nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Implementing a deep object graph with TS7 and LSCS

thomasprommer

Forum Group,

I am interested in hearing people's different implementation approaches to facilitating a true deep object graph in TeamSite using LSCS.

Scenario:

Let's assume the domain model is properly normalized and entities are broken apart into reusable content types:

-> Car Page (holds rendering information about the car page, e.g. featured image etc.)
-> Car (holds core data properties about the car, e.g. construction year etc.)
-> Manufacturer (holds core data properties about the manufacturer, e.g. HQ location etc.)

Now we want to be able to query LCSC to get a car page back including all assigned references and their properties.

<carPage>
   <car>
      <manufacturer/>
   </car>
</carPage>

What are people's thoughts on how to best implement such needs. The approaches we are currently evaluating are:

- Use presentation templates (e.g. xslt includes) to resolves path references and include child data into the ship page entity.
- Use programatic tagging of content to store relationship rather than DCR references. In this case the query would be based on a car ship page ID tag and follow a flat rather than a nested structure.

Are there any other approaches we should consider when aiming to implement the task at hand?

Thank you for your feedback. /Thomas

Find more posts tagged with

Comments

Crsb

I don't understand the presentation template approach. But I think these requirements fit more into the tagging and LSCS query approach, a similar approach would be to use a combination of plain Java, hibernate and an actual database to meet these requirements without making the authoring process too complex for content authors.

thomasprommer

I don't understand the presentation template approach. But I think these requirements fit more into the tagging and LSCS query approach, a similar approach would be to use a combination of plain Java, hibernate and an actual database to meet these requirements without making the authoring process too complex for content authors.

Thanks for the input Crsb.

Within the presentation template approach all references would be stored as DCR paths and the presentation template would use xsl:include statements to fetch and nest the XML content for each reference. This transformation could happen before or after the LSCS query, depending if LSCS allows querying "transformed" DCRs.

The plain Java approach is a valid option as well although we consider it too much of a disconnect from the general TeamSite architecture principals to walk down that road.

santosh013

As per my understanding the best approach is to use Database and it will be the beauty of database design which will show how you are able to fetch contents as per the requirement in your components.
Never the less that the DCT and DD should also be robust to cater your needs as you might be required to update more than one table at a given time.

HTH.

Crsb

Thanks for the input Crsb.
The plain Java approach is a valid option as well although we consider it too much of a disconnect from the general TeamSite architecture principals to walk down that road.

I don't see how it is disconnected from TeamSite architecture principals. You would use DCRs and DD to deploy the data to DB. and you will use LS external components and hibernate or plain java to fetch the data, then the components will process/transform the data. it is similar to deploying the DCRs to LSCS and executing LSCS queries but without the need for too much tagging. It just depends on how complex the data is and the relationship between different pieces of the data, i wouldn't want the user to maintain 100s or even 10s of different tags on each DCR.

thomasprommer

I don't see how it is disconnected from TeamSite architecture principals. You would use DCRs and DD to deploy the data to DB. and you will use LS external components and hibernate or plain java to fetch the data, then the components will process/transform the data. it is similar to deploying the DCRs to LSCS and executing LSCS queries but without the need for too much tagging. It just depends on how complex the data is and the relationship between different pieces of the data, i wouldn't want the user to maintain 100s or even 10s of different tags on each DCR.

Crsb,

I appreciate your thoughts. In your scenario how would references between entities be selectable in the DCT and stored within the DCR?

Thanks

Crsb

As you said, these are just thoughts, I am just thinking here with the few lines of problem description in your initial post.

In terms of how to establish relationships between the different pieces of data, the first things that comes to mind are callserver() and datasources to populate certain items such as a list of manufacturers to select from when creating a car DCR. this means that the manufacturer data is already created and available before creating cars. This should create the relationship between cars and manufacturers and this gets deployed to DB. Now when it comes to rendering the data, i would create a precontroller at the page level or if possible at the site level (if this is a micro-site dedicated to one car model), the precontroller would know what car it is based on site name or page name and fetch the data from the DB (one or multiple tables) and put into PageScope map (e.g PageScope[featuredimage], PageScope[manufacturer]...) where the component has something like < img src="$PageScope[featuredimage]" .../>.