|
|
|
@@ -125,7 +125,7 @@ per-release `**vX.Y.Z:**` narration — CI enforces this
|
|
|
|
|
- `src/core/search/autocut.ts` (default ON in reranked modes) — Weaviate-style autocut: score-discontinuity result-sizing on the cross-encoder rerank separatrix. `applyAutocut(results, scoreOf, cfg)` normalizes the reranker scores, finds the largest consecutive gap, and cuts there when it clears `jumpRatio` (default 0.20); robust to unsorted provider output (cuts on a sorted copy, keeps items in INPUT order via a score threshold), guards `top<=0`/non-finite, never returns empty, and no-ops when <2 results carry a finite `rerank_score` (covers the reranker fail-open path). WHY rerank_score and NOT RRF/cosine: gbrain measured (see `return-policy.ts`) that the RRF rank1→rank2 gap is ~flat whether rank-1 is right or wrong — not a separatrix; the cross-encoder score IS. So autocut runs ONLY where the reranker ran (the floor reaches `balanced`+`tokenmax`; `conservative` is a documented no-op). Exports `AutocutConfig`, `DEFAULT_AUTOCUT` (frozen: `enabled=true, jumpRatio=0.20, minKeep=1`), `AutocutDecision` (`{applied, signal:'rerank'|'none', cut, kept, total, gapRatio}`), `AutocutInput`, `autocutFromConfig`, `resolveAutocut`, `applyAutocut`. Cache-key integration (clean path, not the adaptive-return cache-skip hack): enable+sensitivity flow through `ModeBundle` → `ResolvedSearchKnobs` → `knobsHash` exactly like `graph_signals`. `mode.ts` adds `autocut`/`autocut_jump` (conservative false, balanced/tokenmax true@0.20) AND sets `reranker_top_n_in = searchLimit` for reranked modes (so the reranker scores the full returned set; there is no un-scored tail for autocut to wrongly drop — closes the load-bearing recall finding). Autocut folds into `knobsHash` as its own parts entry (`mode.ts:KNOBS_HASH_VERSION` is the single source of truth for the current hash version; every bump is a one-time global cache cold-miss on upgrade). Wired into `hybridSearch` AFTER adaptive-return, BEFORE the limit slice, first page only; emits `HybridSearchMeta.autocut`. BOTH the cache-miss `finalMeta` and cache-HIT `cachedMeta` rebuilds carry `autocut`+`adaptive_return`+`mode`+`embedding_column`. Preserves alias-hop exact matches: `applyAutocut` takes an optional `preserve` predicate; hybrid passes `r => r.alias_hit === true` so a canonical page injected by `applyAliasHop` after reranking (no `rerank_score`) is never cut. Agent surface: `query` op `autocut` boolean (ceiling override — `false` forces full top-K); `SearchOpts.autocut`; `--explain` shows per-result `rerank_score`, `formatAutocutSummary` renders the decision when search meta is threaded; `gbrain search modes` attribution; metric glossary `autocut.signal`/`autocut.gap_ratio`. Config: `search.autocut`, `search.autocut_jump`. Default-ON backed by an in-repo eval gate — `test/search/autocut-eval.test.ts` (also `bun run eval:autocut`) measures precision-lift-without-recall-regression over labeled qrels fixtures with modeled cross-encoder distributions (no API key, no sibling repo; runs in CI): mean precision 0.33→0.94, recall 1.00→0.95, ZERO recall regression on enumeration queries. Env-overridable floors. Pinned by `test/search/autocut.test.ts` (pure-fn), `test/search/query-op-autocut.test.ts` (agent surface), `test/search/autocut-integration.serial.test.ts` (IRON-RULE behavioral via `rerankerFn` DI seam: cliff trims, flat doesn't, no-reranker no-ops, `autocut:false` ceiling, composes with adaptive-return), `test/search/autocut-eval.test.ts` (the precision/recall gate), and the knobsHash assertions in `test/search-mode.test.ts`.
|
|
|
|
|
- `src/core/ai/recipes/voyage.ts` — Voyage AI openai-compatible recipe. Declares `chars_per_token=1` + `safety_factor=0.5` so the gateway pre-splits Voyage batches at a 60K-character budget (50% of 120K-token cap with the dense-tokenizer ratio), avoiding the backfill loop where tiktoken-grounded budgeting undercounted Voyage's actual token usage. Declares `multimodal_models: ['voyage-multimodal-3']` so the gateway rejects text-only Voyage models pointed at the multimodal endpoint with a clear `AIConfigError` instead of waiting for Voyage's HTTP 400. Recipe docstring at `:7-16` names the seven hosted flexible-dim models that accept `output_dimension` (`voyage-4-large`, `voyage-4`, `voyage-4-lite`, `voyage-3-large`, `voyage-3.5`, `voyage-3.5-lite`, `voyage-code-3`) and notes `voyage-4-nano` is the open-weight variant fixed at 1024-dim that does NOT accept the parameter (negative regression assertion in `test/ai/gateway.test.ts`: `dimsProviderOptions` returns `undefined` for `voyage-4-nano`). `voyage-code-3` is the recommended embedding model for gstack per-worktree code brains (Topology 3 in `docs/architecture/topologies.md`); discoverability surfaces: decision-tree branch in `docs/integrations/embedding-providers.md`, Topology 3 "Recommended embedding model" subsection, runtime nudge from `gbrain reindex --code` against non-code-tuned models. Recipe-shape regression pinned by `test/ai/voyage-code-3-recipe.test.ts`.
