@tangle-network/agent-eval 0.119.0 → 0.120.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +25 -3
- package/README.md +0 -26
- package/dist/analyst/index.d.ts +30 -39
- package/dist/analyst/index.js +9 -5
- package/dist/analyst/index.js.map +1 -1
- package/dist/benchmarks/index.d.ts +11 -5
- package/dist/benchmarks/index.js +7 -7
- package/dist/campaign/index.d.ts +121 -66
- package/dist/campaign/index.js +26 -10
- package/dist/{chunk-RLCJQ6VZ.js → chunk-32BZXMSO.js} +385 -120
- package/dist/chunk-32BZXMSO.js.map +1 -0
- package/dist/{chunk-XJ7JVCHB.js → chunk-3A246TSA.js} +2 -2
- package/dist/{chunk-LMJ2TGWJ.js → chunk-JM2SKQMS.js} +2 -2
- package/dist/{chunk-F6YUH3L4.js → chunk-JN2FCO5W.js} +14 -168
- package/dist/chunk-JN2FCO5W.js.map +1 -0
- package/dist/{chunk-D7AEXSM5.js → chunk-PICTDURQ.js} +2 -2
- package/dist/{chunk-6QIM2EAP.js → chunk-PXD6ZFNY.js} +99 -1
- package/dist/chunk-PXD6ZFNY.js.map +1 -0
- package/dist/{chunk-GERDAIAL.js → chunk-QWMPPZ3X.js} +3 -3
- package/dist/{chunk-7A4LIMMY.js → chunk-S5TT5R3L.js} +229 -97
- package/dist/chunk-S5TT5R3L.js.map +1 -0
- package/dist/{chunk-ZYHJNKI3.js → chunk-ULOKLHIQ.js} +42 -5
- package/dist/chunk-ULOKLHIQ.js.map +1 -0
- package/dist/{chunk-FIUFRXP3.js → chunk-WW2A73HW.js} +49 -91
- package/dist/chunk-WW2A73HW.js.map +1 -0
- package/dist/{chunk-PAHNGS65.js → chunk-XDIRG3TO.js} +175 -11
- package/dist/chunk-XDIRG3TO.js.map +1 -0
- package/dist/contract/index.d.ts +76 -43
- package/dist/contract/index.js +596 -32
- package/dist/contract/index.js.map +1 -1
- package/dist/index.d.ts +92 -75
- package/dist/index.js +38 -35
- package/dist/index.js.map +1 -1
- package/dist/openapi.json +1 -1
- package/dist/rl.d.ts +11 -5
- package/dist/{run-campaign-OBXSN5WK.js → run-campaign-HNFPJET4.js} +2 -2
- package/dist/traces.d.ts +39 -25
- package/dist/traces.js +3 -5
- package/docs/trace-analysis.md +6 -2
- package/package.json +2 -7
- package/dist/chunk-6QIM2EAP.js.map +0 -1
- package/dist/chunk-7A4LIMMY.js.map +0 -1
- package/dist/chunk-F6YUH3L4.js.map +0 -1
- package/dist/chunk-FIUFRXP3.js.map +0 -1
- package/dist/chunk-PAHNGS65.js.map +0 -1
- package/dist/chunk-RLCJQ6VZ.js.map +0 -1
- package/dist/chunk-ZYHJNKI3.js.map +0 -1
- package/dist/primeintellect/index.d.ts +0 -311
- package/dist/primeintellect/index.js +0 -348
- package/dist/primeintellect/index.js.map +0 -1
- /package/dist/{chunk-XJ7JVCHB.js.map → chunk-3A246TSA.js.map} +0 -0
- /package/dist/{chunk-LMJ2TGWJ.js.map → chunk-JM2SKQMS.js.map} +0 -0
- /package/dist/{chunk-D7AEXSM5.js.map → chunk-PICTDURQ.js.map} +0 -0
- /package/dist/{chunk-GERDAIAL.js.map → chunk-QWMPPZ3X.js.map} +0 -0
- /package/dist/{run-campaign-OBXSN5WK.js.map → run-campaign-HNFPJET4.js.map} +0 -0
package/dist/campaign/index.d.ts
CHANGED
|
@@ -1,5 +1,5 @@
|
|
|
1
1
|
import { z } from 'zod';
|
|
2
|
-
import { AxAIService, AxFunction } from '@ax-llm/ax';
|
|
2
|
+
import { AxAIService, AxAgentActorTurnCallbackArgs, AxFunction, AxAIArgs } from '@ax-llm/ax';
|
|
3
3
|
import { AgentProfile } from '@tangle-network/agent-interface';
|
|
4
4
|
|
|
5
5
|
/**
|
|
@@ -1279,12 +1279,14 @@ interface AnalystCost {
|
|
|
1279
1279
|
est_usd_per_run?: number;
|
|
1280
1280
|
/** Models the analyst expects to use (informational). */
