Investigate FHIR bundle equality methods and bundle enhancements

jack-h-wang commented 5 months ago

User Story

As a developer, I want to understand library behavior, so that logical tests function as expected.

Description/Use Case

The HAPI FHIR library implements the bundle.equalsDeep(x) and bundle.equalsShallow(x) methods that test equality of a FHIR bundle against another object. It appears the methods will often return false for bundles that are logically equal, but differ in that bundles created with the same input may have differing UUIDs or creation timestamps. This renders the methods unusable for any practical purpose. We should document why this behavior occurs and determine if changes to the library are needed.

Many instances where we compare bundles for equality currently instead use the CompareFhirData class, which is designed to take these differences into account.

We should also consider if the bundle enhancements we add to FHIR bundles in HL7toFhirTranslator.translate adjust them in a way that makes these methods unworkable.

Risks/Impacts/Considerations

Dev Notes

FhirTransformer.checkForEquality currently relies on the equalsDeep method.
The FhirConverterTests.test getContentFromHL7 alternate profile documents a use of equalsDeep that was expected to be true, but always returns false. equalsShallow also does not work for this test. It should be noted the bundles in this test receive bundle enhancements.
We should re-evaluate whether the bundle enhancements are still necessary or need to be adjusted to account for the HAPI FHIR library's defintion of equality.

Acceptance Criteria

The behavior of the bundle.equalsDeep(x) and bundle.equalsShallow(x) methods are understood and current uses of the methods in code are evaluated.
The current bundle enhancements have been evaluated and adjusted if necessary.

Andrey-Glazkv commented 3 months ago

Hey team! Please add your planning poker estimate with Zenhub @adegolier @arnejduranovic @brick-green @david-navapbc @jack-h-wang @jalbinson @JFisk42 @mkalish @thetaurean

arnejduranovic commented 1 month ago

TODO: The two AC should be split up. Verify we don't already have have a ticket for 2nd AC. For first AC, probably just wait for a more concrete use-case to appear so the ticket is more defined with AC.

arnejduranovic commented 1 month ago

Second AC is part of this ticket: https://github.com/CDCgov/prime-reportstream/issues/14117

CDCgov / prime-reportstream