Null handling by ProperContains, ProperIncludedIn, ProperIncludes, ProperIn evaluators. More tests for arithmetic functions.

[antvaset]

This fixes how the ProperContains, ProperIncludedIn, ProperIncludes, ProperIn evaluators handle nulls. The CQL specification states that:

• The T properly included in List<T> expression (ProperIn(T, List<T>) in ELM) evaluates to null if either argument is null (link).
• Same for List<T> properly included in List<T> (ProperIncludedIn(List<T>, List<T>) in ELM).
• List<T> properly includes T (ProperContains(List<T>, T) in ELM) evaluates to null if either argument is null (link).
• Same for List<T> properly includes List<T> (ProperIncludes(List<T>, List<T>) in ELM).

Also adding new tests for arithmetic functions.

[codecov]

Codecov Report

Attention: Patch coverage is 85.71429% with 1 line in your changes missing coverage. Please review.

Project coverage is 63.74%. Comparing base (179c762) to head (2601a5a).

Files	Patch %	Lines
.../engine/elm/executing/ProperIncludesEvaluator.java	66.66%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #1370      +/-   ##
============================================
+ Coverage     63.69%   63.74%   +0.04%     
  Complexity     2666     2666              
============================================
  Files           492      492              
  Lines         27738    27718      -20     
  Branches       5511     5503       -8     
============================================
- Hits          17669    17668       -1     
+ Misses         7828     7820       -8     
+ Partials       2241     2230      -11     

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

[JPercival]

@antvaset - We have a goal of >=80% test coverage for new/updated code. Can you add tests for the Evaluators that gets the patch up to 80%?

[antvaset]

Hi @JPercival, I've squeezed in some more coverage. This is the most we can do I believe.

[brynrhodes]

The new ArithmeticFunctionsTests are good, but the changes to the "proper" operators are incorrect. List and interval membership operations throughout the language use "equality semantics, with the exception that nulls are considered equal for the purposes of membership". See the "ProperContains" and "ProperIn" operator descriptions in the Logical Specification: https://cql.hl7.org/04-logicalspecification.html#proper-contains

[antvaset]

Thanks for the clarification @brynrhodes. I'm closing this because the fixes to the evaluator logic are incorrect. Also linking here the issue in JIRA https://jira.hl7.org/browse/FHIR-46283 and the pull request with the salvaged arithmetic function tests https://github.com/cqframework/cql-tests/pull/33/files.