Emit the SensorKit Structured Summaries and Declared Payload Formats

[PSchmiedmayer]

♻️ Current situation & Problem

Three SensorKit streams had structured profiles; the rest were admitted only as opaque recordings. A receiver could store the bytes but could not tell how many samples arrived, over what interval, or in what format — the recording DocumentReference carried a media type and nothing else, so application/json could have been any shape at all.

⚙️ Release Notes

• Emit the six new platform-exclusive structured summaries the v0.3.0 guide publishes: messages usage, phone usage, keyboard metrics, sleep session, accelerometer, and photoplethysmography. Each carries content-free counts and durations only; no typed content, contact identity, or sentiment is ever represented.
• Require a declared registry payload format on every recording document. content.format now names the exact registry entry, is validated against the formats its stream admits, and an unregistered code fails closed.
• Mandate the native recording where a summary is lossy (keyboard metrics, accelerometer, PPG) and leave it optional where the summary is complete.
• Carry rawFormats through the generated contract so a producer reads the registry code from the catalog rather than hardcoding it.
• Give the sleep session the exact length of its own interval. The profile previously forbade both a value and a data-absent reason while inheriting the shared rule that a mobile Observation must carry one of them, so it could never validate. SensorKit publishes only the session bounds, so the length of those bounds is the one result available without inventing a stage or a quality score.
• Identify the conformance corpus with Grove testing naming systems instead of study.example.org, which the official Validator rejects outside published examples.
• Re-pin to the v0.3.0 guide cut and regenerate.

📚 Documentation

The payload formats are specified in the sensor guide's Recording formats page, generated from the machine registry; the structured profiles are published in the SensorKit guide.

✅ Testing

• One conversion test per new stream covering profiles, components, effective periods, derivedFrom wiring, output identity, and format codings, plus fail-closed tests for zero durations, negative counts, and unregistered formats.
• Identity vectors extended with the six new output discriminators; the fixture corpus grew to 12 bundles.
• swift build, SwiftLint strict, the package test suite, and the generator --check gates pass, along with the HealthKit suite carried in by the merge.

[Copilot]

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

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 089ad708-e8b6-4ee2-9df8-61e50a37facb

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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.}

[lukaskollmer]

imo the naming here (ie, on this particular type, and also in the rest of the file and the PR) is too specific. we're in the Grove target already, so imo we should drop the Grove prefix from the types and call it eg just SensorFHIRConverter

[lukaskollmer]

most of the types in this file feel like they should be nested elsewhere? (eg the Context, the Conversion, the ConversionError, ...)