Document model and schema

[Matt]

Hi,

As I understand the Substance document model (repo being this one) is a document model much like html has document model implementations. That is to say it is an implementation that drives the model layer of an application hosting a document. 

If I am correct in that, what I would like to get my eyes on is some sort of definition of its schema - as I read somewhere that it is JSON based, is there any way to view, in an inclusive way, what this JSON needs to look like? I assume the Substance composer adheres to that schema, but I would like to consider converting documents from another format into that schema, and for that all I have are some examples found in a downstream fork plus the ability to fiddle with saving documents and poking them as raw files. While I can generalize those examples with manual labour and low fidelity, I wonder if the Substance schema can be directly derived somehow, or at least inferred from a specific area of the code. I see a lot of object oriented javascript code spread around many modules, but not sure where to infer the schema from. Perhaps you can help with that!

I also would hope some of the code can be used to validate a document against the schema, without requiring a UI.

I looked at the compiler repo but could not find any tests there or a link to a schema spec...

Many thanks!

Matan