Commit Graph
6 Commits
Author SHA1 Message Date
amosblomqvist 1d45d77303 feat: graph context injection — the session's prior knowledge, steered in at session start (vision-context-injection.md)
One shot per chat, fired at the first turn's onFinish: the session's
topic (its only typed message) is embedded as an EmbeddingGemma retrieval
query; the semantically closest graph propositions are selected under the
session's token budget (pure budget, no similarity threshold — the
decided doctrine) and delivered as a marker-only steer through the wake
pipeline, queuing behind pending questions like every sub-agent result.

- contracts: ChatSettings.graphContextTokens (default 1024, max 4096,
  step 128, 0 = off); GraphContextData + graphContextText; the
  graph-context JobKind + buildGraphContextSteer (id steer_ctx_<chatId>)
- core/graph-context.ts: brute-force cosine top-K, greedy chars/4 fill
  under budget, most-recent UnderstandingLevel fold, count cap 64,
  parse-don't-cast at the marker seam
- server: embedQuery (role-asymmetric query format, same warm singleton
  — vector-space parity by construction); the runner (startGraphContext-
  Injection one-shot gates; runGraphContextJob validates ledger parity);
  wake branches (chatIdOf/jobIdOf, finalizeJob done, recoverStaleJobs
  re-deliver or contextRespawn); abortJob kills context jobs SILENTLY
  (nothing was promised in the lesson); turn-service trigger seam
- adapter: data-graph-context flattens to the canonical block (malformed
  parts have no model face)
- UI: the budget slider (home dock + /chat/new draft, the session-data
  strip); the quiet SystemLine "~537 · 10 propositions"; dock kind
  "Knowledge context"; model-picker PUT now spreads settings (a model
  pick can no longer clear the sibling override)

Live-smoked end to end on the real embedder + real GLM: seeded ledger
(10 propositions), session on "modal logic: possible worlds and the
necessity box" — job in the dock, queued behind the prober's question,
delivered on the answer, SystemLine in the transcript, wake turn reads
the block and continues the probe. Gates: tsc, lint, 459/459, build.
2026-09-15 07:54:50 +03:00
amosblomqvist bbee72f354 Background agents → sidebar section (owner): the retired floating dock's successor
The old activity dock (fixed bottom-right, every view) was a cross-session
widget parked inside the session view — the same boundary confusion that
got the plan window moved. The sidebar is the app's cross-session surface,
so the agents section lives at its bottom strip now (user-locked design):

- ALWAYS present: empty means empty — the header ('Background agents',
  matching the Sessions label language) sits alone; nothing hides and
  nothing decorates (the total-cost span only renders when populated).
- Populates and grows when processes run: same blobs (sub-agents +
  extraction runs, one poll, inferred cost, elapsed heartbeat), capped by
  max-h-[40dvh] + internal scroll so a pile never crowds the session list.
- Per-blob END button (unchanged abort contract); the detail popover
  becomes an INLINE expansion — the section scrolls, a floating popover
  would clip at its bounds.
- Sidebar-native skin: border-t separator matching the sidebar's edge,
  session-row hover language, no floating-card chrome.
- Sidebar closed = the learner opted out; no ambient signal for now
  (user decision — invisibility until a clearer observability story).
- Narrow-mode count badge/drawer machinery retired with the dock.

app/layout.tsx: ActivityDock mount removed, ActivityProvider stays (the
sidebar consumes the context). Live-verified: populated rows with inline
detail + session link, abort makes the row vanish, idle renders the
header alone, closed sidebar hides the section with nothing floating.
Gates: typecheck, lint, 423/423 tests.
2026-09-14 11:56:22 +03:00
amosblomqvist 3360181e41 Wave-5 reskin batch 5: normalize stray pixel text sizes onto the declared ladder
--text-xxs (10px) joins the ladder; every component-level text-[9|10|11px]
literal in owned components now pulls from the ladder (graph-view.tsx and
home.tsx untouched — graph stream). Cosmetic pins updated where touched.
2026-09-12 23:48:27 +03:00
amosblomqvist 3e74f28e4f Dock END button speaks the real abort contract: already-gone is success, only true failures report
POST /api/activity/abort landed on main (fde46c2). 404 not_found /
409 not_running mean the process reached the desired end state on its
own — refresh and clear the abort state, no error surfaced. Genuine
failures (network/400) show 'Abort failed — try again' on the blob.
2026-09-11 03:34:03 +03:00
amosblomqvist df58bb87f2 The stats bar: context + cost beneath the composer, per-turn detail
One quiet row beneath the composer/question shell (not flush to the
screen bottom; static height, outside the morphing shell):

- LEFT: context window + session cost, clickable → per-turn detail
  popover (context, total, unpriced count, one row per turn with
  model/steps/cost). All folds from core/prices.ts — sessionCost,
  contextWindow, turnCost — no cost math in components.
- RIGHT: model + "Thinking —" — a placeholder slot for the coming
  model/thinking-level CONTROL (later work with its contract changes);
  display-only today, model read from the last turn's usage,
  pretty-printed (Glm 5.3 Flash).
- Data: the record's turns log, seeded on mount, refreshed when the
  local stream settles, kept fresh by the existing record re-syncs.
  No turns yet → "—" everywhere; unpriceable turns never fold into a
  fake $0 (the doctrine).
- formatUsd/formatTokens moved to components/usage-format.ts (shared
  with the activity dock); prettyModel joins them, pure + tested.
2026-09-11 03:15:55 +03:00
amosblomqvist e22f0bddb4 The activity dock: every background process, every view, bottom-right
One reserved corner of the shell, one blob per running BACKGROUND
process — session sub-agents (research/demo/visual) and global
extraction runs. Teaching turns are deliberately NOT here: the dock
exists to surface what the learner can't otherwise see; a running
teaching turn is already fully visible where it happens (user decision).

- Data: ONE poll of GET /api/activity (jobs across ALL chats,
  title-joined, oldest-first + extraction runs), folded client-side
  with core/prices.ts (jobCost). Adaptive interval — 2.5s while
  anything runs, 10s idle, 4× slower on hidden tabs; a failed fetch
  keeps the last feed (no flicker to empty on a blip).
- Per blob: kind icon, elapsed (ticking), cost, activity line, and an
  explicit END on the blob's face. Click → detail popover (context
  line, model, steps, open-session link). The context line is the LAST
  model call's tokens vs the model's limit (JobUsage.lastStep — pending
  agents-stream contract; "—" until it populates), NEVER a cumulative
  sum. Absent data renders "—", never zero (the doctrine).
- Header: total running cost (sum of priced blobs, "—" when nothing
  is priced). Narrow viewports collapse to a count badge → drawer.
- END posts to /api/activity/abort {kind, id} — the backend is being
  composed by master over the streams' mechanisms; until it exists a
  click locks the button and the popover says "Abort not available
  yet" (never a dead control).
- Idle renders NOTHING. Mounted once in the root layout; every view
  gets it for free.

toBlobs / formatElapsed / formatUsd / formatTokens are pure + tested;
SSR render tests pin blobs, the "—" doctrine, the face END, and the
priced total.
2026-09-11 03:12:32 +03:00