chain table redesign (owner refinements, 2026-09-29 evening)

- full-bleed: blur + translucent surface IS the overlay, no inner card;
  own top bar (info + search + close, lh-no-drag so home's drag mask
  never eats the X — the clickability bug)
- three levels, one click each: cluster (birth session; a chain lives in
  its CONCLUSION's cluster) → chain → proposition STATEMENTS revealed
  directly (names never twice; singletons one click); all collapsed by
  default
- proposition names as outlined chips, → connectors between; lonely chip
  = standalone proposition — the '(no premises)' annotation is dead in
  the visible UI (survives only as aria-label); Folder/Waypoints icons;
  rotating chevrons; deep statement indent
- real table look: hairline row + column lines; sortable Name + Created
  headers (click toggles asc/desc; size is text, not a control)
- selection never shifts layout: floating bottom bar (Google-Photos
  pattern) with N propositions → M chains + Clear + Generate
- fuzzy search grouped + auto-expanding clusters

live-verified end-to-end on the owner's dev server; 15 rewritten render
pins; gates: tsc 0, 1179/1179, lint 0, build 0

docs/architecture.md

@@ -382,7 +382,7 @@ Practice turns the graph into problem sets the learner reasons through out loud.
 
 **The jobs pattern — registry, not JobRecords.** Practice runs are NOT Workbench jobs: no `JobStore` records, no steer messages, nothing ever wakes the teacher — a practice run isn't chat-scoped (a set outlives its birth session and has no lesson to re-enter; its result is itself, persisted in its own store and read by its own page). What it shares with the Workbench is the observability surface: a `globalThis` run registry (the module-instance idiom — `__lhPracticeRuns`) with PER-PROBLEM ROWS (owner ruling, 2026-09-29 — matching the extraction per-chunk pattern): `setId:partition` while the partitioner runs ("Partitioning — title"), then one `setId:gen:problemId` per problem ("queued"→"generating", label "Problem i/N — title"), and one `setId:grade:problemId` per graded answer ("queued"→"grading", "Grading i/N — title") — each row its OWN abort controller and its OWN usage accumulation; the row key IS the feed id (the dock dup-key law) and the abort target. ABORT ROUTES BY ROW SHAPE: a gen-row abort kills THAT problem only (siblings continue; the set still settles `ready` with one errored problem), a partition-row abort ends the whole run, a grade-row abort drops that grade (answer kept — one attempt, the problem stays done); set-level abort (delete) composes from row aborts plus the synchronous persisted marks. Model calls — partitioner included — run under the SHARED background-LLM semaphore (`bg-llm-semaphore.ts`): one process-wide pool of 3 permits for ALL background LLM jobs, extraction chunks included (total background LLM concurrency bounded at 3, whoever calls). A boot sweep marks orphaned `partitioning`/`generating` sets as `error "Interrupted (server restarted)"` — once per process, and only when the set's `updatedAt` heartbeat is >120s stale (a live run's 30s FILE watchdog keeps the heartbeat fresh even across dev's per-worker processes — the registry is not shared between them; the guard trio closed a live-stamping bug from the first live test) — the job runner's crash-honesty at set grain. Store writes serialize behind a per-set mutex; the pipeline is the sole writer of `data/practice`.
 
