FI-2750: Add FHIRPath Support

[vanessuniq]

Summary

This PR introduces FHIRPath support, enabling evaluation of FHIRPath expressions on FHIR resources using an external FHIRPath service. The new functionality allows tests to evaluate expressions and receive results from a configurable FHIRPath evaluator.

[codecov]

Codecov Report

Attention: Patch coverage is 89.13043% with 5 lines in your changes missing coverage. Please review.

Project coverage is 79.86%. Comparing base (586d3b8) to head (c5d67dc).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
lib/inferno/dsl/fhirpath_evaluation.rb	91.89%	3 Missing ⚠️
lib/inferno/exceptions.rb	66.66%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #520      +/-   ##
==========================================
+ Coverage   79.83%   79.86%   +0.03%     
==========================================
  Files         252      253       +1     
  Lines       13305    13351      +46     
  Branches     1289     1290       +1     
==========================================
+ Hits        10622    10663      +41     
- Misses       1934     1939       +5     
  Partials      749      749              

Flag	Coverage Δ
backend	91.84% <89.13%> (-0.04%)	⬇️
frontend	74.54% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Jammjammjamm]

This PR shouldn't be updating package-lock.json since it doesn't include any front end changes.

[Jammjammjamm]

An array of what?

[Jammjammjamm]

Is there a reason why test writers would need to have multiple fhirpath services? I was really hoping this would just work out of the box without having to define fhirpath services.

[vanessuniq]

updated

[Jammjammjamm]

Would it be possible/desireable to instantiate the element from FHIR::Models rather than than using the JSON representation of the element?

[vanessuniq]

The result types of a FHIRPath expression are unpredictable as they depend on the specific query or expression being evaluated. The element value can be anything from a primitive type to a complex FHIR data type. element is either a primitive value or a JSON representation of a FHIR element. Transforming element into a FHIR::Model only makes sense when the returned value is consistently a FHIR element, which is not always the case.

[Jammjammjamm]

But why doesn't representing that value as a FHIR model make sense, while representing it as JSON does make sense?

[vanessuniq]

The value of element is not always a json. I should update the description of the element key because it is confusing. The description should be something like primitive value or JSON representation of FHIR element or simply the result value of the FHIRPath expression.

These are the examples of results depending on the expression we are evaluating:

path = Patient.gender='female'
result =  [ {"type": "boolean", "element": true }]

path = Patient.name.given
result = [
    {
        "type": "string",
        "element": "patient"
    },
    {
        "type": "string",
        "element": "sample"
    }
]

path = Patient.name
result = [
    {
        "type": "HumanName",
        "element": {
            "use": "official",
            "family": "Chalmers",
            "given": [
                "patient",
                "sample"
            ]
        }
    }
]

I am not sure if you want to add condition handling to transform to FHIR model when the value is a fhir element.

[Jammjammjamm]

Right, it isn't always JSON. You always have to check the type.

We use fhir_models widely throughout Inferno, so given these two options:

• Return a primitive or a Hash for complex types
• Return a primitive or a FHIR model instance for complex types

I'm wondering if there are technical reasons not to return a FHIR models instance here.

We generally expect (and encourage) test authors to use FHIR models to represent resources (e.g. in the first parameter of evaluate_fhirpath), so why not use them here?

[vanessuniq]

updated