FI-2313 MS Slice Subelement Support

[360dgries]

Summary

Completion of FI-2313

• metadata extraction now collects both the tail end of slices and stores it under slice_name and stores the full id (not path) under slice_id
• areas that referenced slice[:name] correctly use slice_id
• fhir_resource_navigation updated to be able to navigate slices using their discriminator.

[yunwwang]

slice_name is coming from ElementDefinition.sliceName. Example

{
      "id" : "Coverage.class:group",
      "path" : "Coverage.class",
      "sliceName" : "group",
      ...
}

But from the navigator changes, navigator may not need slice_name. So if navigator does not need slice_name, then generator can skip creating this in metadata.

[yunwwang]

This is not simply remove the slice name from the path. What need here is that if there is slice name in the middle, then navigates to the element at the specific path in a spcific slice.

Example, US Core Coverage.class

...
"class": [
  {
    "type": {
      "coding": [
        {
          "system": "http://example.com"
          "code": "local-code"
        }
      ]
    },
    "value": "abc"
  },
  {
    "type": {
      "coding": [
        {
          "system": "http://terminology.hl7.org/CodeSystem/coverage-class"
          "code": "group"
        }
      ]
    },
    "value": "def"
  },
    {
    "type": {
      "coding": [
        {
          "system": "http://terminology.hl7.org/CodeSystem/coverage-class"
          "code": "plan"
        }
      ]
    },
    "value": "ghi"
  },
]

The path class.value navigates to any one of these three elements (most likely the first occurance), while path class:group.value navigates to the 2nd element with value "def"

[360dgries]

Ah, gotchya. I had a misunderstanding of slices, and thought it could only have one. Will patch this, thanks for the catch.

[360dgries]

I've added support for this -- currently it is looking through type.coding.code to pull the slice out. I think the other slices I've looked at are also defined through this path, but let me know if you are aware of any that are not and this needs a more robust solution.

[yunwwang]

CodeableConcept.coding is an array. So here should be

code_concept.coding = [ FHIR::Coding.new.tap{ |coding|

This does not affect unit test result.

[yunwwang]

This should be added back to handle MustSupport choices

[360dgries]

Just went through and re-added any code I had commented out. Should be good now, sorry about that!