Building a fake structure
This is sort of a follow up to my last post (How do you setup different states that depends on EpiServer page structure?). While I like the concept the code you have to write to create the structure is slightly boring to do so i wanted to try and create something that felt a bit more fluent. I spoke a bit with Joel about this and felt inspired enough to give it a go.
The basic setup remains the same from the last post, using StructureMap to inject a IPageSource implementation to your TypedPageData class. What this post is about is how to build the structure that’s returned from the GetChildren call. Is this useful? Not sure. Was it semi-fun to code? Yes.
Building the basic structure
I really like how you build xml-files with XElement. For those of you who has not worked with that it can look something like this
1: new XElement("root",
2: new XElement("Page 1 level 1"),
3: new XElement("Page 2 level 1",
4: new XElement("Page 1 level 2")
5: )
6: );
This reads so much nicer to me than all those AddChildren (or whatever the old way of doing xml-files was. I think I’ve suppressed the memories). It’s nice that you can infer what the parent to the node is by just placing it “below” it’s parent.
Being slightly inspired by this I created a Node class that works in a similar way. The class takes a PageData object, an id of the object and an optional params array of nodes. I can now define my structure like this
1: new Node(new PageData(), 1,
2: new Node(new PageData(), 2),
3: new Node(new PageData(), 3,
4: new Node(new PageData(), 4)
5: )
6: );
Setting PageData values
Simply newing up PageData objects won’t really do us much good and I wanted to create a sort of fluent API for initializing both the default values and any other values you’d want to set on the pagedata objects.
PageDataBuilder (which is a name I came up with on my own and is in no way inspired by any other fictional / not fictional EPiServer related open source projects) is a generic class (where the generic type TTypedPageData has to be TypedPageData) with, among others, the following methods
WithDefault – returns GetDefaultPageData for the page type
WithParameters – Takes an Action<TTypedPageData> so you can can do whatever you please to your page data object
WithCallback – If your TypedPageData class implements the interface ISetDefaultValues it’s called by this method.
Usages of this class can look like this
1: new PageDataBuilder<PageType.ArticleContainer>().WithDefault().Get()
2:
3: new PageDataBuilder<PageType.Article>().WithDefault().And().WithParameters(a => a.Heading = "Article 1").Get()
4:
5: new PageDataBuilder<PageType.Article>().WithDefault().And().WithCallback().Get()
6:
You can use this in your node building like this
1: var nodeCreator =
2: new Node(new PageDataBuilder<PageType.ArticleContainer>().WithDefault().Get(), 0,
3: new Node(new PageDataBuilder<PageType.ArticleContainer>().WithDefault().Get(), 26,
4: new Node(new PageDataBuilder<PageType.Article>().WithDefault().And().WithParameters(a => a.Heading = "Article 1").Get(), 5000),
5: new Node(new PageDataBuilder<PageType.Article>().WithDefault().And().WithCallback().Get(), 5001)
6: )
7: );
If anyone is interested in this and wants to take a look at the code or anything else feel free to drop me a mail at stefan.forsberg (AT) cloudnine . se (not sure how to attach files here).
xoxo
Great post Stefan!
I think this shows great promise and I think it would be really interesting to combine it with a fluent API for setting up fake DataFactory instances.
If you e-mail the code to me (mail@joelabrahamsson.com) I'd be happy to host it on my site so people can download it.
This is so cool!
It would be great to expand it with functionality to generate larger tree structures with fake data.