Split exported MARC schemas

[nichtich]

The current result of export-schema-files into marc-schema.json is a mix of field definitions from several sources. There is at least one case of clash of definitions (Tag997.java in both bltags and kbrtags) leading to a contradictory schema. The Avram export should be split into:

• Common MARC fields
• Custom or additional fields and rules for individual libraries/catalogs
  • b3kat
  • fennica
  • nkcr
  • szte
  • bl
  • dnb
  • gent
  • kbr
  • oclc

[pkiraly]

I have two questions:

• should the JSON have the following top level structure?

{
  "common": {. . .},
  "b3kat": {. . .},
  "fennica": {. . .},
  . . .
}

• there are non field level custom data elements, such as subfields. Should we also separete them? It might be problematic in some cases, e.g. KBR introduced multilingual and numbering subfields for all fields, so we should replicate almost the whole list of fields.

[nichtich]

There should be one valid Avram file for each institution with custom elements, e.g. marc-bibliographic.json, marc-b3kat.json...

The files can either be complete (merge of standard fields and custom fields) or only contain custom fields or fields modified by custom additions. I'd prefer the first to not duplicate data and to have a better overview of custom fields but that's a matter of taste. I created a basic HTML-to-Avram script (call avram-js with option -d html), see https://gbv.github.io/avram-js/ - such tables could be generated for each customization from the schema files.