Merge pull request #262 from MBombeck/release/v1.12.5

release/v1.12.5

CHANGELOG.md

@@ -2,6 +2,13 @@
 
 ## [Unreleased]
 
+## [1.12.5] — 2026-06-05 — correct sleep summary and readiness contributors
+
+### Fixed
+
+- **The sleep summary now reflects the whole night, not a single stage.** The Insights sleep figure and prose were reading one sleep stage (~16 minutes) instead of the reconstructed night total; they now use the night's time-asleep, the same number the dashboard and charts already show.
+- **Readiness contributors no longer read a misleading 0.** Resting heart rate, heart-rate variability, and respiratory rate are hidden (rather than shown as 0) when there isn't yet enough history to establish a baseline; a genuine low sub-score still shows.
+
 ## [1.12.4] — 2026-06-05 — a calmer, more consistent Insights
 
 ### Changed

docs/api/openapi.yaml

@@ -1,7 +1,7 @@
 openapi: 3.1.0
 info:
   title: HealthLog API
-  version: 1.12.4
+  version: 1.12.5
   description: >-
     Self-hosted personal-health-tracking PWA — public API surface for the iOS native client and external ingest.
 

package.json

@@ -1,6 +1,6 @@
 {
   "name": "healthlog",
-  "version": "1.12.4",
+  "version": "1.12.5",
   "description": "Self-hosted personal-health-tracking PWA with Withings integration, AI insights, and doctor-report PDF export.",
   "license": "AGPL-3.0-only",
   "homepage": "https://healthlog.dev",

src/lib/insights/__tests__/metric-status.test.ts

@@ -22,7 +22,20 @@ vi.mock("@/lib/insights/metric-correlation-context", () => ({
   getRelevantCorrelationsForMetric: vi.fn().mockResolvedValue([]),
 }));
 
+vi.mock("@/lib/tz/resolver", async () => {
+  const actual = await vi.importActual<typeof import("@/lib/tz/resolver")>(
+    "@/lib/tz/resolver",
+  );
+  return { ...actual, resolveUserTimezone: vi.fn() };
+});
+
+vi.mock("@/lib/rollups/measurement-read", () => ({
+  loadUserSourcePriority: vi.fn(),
+}));
+
 import { prisma } from "@/lib/db";
+import { resolveUserTimezone } from "@/lib/tz/resolver";
+import { loadUserSourcePriority } from "@/lib/rollups/measurement-read";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
 import { getRelevantCorrelationsForMetric } from "@/lib/insights/metric-correlation-context";
 import { formatPreviousContextForPrompt } from "@/lib/insights/memory";
@@ -68,6 +81,60 @@ beforeEach(() => {
   // are no-ops unless a test overrides them.
   vi.mocked(getRelevantCorrelationsForMetric).mockResolvedValue([]);
   vi.mocked(formatPreviousContextForPrompt).mockReturnValue("");
+  vi.mocked(resolveUserTimezone).mockResolvedValue("Europe/Berlin");
+  vi.mocked(loadUserSourcePriority).mockResolvedValue(null as never);
+});
+
+describe("generateMetricStatus — SLEEP_DURATION night reconstruction (iOS E2)", () => {
+  it("feeds the snapshot the per-night time-asleep total, not a single stage", async () => {
+    vi.mocked(prisma.auditLog.findFirst).mockResolvedValue(null as never);
+    vi.mocked(prisma.user.findUnique).mockResolvedValue({
+      dateOfBirth: null,
+      gender: null,
+    } as never);
+    vi.mocked(prisma.auditLog.create).mockResolvedValue({
+      createdAt: new Date("2026-06-04T07:00:00.000Z"),
+    } as never);
+
+    // One overnight session of granular stages, each row ONE stage (minutes).
+    // Total time asleep = CORE + DEEP + REM = 300 + 90 + 75 = 465 min (7.75 h).
+    // A single stage (DEEP, 90 min) is the largest, and an old per-row read
+    // would have surfaced the LATEST stage (~75 min REM) as "current sleep".
+    const wake = new Date("2026-06-04T06:30:00.000Z");
+    const m = 60 * 1000;
+    const stageRows = [
+      { value: 300, measuredAt: new Date(wake.getTime() - 90 * m), sleepStage: "CORE", source: "APPLE_HEALTH" },
+      { value: 90, measuredAt: new Date(wake.getTime() - 30 * m), sleepStage: "DEEP", source: "APPLE_HEALTH" },
+      { value: 75, measuredAt: wake, sleepStage: "REM", source: "APPLE_HEALTH" },
+    ];
+    vi.mocked(prisma.measurement.count).mockResolvedValue(
+      stageRows.length as never,
+    );
+    vi.mocked(prisma.measurement.findMany).mockResolvedValue(
+      stageRows as never,
+    );
+
+    const capture = { systemPrompt: null, userPrompt: null } as {
+      systemPrompt: string | null;
+      userPrompt: string | null;
+    };
+    stubCompletion("Summary: solide Nacht.", capture);
+
+    const res = await generateMetricStatus({
+      metric: "SLEEP_DURATION",
+      userId: "u1",
+      locale: "de",
+      force: true,
+    });
+
+    expect(res.text).toBeTruthy();
+    expect(capture.userPrompt).toBeTruthy();
+    const prompt = capture.userPrompt as string;
+    // The snapshot must carry the NIGHT total (465 min), never a lone stage.
+    expect(prompt).toContain("465");
+    expect(prompt).not.toContain('"value": 75');
+    expect(prompt).not.toContain('"value": 90');
+  });
 });
 
