AI OnAI Off
Create a class that can hold one of your custom objects. Implement the ISearchContent interface or add extension methods with the same names as the interface properties. Index each object to Find using a scheduled job or even better on some update trigger if possible.
Hook up your class to the Unified Search registry and then you're pretty much done!
http://world.episerver.com/documentation/Items/Developers-Guide/EPiServer-Find/9/DotNET-Client-API/Searching/Unified-search/
I'm carrying out a technical analysis of using EPiFind to drive our websites search page. However, we have one specific page instance which I'm finding it hard to determine if EPiFind will cope with.
We have a page type which is data-driven, there is only one page instance. Given a url oursite.com/dataDrivenPage/foo/ a bit of routing takes place to mean you only hit this data driven page and the /foo/ is a parameter which is used to select a set of images, text and blocks which are then used to populate the page.
We'd need our search to work such that if a user searched for some text that was available on /foo/ they would get a search result for this foo page. The trouble is that behind the scenes there is only a single object and the text is stored in the DB not any properties/fields.
Perhaps my main issue here is I'm not sure if there is common terminology for this type of setup making it hard to search the web/forums for relevant information. If anyone has any links or knows what words are used to refer to this type of thing, please let me know?
Otherwise, I need to figure out if EPiFind will be able to handle this somehow. I expect a web crawler would have no problem dealing with this as it only sees the end-result, so perhaps I could use the crawler connector and turn it onto the episerver website itself (i.e. not using it for indexing external websites, which seems to be what it was created for) in order to effectively index these dynamically created pages.
Or perhaps there are better solutions?
Any help or guidance as to the feasibility of this would be much appreciated.