fix(backend): convert Health Connect blood glucose from mmol/L to mg/dL

[knowald]

Description

Blood glucose values synced from Android via Health Connect arrive in mmol/L (Health Connect's canonical unit for BloodGlucoseRecord), but _build_statistic_bundles stored them verbatim while the blood_glucose series unit is mg_dl. A 110 mg/dL reading was stored as 6.111.

The conversion is gated on the unit string the SDK already sends (unit starts with mmol, case-insensitive), so SDK paths that send mg/dL (the iOS SDK reads glucose with a fixed mg/dL HKUnit) and any other source already sending mg/dL are unaffected. The factor lives in a module constant MMOL_L_TO_MG_DL = Decimal("18.0182").

Tests live in TestSDKImportUnitConversion next to the existing body-fat/height conversion tests: parametrized over mmol/L / MMOL/L (converted, exact Decimal("110.1092202")) and None / "" (stored unchanged), plus a HealthKit mg/dL passthrough test.

Resolves #1119

Notes

• Conversion factor 18.0182 mg/dL per mmol/L - the standard clinical convention (inverse of the rounded 0.0555 mmol/L-per-mg/dL factor). Health Connect's own internal mg/dL converter uses exactly 18.0, so values written to HC in mg/dL round-trip with a ~0.1% offset under this factor (max ~0.25 mg/dL at 250 mg/dL). Noted in a code comment.
• Scope: this fixes the mobile SDK ingestion path only. The Apple Health XML export path (apple_xml/xml_service.py) has the same class of bug for users whose Health app displays glucose in mmol/L (Apple writes the export in the display unit, e.g. unit="mmol<180.15588...>/L"); that path never reads the record's unit attribute. Filed separately so it can be fixed for all affected series types at once rather than special-cased here.
• This touches the same block as fix: convert walking units from healthkit #1114 and the same territory as feat(backend): centralized unit conversion service (concept for #424) #1120. If fix: convert walking units from healthkit #1114 lands first I will move this case into its _normalize_unit helper (it then needs the record's unit passed in). For feat(backend): centralized unit conversion service (concept for #424) #1120: its registry needs a (blood_glucose, mmol/L) entry plus unit-token normalization, since it currently skips samples with unrecognized units.
• Existing data: mmol-stored rows are not migrated here, and value-range heuristics are not safe for glucose (genuine severe-hypoglycemia readings below 30 mg/dL exist, and mmol values can be missed above range cutoffs). The reliable repair path is re-deriving the affected samples from stored raw payloads (store_raw_payload keeps the original records incl. the unit field), matching on external_id. Happy to add that migration on request, here or as a follow-up.

Checklist

General

• My code follows the project's code style
• I have performed a self-review of my code
• I have added tests that prove my fix/feature works (if applicable)
• New and existing tests pass locally
• I have updated relevant documentation in docs/ (or no docs update needed)

Backend Changes

You have to be in backend directory to make it work:

• uv run pre-commit run --all-files passes (ruff + ty pass; prettier hook needs frontend node_modules, no frontend files touched)

Testing Instructions

Steps to test:

1. cd backend && uv run pytest tests/integrations/test_apple_sdk_import.py
2. Sync a Health Connect payload containing a BLOOD_GLUCOSE record with unit: "mmol/L" and inspect the stored sample.

Expected behavior:

Stored blood_glucose value is in mg/dL (6.111 mmol/L -> 110.1092202); HealthKit mg/dL values unchanged.

Summary by CodeRabbit

Release Notes

• New Features

  • Added automatic unit conversion for blood glucose measurements during health data imports. Values are now consistently converted to mg/dL regardless of the source unit format.

• Tests

  • Expanded test coverage to verify blood glucose import and unit conversion functionality works correctly across different input formats.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: fa2b403e-3200-4431-bc5f-0e0a5d1d8184

📥 Commits

Reviewing files that changed from the base of the PR and between 9b962ea and 923eb99.

📒 Files selected for processing (1)

• backend/tests/integrations/test_apple_sdk_import.py

🚧 Files skipped from review as they are similar to previous changes (1)

• backend/tests/integrations/test_apple_sdk_import.py

---

📝 Walkthrough

Walkthrough

Detects incoming blood_glucose records whose record.unit starts with "mmol" (case-insensitive) and multiplies their numeric value by 18.0182 to store mg/dL; tests updated to cover Health Connect conversion and HealthKit pass-through.

Changes

Blood Glucose Unit Conversion

Layer / File(s)	Summary
Conversion implementation backend/app/services/apple/healthkit/import_service.py	Adds MMOL_L_TO_MG_DL = Decimal("18.0182") and scales SeriesType.blood_glucose values when rjson.unit starts with "mmol" (case-insensitive) in _build_statistic_bundles.
Tests: unit handling and helper update backend/tests/integrations/test_apple_sdk_import.py	Extends _record helper to accept unit and adds tests that assert Health Connect (samsung) mmol/L inputs are converted to mg/dL and that HealthKit (apple) mg/dL inputs are stored unchanged.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• the-momentum/open-wearables#917: Related changes to ImportService._build_statistic_bundles and test adjustments for unit/scale handling.

