Make clearer distinction between project-level resources and their products

[caufieldjh]

This registry covers more than KGs: it includes the code used to make them, the relevant data models, mappings, and beyond.
In OBO Foundry, there's a distinction between an ontology and its products, such that a single ontology has multiple products.
e.g., ENVO has multiple products (https://obofoundry.org/ontology/envo.html) including a few different serializations and subsets.

Any KG or graph may have the following:

• The output graph, in multiple formats
• A graph dump (e.g., from Neo4J)
• A hosted endpoint for the graph or some other interface
• The code used to produce the graph
• The framework used for the graph (not a product, but a relationship)
• The data model used in the graph (not a product, but a relationship)
  and beyond.

These may be better modeled in a way like how OBO Foundry does usages, so distinctions between these products and related resources can be clearly specified and presented in-line.

[justaddcoffee]

Does this distinction not already exist? (That is, how do you think this should be improved?)

For example, here's the Monarch KG, which contains a link to the repo https://github.com/monarch-initiative/monarch-ingest and also the products list (which only has one thing, but could have a link to all the various Monarch KG builds)

---
layout: ontology_detail
activity_status: active
id: kg-monarch
title: KG Monarch
description: Monarch Initiative Knowledge Graph
domain: health
preferredPrefix: http
contact:
  orcid: 0000-0002-6601-2165
  github: cmungall
  email: [email protected]
  label: Christopher J. Mungall
homepage: https://kghub.org/kg-monarch/index.html
tracker: https://github.com/monarch-initiative/monarch-ingest/issues
repository: https://github.com/monarch-initiative/monarch-ingest
products:
- id: kg-monarch.tar.gz
  format: kgx
  title: KGX Distribution of KGM
  description: KGX Distribution of KGM
  ontology_purl: https://kg-hub.berkeleybop.io/kg-monarch/current/kg-monarch.tar.gz
uri_prefix: https://w3id.org/kg-monarch/
[snip]

[caufieldjh]

It's primarily a matter of:

• Formalizing it in the schema. The existing schema is fairly loose and largely populated by string fields.
• Expanding the scope to capture the product types, and specifically the distinction between code and usable graph data.
  So yes, the distinction exists, but it isn't specific enough to support re-use (and I'd argue that the distinction between product types could be clearer in OBO Foundry, too, but that's a different project)