 describe("metric-status registry", () => {

src/lib/insights/derived/__tests__/readiness.test.ts

@@ -168,6 +168,52 @@ describe("computeReadiness gating", () => {
     }
   });
 
+  it("emits null (not 0) for a vital with readings below the baseline floor (iOS F2)", async () => {
+    // The user HAS recorded RHR + HRV + respiratory, but only on 2 distinct
+    // days each — below the baseline-band history floor. The contributor must
+    // be NULL (dropped → iOS hides it), never 0 (which would read as a real
+    // worst-case sub-score). Mood + sleep carry the headline so the blend
+    // still reaches the minimum-components floor.
+    const DAY_MS = 24 * 60 * 60 * 1000;
+    const sparse = (type: string, base: number) =>
+      Array.from({ length: 2 }, (_, i) => ({
+        value: base,
+        measuredAt: new Date(NOW.getTime() - (i + 1) * DAY_MS),
+        type,
+      }));
+    moodFindMany.mockResolvedValue(
+      Array.from({ length: 8 }, (_, i) => ({
+        score: 4,
+        moodLoggedAt: new Date(`2026-05-${String(20 + i).padStart(2, "0")}T20:00:00Z`),
+      })),
+    );
+    measurementFindMany.mockImplementation(
+      async (args: { where: { type: string } }) => {
+        const type = args.where.type;
+        if (type === "RESTING_HEART_RATE") return sparse(type, 58);
+        if (type === "HEART_RATE_VARIABILITY") return sparse(type, 65);
+        if (type === "RESPIRATORY_RATE") return sparse(type, 14);
+        if (type === "SLEEP_DURATION") {
+          return [
+            { value: 420, measuredAt: new Date("2026-05-31T06:00:00Z"), sleepStage: "ASLEEP" },
+            { value: 430, measuredAt: new Date("2026-06-01T06:10:00Z"), sleepStage: "ASLEEP" },
+            { value: 410, measuredAt: new Date("2026-06-02T06:05:00Z"), sleepStage: "ASLEEP" },
+          ];
+        }
+        return [];
+      },
+    );
+    const result = await computeReadiness("u1", PROFILE, { now: NOW });
+    expect(result.status).toBe("ok");
+    if (result.status === "ok") {
+      for (const key of ["rhr", "hrv", "respiratory"] as const) {
+        const c = result.value.components.find((x) => x.key === key)!;
+        expect(c.value).toBeNull();
+        expect(c.value).not.toBe(0);
+      }
+    }
+  });
+
   it("provenance source is 'live' for a blend backed only by sleep + mood", async () => {
     // No vital readings (RHR/HRV/resp all gate). Mood entries present + a
     // scorable sleep night → the two present components are both live reads,

src/lib/insights/derived/readiness.ts

@@ -194,8 +194,19 @@ async function readLatestDayMean(
 
 /**
  * Score one baseline-deviation vital component (RHR / HRV / respiratory).
- * Returns null when the baseline is insufficient or there is no recent
- * reading — the component then drops from the blend.
+ *
+ * null-vs-0 contract (iOS v0.14.3 F2)
+ * -----------------------------------
+ * A contributor is `null` ONLY when there is no usable signal: the baseline
+ * is not established (insufficient history / band failed), there is no recent
+ * reading, or the inputs are non-finite. A `null` contributor drops from the
+ * blend and the client hides it. A FINITE number — including a genuine `0` —
+ * means the baseline IS established and the latest reading really does deviate
+ * that far (≥ 2 spreads in the unfavourable direction maps to 0; see
+ * `scoreDeviation`). We deliberately never emit `0` as a "missing data"
+ * sentinel: the previous behaviour of letting a non-finite `today`/`center`
+ * fall through `clamp100` could surface a misleading 0 for an account that had
+ * the metric but no scorable baseline — that case is now `null`.
  */
 async function scoreVitalDeviation(
   userId: string,
@@ -225,7 +236,15 @@ async function scoreVitalDeviation(
     };
   }
   const today = await readLatestDayMean(userId, type, windowDays, now);
-  if (today == null) {
+  // No recent reading, or a non-finite baseline center / today value: there
+  // is no scorable deviation. Return null (hidden) rather than letting a NaN
+  // fall through `scoreDeviation` → `clamp100`, which would surface a
+  // misleading 0 for an account that actually has the metric (iOS F2).
+  if (
+    today == null ||
+    !Number.isFinite(today) ||
+    !Number.isFinite(baseline.value.center)
+  ) {
     return {
       value: null,
       source: baseline.provenance.source,

