FI-2460: Support for HL7 validator wrapper #488
Conversation
There was a problem hiding this comment.
This seems to be working well for me, but I expect your FHIR core PR will need to get in before we could use this in production. We should see what @arscan thinks as well. I think the next steps are something like the following:
- Make a subclass of the g10 suite, change its id, and have one use the current validator and one use the new validator so that we'll be able to support both during a transition period. I'm not sure whether we'll want the "real" g10 suite to use the new validator to ensure it gets broad testing or the current one to minimize any potential disruption. I guess this would need two version of the configuration checker as well.
- Update the US Core and SMART App Launch test suites to use the new validator so that we'll be ready to completely cut over and remove the current validator from this repo.
| case (us_core_version_requirement[:us_core_version]) | ||
| when G10Options::US_CORE_3 | ||
| igs('hl7.fhir.us.core#3.1.1') | ||
| us_core_message_filters = USCoreTestKit::USCoreV311::USCoreTestSuite::VALIDATION_MESSAGE_FILTERS |
There was a problem hiding this comment.
I don't understand why this assignment was changed.
There was a problem hiding this comment.
It just felt strange to have the igs line in the middle of the case when the result of that case is assigning to a separate variable. I'm probably going to revert this anyway because I suspect there's a better way to define the IG for each option.
There was a problem hiding this comment.
Having the igs lines there wouldn't interfere with the original assignment style.
|
How will the "two suites" approach look from a user perspective? Will we show the two suites on the start page and the user has to pick one? |
We'll want to discuss that with @arscan |
dehall
changed the title
* FI-2532: add feature flag to switch between inferno validator and hl7 validator wrapper * standardize name * default hl7 validator to false and comment out its services
dehall
force-pushed
the
fi-2460-hl7-validator
branch
from
6364702 to
e05dab1
Compare
This PR enables support for the the g10-test-kit to use the original Inferno validator wrapper or the HL7 validator wrapper, toggled by the env var
USE_HL7_RESOURCE_VALIDATOR. From the user perspective, there should be no visible change. If I've done this right then it should be plug & play.Summary of changes:
validatorblock in test suite to a conditional that selectsvalidatororfhir_resource_validatorbased on a new env varUSE_HL7_RESOURCE_VALIDATOR. Addigssetting to this block if using the HL7 validator.cli_contextblock to specify validator settings:txServer nildisables use of a TX server. Aligns withDISABLE_TXenv var in the inferno validatordisplayWarnings truesets any "display issue" messages to be warnings instead of errors. Aligns withDISPLAY_ISSUES_ARE_WARNINGSenv var in the inferno validatordisableDefaultResourceFetcher truedisables fetching and validating against any unknown profiles found in a resource. (This is now a default in inferno-core, not yet released though. No harm in keeping this setting here temporarily or forever)New error messages that need to be ignored -- there are two messages that occur per usage of a mime type (CapabilityStatement, DocumentReference, some uses of DiagnosticReport)
