Add flatten() method to JSON-LD API

[lanthaler]

Should we add a flatten() method to JSON-LD API

[gkellogg]

I think it would be useful, and we already have the algorithm described. OTOH, it's that much more complexity to the API and something else that needs specific support.

[lanthaler]

The question is how/if we support graphs here or just flatten the merged graph. If we don't support graphs here, the result would basically be the same as the output of the subject map generation algorithm.

[dlongley]

Maybe we can support this via a specific frame, rather than an entirely new API call. For example, maybe a frame of {"@graph": null, "@embed": false} returns flattened output for the value of "@graph":

{
  "@context": ...,
  "@graph": <flattened subjects here>
}

[gkellogg]

Yes, I was thinking that flattening could be done using a trivial frame too. This looks good to me.

[lanthaler]

We could leverage framing internally but I think it would still make sense to expose an API call that returns just an array of subject definitions. Perhaps the only parameter for this new call would be which graph should be flattened (default: the merged graph). This would then be as simple as returning transforming the subject map into an array.

[lanthaler]

Which would btw. mean that everything is in expanded form which, I think, would make sense.

[jmandel]

+1 for exposing flatten functionality through the API.

Among other reasons, I want it to support a function I've prototyped called (for lack of a better term) jsonld.objectify(graph, context):

1. Flatten the graph [built-in call would be nice here]
2. Apply context
3. Replace all id references with in-memory object links

[lanthaler]

So a flatten(graph) would satisfy your needs? graph would default to the merged graph (@merged), and you could either set it to the default graph (either @graph or @default) or any other graph identifier.

[jmandel]

My use cases so far don't require named graphs, so this would be fine for me.

On Thu, May 17, 2012 at 1:19 AM, Markus Lanthaler
reply@reply.github.com
wrote:

So a flatten(graph) would satisfy your needs? graph would default to the merged graph (@merged), and you could either set it to the default graph (either @graph or @default) or any other graph identifier.

---

Reply to this email directly or view it on GitHub:
#109 (comment)

[lanthaler]

I added a flatten() method to my processor. You can test it at my playground. Let me know if that's what you had in mind.

[lanthaler]

RESOLVED: Support a 'flatten' option to .expand() and .compact(). The value of the option will be the graph to flatten and return. The default value will be false, to ensure that that the JSON-LD input document is not flattened.