valueset validation

[rho-sk]

What library do you use for valuesets validation. I am looking for python implementation if possible. Thnx.

[miguelzapaton]

That's an interesting point. In the following spec (in the chapter "Annex B JSON schema")...

https://ec.europa.eu/health/sites/health/files/ehealth/docs/digital-green-certificates_dt-specifications_en.pdf

... values are coded in JSON Schema notation (draft-07) which also would allow validating data with the same standard mechanism (Ajv).

Or did I miss anything and the "ValueSets" are a new feature of JSON Schema 2020_12 and can be also validated with Ajv?

Currently it seems that the samples in the test/invalid folder are only validated by structure, aren't they?

If this is the case, what's the reason and advantage that you work with the "ValueSet" notation instead of JSON Schema data validation right now?

Cheers!

Miguel

[jschlyter]

The value sets are not part of the JSON schema, they are indirect lookup tables used since JSON schemata cannot have external references to enum values.

We initially had the the values declared as enum, but choose this path as the list of allowed values are expected to change more frequently than the schema itself. Frequent schema changes was considered more difficult to work with for implementors.

[miguelzapaton]

I see your point about frequent schema changes.

Regarding external references please feel free to have a look at my proof of concept fork to see how enum validation can be done with $ref.:

https://github.com/miguelzapaton/ehn-dgc-schema

The use of "const" combined with "oneOf" is just an alternative JSON Schema syntax for enum.

As you can see, validation fails now, when there's no corresponding vaccine manufacturer entry in valueschemas/vaccine-mah-manf.json

This would allow specifying a standard mechanism to fully verify the DGC-JSON.

Miguel

[gabywh]

ValueSets:

• more than just not [directly] part of the schema - they are generated from an EU-wide data source (JRC) which publishes e.g. which vaccines are valid / recognized, start date from when they will be recognized (can be in the future), end date by which they will be discontinued etc etc. These fields are given as active/inactive in the lookup tables (there is in some cases more than just start date / end date to determine if e.g. a vaccine is considered to be active / inactive) but active / inactive is what is of importance for us in - primarily - the verifier app. It is intended to generate the lookup tables on a 24-hour (probably early a.m. for a given EU timezone) cycle, updating such fields as active / inactive.
• separation of concerns: it should be possible to download an updated value set without having to re-install the complete application. Having the lookup tables external achieves this
• lots more information in the lookup tables that you actually would not want to repeatedly include in the QR code itself, in part due to space constraints for the QR, in part because each time a lookup value changes (e,g, goes from inactive to active) you would also need to generate a new QR code.

please feel free to have a look at my proof of concept fork to see how enum validation can be done with $ref.:

feel free to issue a pull request from your fork to this repo and we can take a look

HTH, Gaby

[gabywh]

What library do you use for valuesets validation. I am looking for python implementation if possible. Thnx.

@rho-sk you could try e.g. this https://python-jsonschema.readthedocs.io/en/stable/validate/ However, do please be aware that "validate the JSON schema" is only one half of the story - you will almost certainly also need some business rules, see e.g. https://github.com/ehn-digital-green-development/ehn-dgc-schema/wiki/FAQ#what-do-the-typical-processing-stages-look-like-for

[martin-lindstrom]

For Java-users: The https://github.com/ehn-digital-green-development/dgc-java libraries now has support for value-set validation, both "static" where files hard-wired into the jar are used and "dynamic" where you use a reloadable value-set.

[gabywh]

@rho-sk I'll close if no further comments / questions from your side?

[rho-sk]

What library do you use for valuesets validation. I am looking for python implementation if possible. Thnx.

@rho-sk you could try e.g. this https://python-jsonschema.readthedocs.io/en/stable/validate/ However, do please be aware that "validate the JSON schema" is only one half of the story - you will almost certainly also need some business rules, see e.g. https://github.com/ehn-digital-green-development/ehn-dgc-schema/wiki/FAQ#what-do-the-typical-processing-stages-look-like-for

I already implemented schema and value validation in python for that, thx.

[sanghvibk]

I already implemented schema and value validation in python for that, thx.

I am also looking for Python library to perform Valueset validation. Would appreciate if I could get some pointers.

Thank you. Bhavin