src/lib/insights/metric-status.ts

@@ -59,6 +59,12 @@ import {
   summarizeSeries,
 } from "@/lib/insights/status-shared";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
+import {
+  reconstructSleepNights,
+  type SleepStageRow,
+} from "@/lib/analytics/sleep-night";
+import { loadUserSourcePriority } from "@/lib/rollups/measurement-read";
+import { resolveUserTimezone } from "@/lib/tz/resolver";
 import {
   readFreshStatusText,
   resolveReadOnlyStatusMiss,
@@ -206,19 +212,64 @@ export async function generateMetricStatus(args: {
   // Bounded raw read for the derived recent stats; the monthly/yearly
   // tail comes from the rollup tier (full-history fallback on a miss),
   // exactly as the specialised cards do.
-  const measurements = await prisma.measurement
-    .findMany({
-      where: { userId: args.userId, type: meta.measurementType, deletedAt: null },
-      orderBy: { measuredAt: "desc" },
-      take: 365,
-      select: { value: true, measuredAt: true },
-    })
-    .then((rows) => rows.reverse());
+  //
+  // SLEEP_DURATION is the one type stored ONE ROW PER STAGE per night
+  // (IN_BED / AWAKE / CORE / DEEP / REM, minutes each), so a raw per-row
+  // read would feed the status prose a single ~16 min stage as the
+  // "current sleep" instead of the night total. Route it through
+  // `reconstructSleepNights` (the same helper the dashboard / series /
+  // targets use) so every derived value here — `series`, `summary`,
+  // `latest` — is a per-night TIME-ASLEEP total in minutes (the registry
+  // unit + normalRange are already minutes). The night reconstruction
+  // clusters stages into sessions, dedups multi-source nights via the
+  // user's `sleep` priority ladder, and handles the midnight split. Gated
+  // to SLEEP_DURATION only so no other metric's status path changes.
+  const isSleep = meta.measurementType === "SLEEP_DURATION";
 
-  const points = measurements.map((m) => ({
-    measuredAt: m.measuredAt,
-    value: m.value,
-  }));
+  const points: Array<{ measuredAt: Date; value: number }> = [];
+  let measurements: Array<{ measuredAt: Date }> = [];
+  if (isSleep) {
+    const [tz, priorityJson, sleepRows] = await Promise.all([
+      resolveUserTimezone(args.userId),
+      loadUserSourcePriority(args.userId),
+      prisma.measurement.findMany({
+        where: {
+          userId: args.userId,
+          type: "SLEEP_DURATION",
+          deletedAt: null,
+        },
+        orderBy: { measuredAt: "desc" },
+        take: 365 * 6, // ≈ 5 stage rows/night × source-count for a year
+        select: { value: true, measuredAt: true, sleepStage: true, source: true },
+      }),
+    ]);
+    const nights = reconstructSleepNights(
+      sleepRows as SleepStageRow[],
+      tz,
+      priorityJson,
+    ).filter((n) => n.asleepMinutes > 0);
+    for (const n of nights) {
+      points.push({ measuredAt: n.measuredAt, value: n.asleepMinutes });
+    }
+    measurements = nights.map((n) => ({ measuredAt: n.measuredAt }));
+  } else {
+    const rows = await prisma.measurement
+      .findMany({
+        where: {
+          userId: args.userId,
+          type: meta.measurementType,
+          deletedAt: null,
+        },
+        orderBy: { measuredAt: "desc" },
+        take: 365,
+        select: { value: true, measuredAt: true },
+      })
+      .then((r) => r.reverse());
+    for (const m of rows) {
+      points.push({ measuredAt: m.measuredAt, value: m.value });
+    }
+    measurements = rows.map((m) => ({ measuredAt: m.measuredAt }));
+  }
   const series = applyPayloadBudget(points, { now });
   const graded = await buildGradedSeriesWithRollups(
     args.userId,