Commit Graph
3 Commits
Author SHA1 Message Date
Marc-André Bombeck 6182d45354 fix(i18n): print server-built sentences in the reader's notation
The decimal lint covered components and app only, and sentences for
people are built in src/lib too: the weekday weight correlation printed
"0.8 kg" into a German sentence, and the no-key fallback and the period
narrative's fallback spelled fractions with a full stop. Those three use
the new formatDecimal from format-locale with the reader's locale. The
rule now covers src/lib and src/hooks; text for the model, operator log
lines, SVG data, zone checks and wire values a client re-formats carry a
reasoned directive.
2026-10-20 16:52:08 +02:00
Marc-André BombeckandClaude Opus 4.8 b6126db754 feat(insights): extend the anti-repetition diversity context to the headline cards
The v1.12.1 diversity block (variety lead, data-strength line, steady-run
repetition signal, FDR-screened cross-metric relations) reached only the
generic archetype cards. The seven hand-tuned headline cards — the ones users
see most — ran at temperature 0.3 with only the previous-context block, so they
stayed prone to repetitive openers and forced filler steps. Thread the same
assessmentContextBlock through the specialized get*UserPrompt builders and
raise their temperature to 0.45 to match the archetypes.

Soften the unconditional "close with one step" across the specialized sections
and the activity / body-composition / sleep archetypes to the base prompt's
conditional form — close with a step only when the finding implies one,
otherwise affirm and name one thing to watch — so a steady metric no longer
manufactures a filler step. EN/DE parity preserved; guard computeSteadyRun
against an absent graded slice.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-05 13:58:28 +02:00
Marc-André Bombeck 05cff531e4 feat(insights): add relational context, repetition-awareness, and controlled variety to assessments
The per-metric assessment cards are tightly grounded but their form had
converged: every card opened the same way, forced a "one doable step"
even when nothing was actionable, and a metric that stayed steady for
weeks produced near-identical paragraphs. Squeeze more value from data
already computed server-side, without touching the grounding floor or any
safety contract.

- Surface FDR-surviving cross-metric correlations on the per-metric card.
  The discovery engine already screens behaviour×outcome pairs with a real
  Pearson + exact p-value + Benjamini-Hochberg FDR control; that
  intelligence only reached the period narrative. A new read-only consumer
  runs the same full-matrix discovery and filters to the pairs that
  involve the current metric, then feeds the engine's own conservative,
  descriptive, never-causal interpretation strings into the prompt as
  grounded relations. Best-effort: a correlation hiccup can never block the
  generation it only decorates.

- Streak-aware repetition. Derive a steady-run length from the graded
  series (consecutive recent weeks within a band of the user's own
  baseline) and tell the model when it has already reported "no material
  change" N times running, so it acknowledges the continuity in one clause
  and pivots to a different facet instead of restating the same level.

- Stop forcing a step; add controlled variety. The closing step is now
  conditional — when nothing is genuinely actionable the model affirms and
  names one thing to watch rather than manufacturing the filler that lets
  platitudes back in. A deterministic per-render variety token (seeded from
  user + metric + day, never Math.random / Date.now) rotates the opening
  angle so consecutive cards and days don't read identically, and the
  phrasing temperature lifts 0.3 → 0.45 while the facts stay pinned by the
  snapshot and the forbidden-phrase guards.

- Surface the computed data strength (n + recency) into the prose prompt so
  it hedges on thin data instead of guessing what "few" means, matching the
  UI confidence badge.

- Add locale-matched few-shot examples (a grounded assessment + a labelled
  banned-filler counter-example) to the assessment system prompt, which
  most helps weaker local-provider models follow the contract.

The own-baseline grounding, computed-not-hallucinated stats, schema-
enforced output, filler-phrase ban, and correlation-as-association
discipline are all preserved and re-pinned by new guard tests.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
(cherry picked from commit c9d04cab1523e0e57a3f50c87c9f88684b01b26f)
2026-06-05 03:25:22 +02:00