Fix handling of function without library for Equal

[lmd59]

Summary

CMS334 https://github.com/cqframework/ecqm-content-qicore-2024/tree/main/bundles/measure/CesareanBirthFHIR has an ELMFunctionRef without a libraryName that gets hit by interpretEqual. This makes interpretFunctionRef return an undefined which makes isOfTypeGracefulError choke.

I would prefer to fix this by making interpretFunctionRef return definitive types, but it looks like everywhere else (and now in interpretEqual) we do a check for undefined via a lovely == null check after it's called. So that's what we're doing here for now, but it would be nice to clean some of this up further in the future... should we add a ticket?
This PR also makes isOfTypeGracefulError a bit more robust.

New behavior

No longer breaks on CMS334.

Code changes

• See summary

Testing guidance

• npm run check
• npm run cli -- dataRequirements -m {path2CMS334-VS.json} (I added valuesets to the measure locally, but you can also use the command line option to fetch valuesets during calculation)

[github-actions]

Coverage report

St.❔	Category	Percentage	Covered / Total
🟢	Statements	85.57% (-0.05% 🔻)	2456/2870
🟡	Branches	73.39% (-0.02% 🔻)	2311/3149
🟢	Functions	87.95%	438/498
🟢	Lines	85.85% (-0.06% 🔻)	2372/2763

Show files with reduced coverage 🔻

St.❔	File	Statements	Branches	Functions	Lines
🟢	... / GracefulError.ts	100%	80% (-20% 🔻)	100%	100%
🟢	... / QueryFilterParser.ts	86.42% (-0.33% 🔻)	80.77% (-0.21% 🔻)	97.62%	86.19% (-0.33% 🔻)

Test suite run success

459 tests passing in 31 suites.

Report generated by 🧪jest coverage report action from 1fc6393

[elsaperelli]

I think these changes make sense for now in order to successfully run CMS334. I think making a ticket to clean up the any return type and the propRef == null statements is a good idea.

Looks like it just needs a rebase!

[hossenlopp]

This is a partial fix. The actual issue that causes the crash is in the interpretFunctionRef function. It could use an else part to handle the situation when it is a reference to a function in the same library. It should be returning a GracefulError like it does when the libraryName does exist and can't be determined what it does

[lmd59]

This is a partial fix. The actual issue that causes the crash is in the interpretFunctionRef function. It could use an else part to handle the situation when it is a reference to a function in the same library. It should be returning a GracefulError like it does when the libraryName does exist and can't be determined what it does

Yeah- that was what I was trying to say about interpretFunctionRef in the summary above. But this == null check is the way that we handle it for every other interpret* function. I think we should be consistent for now but make a ticket to overhaul interpretFunctionRef and the interpret* functions that call it so they're all handling it correctly (Note: this will probably involve a fair bit of unit test updates as well)