Closed mvahowe closed 3 years ago
Relates to #139
To progress this I think we need @FoolRunning to tell us where we might expect multiple values, rather than the single value assumed in the current schema.
Sorry, I forgot about this. In Paratext, the following settings can contain multiple entries/values:
Also, the bookSourceForMarkerXt and bookSourceForMarkerR settings have values that are an enumeration: LongName, ShortName, or Abbreviation (and can not be blank/null).
Also (again), the referenceExtraMaterial and referenceFinalPunctuation values are optional and can be blank/null - in which case validation fails if those things were to be specified in a reference. Since we're not doing validation, it might be possible to just ignore these two settings.
@FoolRunning Thanks! When you say "not doing validation", you mean against the versification for that entry? If so, I agree that we're not doing it. (We do need to validate the syntax of the references.)
@mvahowe, I thought this was just to get the loc attributes set. If so, then we just have to be able to parse the reference, right?
This PR provides a schema and example document for USFM reference parsing settings as JSON.
The schema is underspecified in terms of what should be required and how those fields may be constrained (eg I suspect most of them should be non-alphanumeric characters).
Tests should pass because they don't actually test the schema or test document, yet.