If you’re unfamiliar with trees as an abstract data type, you might want to review the concepts involved.
As a web developer, though, you probably already have a good understanding of trees as filesystem directories or paths. Wagtail pages can create the same structure, as each page in the tree has its own URL path, like so:
/
people/
nien-nunb/
laura-roslin/
events/
captain-picard-day/
winter-wrap-up/
The Wagtail admin interface uses the tree to organise content for editing, letting you navigate up and down levels in the tree through its Explorer menu. This method of organisation is a good place to start in thinking about your own Wagtail models.
It might be handy to think of the Page-derived models you want to create as being one of two node types: parents and leaves. Wagtail isn’t prescriptive in this approach, but it’s a good place to start if you’re not experienced in structuring your own content types.
Parent nodes on the Wagtail tree probably want to organise and display a browse-able index of their descendants. A blog, for instance, needs a way to show a list of individual posts.
A Parent node could provide its own function returning its descendant objects.
class EventPageIndex(Page):
# ...
def events(self):
# Get list of live event pages that are descendants of this page
events = EventPage.objects.live().descendant_of(self)
# Filter events list to get ones that are either
# running now or start in the future
events = events.filter(date_from__gte=date.today())
# Order by date
events = events.order_by('date_from')
return events
This example makes sure to limit the returned objects to pieces of content which make sense, specifically ones which have been published through Wagtail’s admin interface (live()) and are children of this node (descendant_of(self)). By setting a subpage_types class property in your model, you can specify which models are allowed to be set as children, and by setting a parent_page_types class property, you can specify which models are allowed to be parents of this page model. Wagtail will allow any Page-derived model by default. Regardless, it’s smart for a parent model to provide an index filtered to make sense.
Leaves are the pieces of content itself, a page which is consumable, and might just consist of a bunch of properties. A blog page leaf might have some body text and an image. A person page leaf might have a photo, a name, and an address.
It might be helpful for a leaf to provide a way to back up along the tree to a parent, such as in the case of breadcrumbs navigation. The tree might also be deep enough that a leaf’s parent won’t be included in general site navigation.
The model for the leaf could provide a function that traverses the tree in the opposite direction and returns an appropriate ancestor:
class EventPage(Page):
# ...
def event_index(self):
# Find closest ancestor which is an event index
return self.get_ancestors().type(EventIndexPage).last()
If defined, subpage_types and parent_page_types will also limit the parent models allowed to contain a leaf. If not, Wagtail will allow any combination of parents and leafs to be associated in the Wagtail tree. Like with index pages, it’s a good idea to make sure that the index is actually of the expected model to contain the leaf.
Your Page-derived models might have other interrelationships which extend the basic Wagtail tree or depart from it entirely. You could provide functions to navigate between siblings, such as a “Next Post” link on a blog page (post->post->post). It might make sense for subtrees to interrelate, such as in a discussion forum (forum->post->replies) Skipping across the hierarchy might make sense, too, as all objects of a certain model class might interrelate regardless of their ancestors (events = EventPage.objects.all). It’s largely up to the models to define their interrelations, the possibilities are really endless.
For going beyond the basics of model definition and interrelation, it might help to know how Wagtail handles requests and constructs responses. In short, it goes something like:
Site record will handle this request.route() method and letting each page model decide whether it will handle the request itself or pass it on to a child page.RouteResult object from route(), which identifies the page along with any additional args/kwargs to be passed to serve().serve(), which constructs a context using get_context()
serve() finds a template to pass it to using get_template()
serve() and Django responds to the requester.You can apply custom behaviour to this process by overriding Page class methods such as route() and serve() in your own models. For examples, see Recipes.
Page publishing can be scheduled through the Set schedule feature in the Status side panel of the Edit page. This allows you to set up initial page publishing or a page update in advance. In order for pages to go live at the scheduled time, you should set up the publish_scheduled management command.
The basic workflow is as follows:
Note that you have to click Publish after setting the go-live at field for the revision to be scheduled. Saving a draft revision with the go-live at field without clicking Publish will not schedule it to be published.
The History view for a given page will show which revision is scheduled and when it is scheduled for. A scheduled revision in the list will also provide an Unschedule button to cancel it.
In addition to scheduling a page to be published, it is also possible to schedule a page to be unpublished by setting the expire at field. However, unlike with publishing, the unpublishing schedule is applied to the live page instance rather than a specific revision. This means that any change to the expire at field will only be effective once the associated revision is published (i.e. when the changes are applied to the live instance). To illustrate:
Consider a live page that is scheduled to be unpublished on e.g. 14 June. Then sometime before the schedule, consider that a new revision is scheduled to be published on a date that’s earlier than the unpublishing schedule, e.g. 9 June. When the new revision goes live on 9 June, the expire at field contained in the new revision will replace the existing unpublishing schedule. This means:
Consider another live page that is scheduled to be unpublished on e.g. 14 June. Then sometime before the schedule, consider that a new revision is scheduled to be published on a date that’s later than the unpublishing schedule, e.g. 21 June. The new revision will not take effect until it goes live on 21 June, so the page will still be unpublished on 14 June. This means:
The same scheduling mechanism also applies to snippets with DraftStateMixin applied. For more details, see Saving draft changes of snippets.
© 2014-present Torchbox Ltd and individual contributors.
All rights are reserved.
Licensed under the BSD License.
https://docs.wagtail.org/en/stable/reference/pages/theory.html