRenderer',
)

In theory, I don't care whether your JACKFROST_RENDERERS are functions or classes; if it's a class it must implement __call__. Either way, it should, when called, return a number of URL paths to be consumed.

Renderers for models

If you have a model which has a get_absolute_url method, your renderer can be as simple as:

from jackfrost.models import ModelRenderer

class MyModelRenderer(ModelRenderer):
    def get_model(self):
        return MyModel

If you need to customise the queryset, there is a get_queryset method which can be replaced. There is also a get_urls method, if you need to go totally off-reservation.

Reading from sitemaps

Giving jackfrost the dotted path to a standard Django sitemap as one of the JACKFROST_RENDERERS should do the right thing, and get the URLs out of the sitemap itself without you needing to do anything or write a new renderer.

Reading from django-medusa

In theory, giving jackfrost the dotted path to a subclass of the django-medusa BaseStaticSiteRenderer should do the right thing, and get the URLs out of the medusa renderer itself, without you doing anything. It will avoid going through the medusa rendering process, instead it'll go through mine.

Reading from Django RSS Feeds

Giving jackfrost the dotted path to a subclass of a Feed should do the right thing, and get the URLs out by asking the Feed for the item_link for everything in items, without you doing anything.

Writing a renderer

The most basic renderer would be:

def myrenderer_yielding():
    yield reverse('app:name')

or:

def myrenderer():
    return [reverse('app:name')]

Renderers may also be classes:

class MyRenderer(object):
    __slots__ = ()

    def __init__(self):
        pass

    def __call__(self):
        yield reverse('app:name')

Listening for renders

There are 8 signals in total:

• build_started is fired when the management command is run.
• reader_started is fired when a URLReader instance begins working.
• read_page is fired when a URLReader successfully gets a URL's content.
• reader_finished is fired when a URLReader instance completes.
• writer_started is fired when a URLWriter instance begins working.
• write_page is fired just after the content is written to the storage backend.
• writer_finished is fired when the URLWriter completes
• build_finished fires at the end of the management command.

Rendering on model change

Additionally, there is a listener, jackfrost.receivers.build_page_for_obj which is suitable for being used as a pre_save or post_save receiver for a Model instance, and will attempt to build just the get_absolute_url for that object.

Defining when a model may build

If a Model instance implements a jackfrost_can_build method, this is checked before building the static page. If jackfrost_can_build returns False, the page won't get built. Any other value will result in it being built.

Defining different URLs

If a Model instance implements a jackfrost_urls method, this is used instead of the get_absolute_url, and should return an iterable of all the URLs to consider building.

If the Model instance has a get_list_url method, that page will also be built. Useful for updating any ListView pages, etc.

Extras

Where possible, jackfrost will attempt to compensate for redirects (301, 302 etc) by writing an HTML page with a <meta refresh> tag pointing at the final endpoint. The template used is called 301.html.

Additionally, static pages for 401, 403, 404 and 500 errors will be built from their respective templates, if they exist. Useful if you want to wire up Apache ErrorDocument directives or whatever.

Running the tests (87% coverage)

Given a complete clone:

python setup.py test