Suggested reviewers

• KaliszS
• bartmichalak
• czajkub

Poem

🐰 I hopped through records, sniffed the unit line,
mmol whispered softly, "convert me fine."
Multiply by eighteen and a tad more,
mg/dL settles at the data's core.
Hooray — sweet numbers hop back to shore.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 42.86% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title mentions converting Health Connect blood glucose from mmol/L to mg/dL, which aligns with the primary change detected in the code: adding unit conversion logic for blood glucose records in the import service.
Linked Issues check	✅ Passed	The PR implements option A from issue #1119: detects mmol/L units and converts blood glucose values to mg/dL using the factor 18.0182, with comprehensive test coverage for both Health Connect and HealthKit paths.
Out of Scope Changes check	✅ Passed	All changes are scoped to implementing the blood glucose unit conversion for Health Connect and adding appropriate test coverage; no unrelated modifications detected.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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 and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🧹 Nitpick comments (1)

backend/app/services/apple/healthkit/import_service.py (1)

140-143: ⚡ Quick win

Move conversion factor to module-level constant.

The Decimal("18.0182") conversion factor is allocated on every call to _build_statistic_bundles, which processes all incoming records. Based on learnings, constant data structures should be defined at module or class level to avoid repetitive allocations in hot paths.

♻️ Suggested refactor

At module level (e.g., after imports, around line 40):

+# Unit conversion constants
+MMOL_L_TO_MG_DL = Decimal("18.0182")  # Blood glucose conversion factor
+
 class ImportService:

Then in the conversion logic:

         # Health Connect reports blood glucose in mmol/L; the series unit is mg/dL.
         if series_type == SeriesType.blood_glucose and (rjson.unit or "").lower().startswith("mmol"):
-            value = value * Decimal("18.0182")
+            value = value * MMOL_L_TO_MG_DL

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@backend/app/services/apple/healthkit/import_service.py` around lines 140 -
143, Move the hard-coded conversion factor Decimal("18.0182") to a module-level
constant and use that constant inside the _build_statistic_bundles function:
define a descriptive constant (e.g., BLOOD_GLUCOSE_MGDL_PER_MMOLL) near the top
of the module and replace the inline Decimal("18.0182") in the
SeriesType.blood_glucose conversion branch (the block that checks (rjson.unit or
"").lower().startswith("mmol")) to reference the new constant so the Decimal is
allocated once at import time rather than on every call.

Source: Learnings

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@backend/tests/integrations/test_apple_sdk_import.py`:
- Around line 376-403: The test payload in
test_healthkit_blood_glucose_stored_unchanged is missing the "source" object
which causes import_service._build_statistic_bundles to call
extract_device_info(rjson.source) with None; update the test record to include a
minimal "source" object (matching other tests) such as {"name": "...",
"bundleIdentifier": "..."} so extract_device_info receives a populated source
and the test no longer fails due to a missing attribute.

---

Nitpick comments:
In `@backend/app/services/apple/healthkit/import_service.py`:
- Around line 140-143: Move the hard-coded conversion factor Decimal("18.0182")
to a module-level constant and use that constant inside the
_build_statistic_bundles function: define a descriptive constant (e.g.,
BLOOD_GLUCOSE_MGDL_PER_MMOLL) near the top of the module and replace the inline
Decimal("18.0182") in the SeriesType.blood_glucose conversion branch (the block
that checks (rjson.unit or "").lower().startswith("mmol")) to reference the new
constant so the Decimal is allocated once at import time rather than on every
call.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 86bcf0d5-40d1-4ab9-ac46-2e83aa0c74dd

📥 Commits

Reviewing files that changed from the base of the PR and between 85bb567 and fcd33f5.

📒 Files selected for processing (2)

• backend/app/services/apple/healthkit/import_service.py
• backend/tests/integrations/test_apple_sdk_import.py

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@backend/app/services/apple/healthkit/import_service.py`:
- Line 42: Add an explicit type annotation for the module-level constant
MMOL_L_TO_MG_DL: declare it as a Decimal (e.g., MMOL_L_TO_MG_DL: Decimal =
Decimal("18.0182")) so static type checkers and IDEs recognize its type; update
the import if Decimal is not already imported and ensure the constant remains
unchanged otherwise.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 0b0447ea-ddf4-46a7-80c0-69f88e5b90f0

📥 Commits

Reviewing files that changed from the base of the PR and between fcd33f5 and 9b962ea.

📒 Files selected for processing (2)

• backend/app/services/apple/healthkit/import_service.py
• backend/tests/integrations/test_apple_sdk_import.py

🚧 Files skipped from review as they are similar to previous changes (1)

• backend/tests/integrations/test_apple_sdk_import.py