-**API + UI.** Creation is GRAPH-DOMAIN, not chat-domain (owner ruling 2026-09-29: "problem sets always take in nodes — graph objects"): `POST /api/practice/sets` takes `{nodeIds, params, title?, note?}` ONLY — no chatId, no clusterId; the UI resolves a picked cluster to its member node ids and passes the cluster's label as the title. The `note` is the learner's optional free-form creation input (owner, same evening: more learner control, no other purpose): trimmed, capped at 2000 characters, persisted on the set, and carried to the partitioner and every generator as `THE LEARNER'S NOTE` — an additional input the agents WEIGH but are never obliged to follow (the doctrine pins the stance: it never moves a problem off its rung or out of its band, and the partition laws always win). `GET /api/graph/clusters` serves the picker: the birth-session partition itself (`birthSessionPartition`, same ordering — size desc, ties by id) — EVERY group regardless of size (owner ruling, same evening: MIN_CLUSTER_SIZE is a hull-drawing rule; practice availability must not inherit a visual threshold), each as `{key, kind: "chat"|"notes", label, nodeIds, nodeCount}`; a cluster whose naming record is gone keeps `label: null` — honest, never guessed. 202 `{id}` with generation fire-and-forget; 400 empty selection (empty intersection with live nodes — the vanished-nodes edge) / bad params; 503 model/key refusal. `GET /api/practice/sets` · `GET/DELETE /api/practice/sets/[id]` (delete aborts live runs first) · `POST /api/practice/sets/[id]/answers` (202; 409 when not ready; 400 re-answering a submitted problem — one attempt). Surfaces: `/practice` (the list — the chat-pattern top bar with the sidebar reopen toggle and the New action at the right edge, status/counts rows, delete confirm, the new-set dialog with the named-level controls (a 5-detent ladder slider, a 3-detent cue slider, a Theory|Application toggle — each showing the level's name + definition), the cluster picker, and the optional free-form note textarea (typing-only). The SECOND creation entry (owner plan, same evening): the graph view's **chain table** — a `Chains` button in the graph chrome (empty graph: hidden) opens a full-screen overlay (the session-overlay idiom: absolute inside the graph area, Esc/backdrop dismiss) listing EVERY reasoning chain via the core `formatChainLine` one-liner (`"Premise one, Premise two → Conclusion"`, singletons marked `"(no premises)"`), each row expanding to its member propositions and each member to its full statement; columns are birth session(s) (the clusters' honest label join — chat title / notes path / muted `unlabeled`) and created date; NO understanding-levels column (explicitly not built). Selection is checkboxes on whole chains or single propositions — a flat node-id set (a node in several chains is one selection) — with a fuzzy search (names, aliases, statements) that filters the LIST only. The pinned Selected section derives the agents' ACTUAL input live (`chainsForSelection` over the selection: a premise picked without its conclusion shows as the singleton it becomes) as `N propositions → M chains` + the derived lines, with Clear and **Generate problem set**, which opens the SAME creation dialog in **custom mode**: no cluster picker — a selection summary (the derived shape + the dated default title the pipeline will stamp) — posting `{nodeIds, params, note}` with no title. The overlay is read-only toward the graph and `/practice/[id]` (the exam page — the chat-pattern top bar, a levels + status line, numbered problems, markdown statements with math healing at the display seam, a textarea + focus-gated dictation bar per problem with its OWN Submit button — one attempt per problem, instant solution/rubric reveal, the score landing live). Entry: the sidebar's Practice nav row.
+**API + UI.** Creation is GRAPH-DOMAIN, not chat-domain (owner ruling 2026-09-29: "problem sets always take in nodes — graph objects"): `POST /api/practice/sets` takes `{nodeIds, params, title?, note?}` ONLY — no chatId, no clusterId; the UI resolves a picked cluster to its member node ids and passes the cluster's label as the title. The `note` is the learner's optional free-form creation input (owner, same evening: more learner control, no other purpose): trimmed, capped at 2000 characters, persisted on the set, and carried to the partitioner and every generator as `THE LEARNER'S NOTE` — an additional input the agents WEIGH but are never obliged to follow (the doctrine pins the stance: it never moves a problem off its rung or out of its band, and the partition laws always win). `GET /api/graph/clusters` serves the picker: the birth-session partition itself (`birthSessionPartition`, same ordering — size desc, ties by id) — EVERY group regardless of size (owner ruling, same evening: MIN_CLUSTER_SIZE is a hull-drawing rule; practice availability must not inherit a visual threshold), each as `{key, kind: "chat"|"notes", label, nodeIds, nodeCount}`; a cluster whose naming record is gone keeps `label: null` — honest, never guessed. 