🚧 Map Concepts between KF and FHIR

[liberaliscomputing]

This PR is part of an ongoing effort to map Kids First concepts and FHIR resources/profiles. The relevant ticket is #74.

[fiendish]

target service ID isn't known before it goes into the server. what do we do about this?

[liberaliscomputing]

Yeah, the use of TARGET_SERVICE_ID is like place-holding at the moment. As far as I understand, the 'id' field is the one used when we hit each of the REST endpoints like BASE_URL/Patient/id. So, it seems like we need to define some consistent, global scheme of 'id'. Any thoughts on this, @znatty22 and @ShahimEssaid?

[znatty22]

@liberaliscomputing

Couple general comments:

1. I think each of your python modules should not be named after FHIR base types since they are actually representing the types defined in the Phenopackets FHIR model. So for example, I would get ride of condition.py and rename it to disease.py.

2. It doesn't look like your payloads are following the FHIR payload strucutre. For example, there are lot of attributes which should NOT be first-level attributes since they are not part of the base model for the entity. They are extensions and therefore should be inside an extensions element in the payload.

3. Also, it looks like every payload is missing profiles.

Here is an example for disease.py which addresses all of this:

def disease_onset(x):
    """
    https://aehrc.github.io/fhir-phenopackets-ig/ValueSet-onset.html
    """
    pass


def disease_tumor_stage(x):
    """
    https://aehrc.github.io/fhir-phenopackets-ig/ValueSet-tumor-stage.html
    """
    pass


def condition_code(x):
    """
    http://hl7.org/fhir/R4/valueset-condition-code.html
    """
    pass


def body_site(x):
    """
    http://hl7.org/fhir/R4/valueset-body-site.html
    """


disease = {
    'resourceType': 'Condition',
    'meta: {
        'profile': ['http://ga4gh.fhir.phenopackets/StructureDefinition/Individual']
    },
    'id': DIAGNOSIS.TARGET_SERVICE_ID,
    'extension': [
        {
            'text': 'coded-onset',
            'coding': [
                {
                    'system': 'http://foo/coded-onset',
                    'code': 'bar'
                    'display': 'Pediatric onset'

                }
            ]
        },
        {
            'text': 'disease-tumor-stage',
            'coding': [
                {
                    'system': 'http://foo/disease-tumor-stage',
                    'code': 'bar'
                    'display': 'Disease Tumor Stage'

                }
            ]
        }
    ],
    'code': {
        'coding': None,  # condition_code()
        'text':	DIAGNOSIS.NAME
    },
    ...
}

[fiendish]

 'extension': [
         {
             'text': 'CodedOnset',
             'url': 'http://ga4gh.org/fhir/phenopackets/StructureDefinition/CodedOnset',
             'valueCodeableConcept': {
                 'coding': [
                     {
                         'system': 'http://purl.obolibrary.org/obo/hp.owl',
                         'code': 'HP:0410280',
                         'display': 'Pediatric onset'
                     }
                 ]
             }
         }
     ],

doesn't look like

'extension': [
        {
            'text': 'coded-onset',
            'coding': [
                {
                    'system': 'http://foo/coded-onset',
                    'code': 'bar'
                    'display': 'Pediatric onset'

                }
            ]
        },

@znatty22 ?

[znatty22]

@fiendish @liberaliscomputing

Oops, did that too hastily. The first one is correct.

[ShahimEssaid]

The "member" is missing the "family-member-phenopacket" extension shown here:

https://aehrc.github.io/fhir-phenopackets-ig/StructureDefinition-Family-definitions.html#Group.member.extension:family-member-phenopacket

Also, the new IG doesn't show the ValueSet/CodeSystme that is used in the "family-member-type" extension. I think it was shown in the older IG format.

[ShahimEssaid]

Same comment as in other files regarding "text" not being directly under Extension

[ShahimEssaid]

Same

[ShahimEssaid]

Same

[ShahimEssaid]

This isn't really necessary if there is a direct reference. See: http://hl7.org/fhir/R4/references-definitions.html#Reference.identifier

[ShahimEssaid]

Add 'system'

[ShahimEssaid]

same as above. not needed if 'reference' is available.

oops

[fiendish]

what is this type field for? "Patient" is already part of the reference url component. I think if anything the reference field should be f"{iRType}/{individuals[subject_id]['id']}". .

[fiendish]

Can everything starting with . in this file just be replaced by .* instead?