switch the type of Element.id from id to string for R4B

[cybernop]

According to the FHIR documentation Element.id should be of type string.

StructureDefinitions with type slicing will have entries in the differential and snapshot part in the format like Observation.category:VSCat. This is valid according to the specification. But before the element was of type id which caused validation errors when parsing StructureDefinitions like this.

[nazrulworld]

@cybernop Thanks a lot for your PR. Please have a look here https://github.com/nazrulworld/fhir.resources?tab=readme-ov-file#resourceid-aka-fhirtypesid-constraint-extensibility

Here is an example of patch https://github.com/nazrulworld/fhir.resources/blob/main/fhir/resources/core/patch/patch_r4b_test.py

[cybernop]

@cybernop Thanks a lot for your PR. Please have a look here https://github.com/nazrulworld/fhir.resources?tab=readme-ov-file#resourceid-aka-fhirtypesid-constraint-extensibility

Here is an example of patch https://github.com/nazrulworld/fhir.resources/blob/main/fhir/resources/core/patch/patch_r4b_test.py

@nazrulworld Thanks for the reply. But I do not fully understand what you are trying to tell me. Do you want to tell me error is on the user side or should should I add a test case? I do not want to to change the length of Id nor do I want to create a patch for something. I just wanted to fix an error in the data type definition.

In my opinion the type of Element.id in R4B is incorrect. In the definition it is of type string and all other version (DSTU2, STU3 and R5) use the correct data type for Element.id. Only in R4B id was used incorrectly which prevents using the library in my project because of invalid parsing errors.

[cybernop]

I just want to stress that this is an issue of incorrect type and not the length of ID

[Horstage]

That's especially annoying since it's only the case for R4B, not for DSTU2, STU3 or R5.
The FHIR definition for Elements states, that Element.id is a string in all versions.

[Horstage]

Types have changed in f604fb1.
It is now StringType

[cybernop]

Updated the type to use the new StringType

[cybernop]

I want to stress how important this issue is, as in Germany we only use R4(B) at the moment and this issue prevents us to use the library for any StructureDefinition

[nazrulworld]

@cybernop This PR should be merged with the branch 7.X.X, right?

[Horstage]

I took the liberty to create a separate PR #165 for 7.x.x as it should be fixed in both 7.x.x and main, but the commits differ slightly.

[Horstage]

My last comment was incomplete, sorry.
Comparing with fhir/resources/element.py, fhirtypes.StringType | None is the correct type.

[cybernop]

Yes you are right, I adjusted my PR

[cybernop]

@nazrulworld This change should be merged into main.

PR #165 by @Horstage covers my original modification compatible to 7.X.X and should be merged into that branch.

[nazrulworld]

This new release should have been your PR included https://pypi.org/project/fhir.resources/8.0.0b4/