Feature: Include a profile in another profile

[ptsefton]

For the Language data commons profile we need two entry points -- describing Collections and describing Objects. These have different top-level elements: RepositoryCollection vs RepositoryObject but share a lot.

We could have Collection mode import an Object mode which would add its class definitions (exact merging behaviour would have to be worked-through)

OR we could allow a single profile to have more than one root, with an associated conformsTo)

SO this:

 "conformsToUri": [
   "https://purl.archive.org/language-data-commons/profile#Collection"
 ],
  "rootDataset": {
    "type": ["Dataset", "RepositoryCollection"]
  },

Would become something like:

  "rootDataEntity": [
   {
        "type": ["Dataset", "RepositoryCollection"],
        "conformsToUri": ["https://purl.archive.org/language-data-commons/profile#Collection"]
    },
   {
        "type": ["Dataset", "RepositoryObject"],
        "conformsToUri": ["https://purl.archive.org/language-data-commons/profile#Object"]
    }
  ],

[ptsefton]

I am preferring the second approach, but it would mean changing our existing profiles (no big deal at this point) and updating Crate-O's validation behaviour -- giving more options for adding missing root types / conformsTo URIs

[ptsefton]

Maintenance is probably the biggest concern (as pointed out by @mraadgev), so I think the best way to do this is to work with a single all-inclusive profile for each domain (eg Language Data Commons) -- this can have multiple entry points as described above so you can use it describe a Collection or an Object for example. We can then use some kind of filtering process to generate cut-down sub profiles that have the same help text etc. This could be done by example, eg input an example document and generate a sub profile that only has the stuff in it used by that example. We'll need an additional way to link these sub-profiles to the documents when we get around to this.