|
|
|
|
|
- `src/core/ai/recipes/anthropic.ts` — Anthropic recipe (chat + expansion touchpoints). Canonical id is `claude-sonnet-4-6` (no date suffix); a reverse alias `claude-sonnet-4-6-20250929 → claude-sonnet-4-6` keeps stale user configs working (rescues `facts.extraction_model` and `models.dream.synthesize`). Recipe-shape regression pinned by `test/anthropic-model-ids.test.ts`.
|
|
|
|
|
- `src/core/model-pricing.ts` — single source of truth for paid-cloud CHAT/completion model pricing (USD per 1M tokens, input | output). `CANONICAL_PRICING` is a `provider:model`-keyed table (Anthropic Opus 4.8/4.7/4.6 `$5/$25`, Sonnet 4.6 `$3/$15`, Haiku 4.5 `$1/$5` both dateless + dated, plus OpenAI / Google / Together / DeepSeek panel models). `canonicalLookup(modelId)` resolves bare (`claude-opus-4-8`), colon (`anthropic:claude-opus-4-8`), and slash (`anthropic/...`) forms — bare ids default to the `anthropic:` provider; nested OpenRouter ids (`openrouter:anthropic/...`) intentionally MISS so OpenRouter markup isn't repriced as the inner vendor. Every other chat-pricing table is a DERIVED view of this one (NOT a hand-copied duplicate), so cross-table price drift is structurally impossible. Embeddings live separately in `embedding-pricing.ts` (different unit). Pinned by `test/model-pricing.test.ts` whose drift guard asserts each derived view equals canonical and that the cross-modal panel models are all present.
|
|
|
|
|
- `src/core/model-pricing.ts` — single source of truth for paid-cloud CHAT/completion model pricing (USD per 1M tokens, input | output). `CANONICAL_PRICING` is a `provider:model`-keyed table (Anthropic Opus 5/4.8/4.7/4.6 `$5/$25`, Sonnet 4.6 `$3/$15`, Haiku 4.5 `$1/$5` both dateless + dated, plus OpenAI / Google / Together / DeepSeek panel models). `canonicalLookup(modelId)` resolves bare (`claude-opus-4-8`), colon (`anthropic:claude-opus-4-8`), and slash (`anthropic/...`) forms — bare ids default to the `anthropic:` provider; nested OpenRouter ids (`openrouter:anthropic/...`) intentionally MISS so OpenRouter markup isn't repriced as the inner vendor. Every other chat-pricing table is a DERIVED view of this one (NOT a hand-copied duplicate), so cross-table price drift is structurally impossible. Embeddings live separately in `embedding-pricing.ts` (different unit). Pinned by `test/model-pricing.test.ts` whose drift guard asserts each derived view equals canonical and that the cross-modal panel models are all present.
|
|
|
|
|
- `src/core/anthropic-pricing.ts` — bare-keyed Anthropic VIEW of `model-pricing.ts` (the `anthropic:` canonical entries with the prefix stripped). Kept distinct because many callers look up by bare Claude id and because `estimateMaxCostUsd(modelId, inTokens, maxOutTokens)` carries the null-on-miss contract the dream-cycle budget gate depends on (non-Anthropic ids return null → caller warns `BUDGET_METER_NO_PRICING` once and runs unbounded). `estimateMaxCostUsd` routes bare/colon/slash ids through `splitProviderModelId`. Do NOT hand-edit prices here — the map is derived from canonical, so it cannot drift. `ANTHROPIC_PRICING` is consumed by `budget/budget-tracker.ts`, `minions/batch-projection.ts`, and `cycle/budget-meter.ts`.
|
|
|
|
|
- `src/core/takes-quality-eval/pricing.ts` — fail-closed budget pricing for `eval takes-quality run --budget-usd N`. `MODEL_PRICING` is a curated `provider:model` allowlist (default panel + likely overrides) whose VALUES are derived from `model-pricing.ts` via `canonicalLookup`; an allowlisted id missing from canonical throws at module load. Schema is `{input_per_1m, output_per_1m}`. A model NOT on the allowlist aborts the run with an actionable error rather than guessing (distinct from `cross-modal-eval/runner.ts`, which silently estimates zero on unknown models — both now source numbers from canonical).
|
|
|
|
|
- `src/core/budget/budget-tracker.ts` — keystone primitive for the brainstorm cost-cathedral wave. One typed error (`BudgetExhausted` with `reason: 'cost' | 'runtime' | 'no_pricing'`), one schema-stable audit JSONL at `~/.gbrain/audit/budget-YYYY-Www.jsonl`. Contracts: `record()` throws when cumulative spend exceeds cap (the cap is a real ceiling, not a suggestion); `reserve()` hard-fails with `reason: 'no_pricing'` when `maxCostUsd` is set AND the model is missing from pricing maps (warn-once preserved when cap is unset); `extractUsageFromError(err, fallback)` returns `err.usage` when the SDK provides it, else the pessimistic fallback (caller passes `maxOutputTokens`, not the optimistic pre-call estimate). `onExhausted(cb)` fires once synchronously BEFORE the throw propagates so callers can persist checkpoints. Replaces three parallel copies (inline brainstorm class, cycle/budget-meter, eval-contradictions). Adapts the old `BudgetMeter` (public shape preserved + `schema_version: 1` stamped on every dream-budget audit line). Pinned by 18 unit cases.
|
|
|
|
|
Masa