LG-4 — CORE-Logos Studio: Alignment tab (centerpiece) + holonomy-absent

Date: 2026-06-14 Wave: read-only CORE-Logos Studio (L1–L5). Parent pack: docs/handoff/core-logos-studio-readonly-briefs-2026-06-14.md. Plan: docs/workbench/core-logos-studio-plan.md. Depends on: LG-2 merged (/logos route shell + the 10 Logos* TS interfaces mirrored). Rebase on it. May be authored in parallel with LG-3 but merges sequentially with it (shared train files: routes.ts / App.tsx / api.ts / evidenceAddress.ts) — rebase-union + retest. Scope (hard): the Alignment tab (the wave centerpiece) + confirming the honest holonomy-absent state. No contents tabs (LG-3). No mutation. Reads only GET /logos/packs/{id}/alignment (list[LogosAlignmentRow]).

Non-negotiables

Determinism is the whole point. Render with the existing deterministic DAG primitive (src/design/components/Dag/Dag.tsx, layered longest-path via layout.ts). No force-directed / nondeterministic layout, no motion-as- cognition. A golden-file layout test is required (mirror Dag/Dag.layout.golden.json).
No theater, no recompute. Render only the endpoint's fields.

Build

Alignment tab

Add an Alignment tab to the existing /logos TabBar. The trilingual resonance graph (he → grc → en) is the strongest real data in the wave — make it the centerpiece. Consumer pattern: src/app/proposals/ProposalChainViewer.tsx (the existing Dag consumer).
Per-edge card / row (LogosAlignmentRow): source_id · target_id · relation · weight · evidence_ids · target_pack_id · invalid-target warning when invalid_target is true.
Surface the invalid targets honestly. LG-1 reports 5 undeclared en-collapse-* anchors (breath/heart/holy/soul/time) as invalid_target in grc_logos_cognition_v1; the declared 3 (love/justice/peace) resolve. Show the warning — do not hide or smooth it. This is a reviewer-visible "the geometry tells the truth" moment, not a defect to mask.
Answer the plan's four questions in the tab: what does this Hebrew root align with / what Greek relation carries the same pressure / what English surface receives it / what evidence supports the edge (evidence_ids, e.g. John1:1 / Gen1:1).

Evidence subject (`src/app/evidenceAddress.ts`)

Add logos_alignment_edge → logos:<pack_id>:alignment:<edge_id>, where edge_id is the LG-1 deterministic id (sha256(source|target|relation)[:16], already on LogosAlignmentRow).
Union member, equality, encode/parse, right-inspector projection, copyable pointer, ?inspect= sync, chain-rail derivation. Round-trip test.

Holonomy — confirm absent-state (no new tab)

No Holonomy tab, no proof card, no success state. Overview already shows holonomy_case_count 0 + missing_evidence (LG-2); Safety shows missing_holonomy_refs unknown. LG-4's only holonomy work is a guard that no holonomy proof affordance leaks in — add a test asserting the /logos surface renders no holonomy success/proof element. (Real holonomy cards are the future W-Holonomy wave once holonomy.jsonl exists.)

Acceptance / tests

Golden-file layout test for the alignment DAG (deterministic node/edge layout; mirror the Dag.layout.golden.json pattern).
Invalid-target render test: an invalid_target edge shows the warning; a resolved edge does not. Assert the 5 undeclared en-collapse-* anchors warn.
evidenceAddress.test.ts — logos_alignment_edge round-trip.
Holonomy-absent guard: no proof/success element on the /logos surface.
routes.test.tsx / schemaDrift / enumCoverage stay green. No new engine schemas — NOT_YET_MIRRORED untouched.
pnpm build clean.
Run the FULL pnpm exec vitest run before push — it must pass AND exit. The CI workbench-ui job gates on the whole suite; focused per-file runs miss a PR's own new test files and cross-cutting tests (LG-2 lesson: stale route-count assertions only surfaced in the full run, never the focused one).

Out of scope (separate brief)

LG-3 — Lexicon / Glosses / Morphology contents tabs.
W-Holonomy / W-Forge — future waves (author holonomy.jsonl; universal ProposalArtifact envelope + Patch Forge). Not this wave.

4.3 KiB Raw Blame History Unescape Escape