202 `{id}` with generation fire-and-forget; 400 empty selection (empty intersection with live nodes — the vanished-nodes edge) / bad params; 503 model/key refusal. `GET /api/practice/sets` · `GET/DELETE /api/practice/sets/[id]` (delete aborts live runs first) · `POST /api/practice/sets/[id]/answers` (202; 409 when not ready; 400 re-answering a submitted problem — one attempt). Surfaces: `/practice` (the list — the chat-pattern top bar with the sidebar reopen toggle and the New action at the right edge, status/counts rows, delete confirm, the new-set dialog with the named-level controls (a 5-detent ladder slider, a 3-detent cue slider, a Theory|Application toggle — each showing the level's name + definition), the cluster picker, and the optional free-form note textarea (typing-only). The SECOND creation entry (owner redesign, same evening): the graph view's **chain table** — a `Chains` button in the graph chrome (empty graph: hidden) opens a FULL-BLEED table overlay (absolute inside the graph area; the blur + translucent background IS the surface, no inner card; its own top bar carries the info + the close button, `lh-no-drag` so home's window-drag mask never eats it; Esc or X dismiss). The hierarchy is three levels, one click each: CLUSTER (the birth session — a chat or a notes file; a chain lives where its CONCLUSION was born) → REASONING CHAIN → PROPOSITION STATEMENTS — expanding a chain reveals its members' statement text directly (names live in the chain line, never twice; a singleton reveals its one statement in a single click); everything starts collapsed. The table look (owner ruling: "it should really resemble a table"): drawn hairline row + column lines, sortable column headers on Name + Created (click toggles asc/desc; the size column is text, not a control; default Name A→Z). In the chain line every proposition name is a faint-outlined CHIP with the → a clearly separated connector between chips — a lonely chip reads as a standalone proposition (the "(no premises)" annotation is dead in the visible UI; it survives only as the row's aria-label); icons: clusters FolderIcon (open-folder when expanded), chains WaypointsIcon, propositions chip-only; chevrons rotate with expansion. Selection is checkboxes on whole clusters, chains, or single propositions — a flat node-id set (a node in several chains is one selection) — accumulated in a FLOATING BOTTOM BAR (the Google-Photos pattern: appears only when something is selected, never shifts the table layout) showing the agents' ACTUAL input (`chainsForSelection` over the selection: a premise picked without its conclusion counts as the singleton it becomes) as `N propositions → M chains`, with Clear and **Generate problem set** (the one action for now — the bar is a neutral action slot), which opens the SAME creation dialog in **custom mode**: no cluster picker — a selection summary (the derived shape + the dated default title the pipeline will stamp) — posting `{nodeIds, params, note}` with no title. A fuzzy search (names, aliases, statements) filters chains grouped + auto-expanded under their clusters (non-matching clusters hide; the collapsed view returns on clear). NO understanding-levels column (explicitly not built). The overlay is read-only toward the graph and `/practice/[id]` (the exam page — the chat-pattern top bar, a levels + status line, numbered problems, markdown statements with math healing at the display seam, a textarea + focus-gated dictation bar per problem with its OWN Submit button — one attempt per problem, instant solution/rubric reveal, the score landing live). Entry: the sidebar's Practice nav row.
 
 **Model wiring.** Three roles — `partitioner`, `problemGenerator`, `grader` — all pinned to the **strong** slot (quality-critical, one settings edit from cheap) and in `BACKGROUND_MODEL_ROLES` (per-role overrides). The stage agents are single-submit loops on the shared `runSubmitAgent` seam (`submit-agent.ts`, extracted from extraction for exactly this): a `validate` hook turns a bad submit into a tool-input error the model retries against, within `maxTurns 4`; exhaustion fails the stage honestly (the set says so, the chains re-run cleanly on a new set).