Convert Health Data through Grove Instead of the App's FHIR Layer

[PSchmiedmayer]

♻️ Current situation & Problem

My Heart Counts previously owned parts of the FHIR exchange contract itself: it converted HealthKit and SensorKit records, assembled upload payloads, projected questionnaire answers into HealthKit, and maintained legacy questionnaire rendering alongside Grove's native implementation.

That duplicated the Grove FHIR Implementation Guides and allowed the app, Grove, and the published contract to evolve independently. It also left successful uploads and source-query acknowledgements insufficiently coupled: an interrupted upload or account transition could make it difficult to prove which source records had been published and which anchor could safely advance.

This migration makes Grove the FHIR implementation boundary while keeping the app responsible for collection, durable staging, account ownership, and transport. Clinical records remain lossless for both DSTU2 and R4.

It follows #202, #213, and #216: #213 carries the Grove 0.3.0-beta.8 UI adoption and #216 adds the UK consent-selection fixes. This PR carries the FHIR port. It adopts Grove SchmiedmayerLab/Grove#67, aligned with the producer contract in SchmiedmayerLab/grove-fhir#42.

⚙️ Release Notes

• Convert HealthKit and SensorKit inputs through Grove's typed FHIR exchange APIs and validate the resulting exchange graph before upload.
• Preserve exact source identity and native metadata in the governed Grove fields, including full SensorKit Device Usage payloads rather than an app-defined summary.
• Stage SensorKit source evidence, the FHIR Bundle, and the acknowledgement state as one durable publication batch; advance the source anchor only after the matching upload succeeds.
• Fence staged exchanges to the signed-in account generation and captured Firestore destination so sign-out or account replacement cannot retarget an in-flight upload.
• Use deterministic identifiers and retry digests so retries reproduce the same exchange instead of creating duplicate clinical resources.
• Project questionnaire quantities and blood-pressure groups through the native Grove Questionnaire model, including UCUM validation and one HealthKit correlation for the two blood-pressure components.
• Replace the legacy Grove Questionnaire and ResearchKit-on-FHIR rendering path with native Grove Questionnaire views and composition.
• Keep DSTU2 and R4 clinical-record bytes and their declared FHIR release intact through the Grove transport document.
• Remove app-local FHIR extensions and conversion code that are now authoritative in Grove.
• Derive every identifier system from https://myheartcounts.stanford.edu/fhir, and bundle a known enrollment as ResearchStudy, PlanDefinition (…/PlanDefinition/<study uuid>|<study revision>) and ResearchSubject, with the study UUID in lower case.
• Store every HealthKit graph and retraction under its sample's UUID, so Firebase ingestion keys them as before.
• Retract deleted samples through Grove's HealthKitConverter.retraction(for:context:retractedAt:), and keep each deletion's lower bound (deletedAfter, the previous query's start) for Grove Proof-of-concept: on-device LLM generated nudges #94's retraction period.
• Log refusals with their registry code, and conversion warnings per exchange graph.
• Render questionnaires from the merged multilingual resource in the participant's language, and export responses with language through renderedIn:; questionnaire time functions run on an explicit clock.

📚 Documentation

The implementation follows the Grove FHIR Implementation Guides and keeps contract reasoning next to the conversion and staging boundaries. The branch pins the exact development revisions used for verification: Grove 66f0a88e, Study Definitions 490ffab0 (SchmiedmayerLab/MyHeartCounts-StudyDefinitions#47) and Firebase 12f2992c (SchmiedmayerLab/MyHeartCounts-Firebase#217).

✅ Testing

• At e151f6ea: SwiftLint strict passes, the app unit tests pass (95 tests in 25 suites), MyHeartCountsShared passes on macOS (13) and on Linux with Swift 6.4 (5), and the HealthDashboard, TaskHandling and survey-extraction UI tests pass against the Firebase emulators (12 run, 2 skipped by the tests themselves).
• Earlier: the complete MyHeartCounts Unit Tests plan passed locally with 51 tests across 15 suites.
• The HealthSampleProcessing suite also passes across five repeated executions of its persistence tests.
• swift test --package-path MyHeartCountsShared passes locally: 14 tests across 4 suites.
• The Study Definitions package passes locally: 2 tests, including questionnaire-catalog conformance.
• Repository-wide swiftlint lint --strict and git diff --check pass.
• The Grove GroveHealthKitFHIR target passes locally: 123 tests across 16 suites.

Targets #216 in the stack. Merge conflicts will be resolved separately, including reconciling the FHIR StudyDefinitions dependency with the UK consent placeholder from SchmiedmayerLab/MyHeartCounts-StudyDefinitions#48. No branch merge or rebase was performed when inserting #216. This pull request remains a draft.

Code of Conduct & Contributing Guidelines

By creating and submitting this pull request, you agree to follow our Code of Conduct and Contributing Guidelines:

• I agree to follow the Code of Conduct and Contributing Guidelines.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Advanced

Run ID: 14280b28-f97b-4ad2-a54f-0c2efa21b2f4

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}