|
|
1281
1281
|
models?: string[];
|
|
1282
|
+
/** Maximum post-cancellation wait for provider usage. Model analysts default to 5 seconds. */
|
|
1283
|
+
settlement_timeout_ms?: number;
|
|
1282
1284
|
}
|
|
1283
1285
|
interface AnalystContext {
|
|
1284
1286
|
runId: string;
|
|
1285
1287
|
/** Stable correlation id so logs from a single registry.run() share a tag. */
|
|
1286
1288
|
correlationId: string;
|
|
1287
|
-
/**
|
|
1289
|
+
/** Enforced wall-clock deadline (epoch ms). */
|
|
1288
1290
|
deadlineMs?: number;
|
|
1289
1291
|
/** Per-analyst USD budget. Analysts MAY check before issuing LLM calls. */
|
|
1290
1292
|
budgetUsd?: number;
|
|
@@ -1453,6 +1455,10 @@ interface Scenario {
|
|
|
1453
1455
|
kind: string;
|
|
1454
1456
|
tags?: string[];
|
|
1455
1457
|
}
|
|
1458
|
+
/** Redacted identity of a complete scenario payload retained in campaign results. */
|
|
1459
|
+
interface CampaignScenarioIdentity extends Pick<Scenario, 'id' | 'kind'> {
|
|
1460
|
+
scenarioDigest: `sha256:${string}`;
|
|
1461
|
+
}
|
|
1456
1462
|
/** Context handed to every dispatch invocation. Scoped — every
|
|
1457
1463
|
* trace/span carries the cellId, every artifact write lands under the cell's
|
|
1458
1464
|
* artifact root, the cost meter accumulates per cell. */
|
|
@@ -2035,7 +2041,11 @@ interface CampaignAggregates {
|
|
|
2035
2041
|
interface CampaignResult<TArtifact = unknown, TScenario extends Scenario = Scenario> {
|
|
2036
2042
|
/** sha256(scenarios, judges, dispatch source ref, optimizer config, seed). Stable identity for reruns. */
|
|
2037
2043
|
manifestHash: string;
|
|
2044
|
+
/** Canonical identity of the exact scenario payloads and replicate count. */
|
|
2045
|
+
splitDigest: `sha256:${string}`;
|
|
2038
2046
|
seed: number;
|
|
2047
|
+
/** Replicates designed for every scenario in this campaign. */
|
|
2048
|
+
reps: number;
|
|
2039
2049
|
startedAt: string;
|
|
2040
2050
|
endedAt: string;
|
|
2041
2051
|
durationMs: number;
|
|
@@ -2049,11 +2059,9 @@ interface CampaignResult<TArtifact = unknown, TScenario extends Scenario = Scena
|
|
|
2049
2059
|
prUrl?: string;
|
|
2050
2060
|
runDir: string;
|
|
2051
2061
|
artifactsByPath: Record<string, string>;
|
|
2052
|
-
/**
|
|
2053
|
-
*
|
|
2054
|
-
|
|
2055
|
-
* want narrowed types when extending `CampaignResult`. */
|
|
2056
|
-
scenarios: Array<Pick<TScenario, 'id' | 'kind'>>;
|
|
2062
|
+
/** Redacted identities that let consumers verify the exact scenario payloads
|
|
2063
|
+
* without retaining customer task content in the result. */
|
|
2064
|
+
scenarios: Array<CampaignScenarioIdentity & Pick<TScenario, 'id' | 'kind'>>;
|
|
2057
2065
|
}
|
|
2058
2066
|
|
|
2059
2067
|
/**
|
|
@@ -2157,11 +2165,11 @@ interface AnalyzeTracesInput {
|
|
|
2157
2165
|
question: string;
|
|
2158
2166
|
}
|
|
2159
2167
|
interface AnalyzeTracesResult {
|
|
2160
|
-
/** The
|
|
2168
|
+
/** The actor's submitted prose answer. */
|
|
2161
2169
|
answer: string;
|
|
2162
|
-
/** Bulleted findings
|
|
2170
|
+
/** Bulleted findings from the actor's structured completion. */
|
|
2163
2171
|
findings: string[];
|
|
2164
|
-
/** Per-
|
|
2172
|
+
/** Per-turn snapshots captured via `actorTurnCallback`. */
|
|
2165
2173
|
turns: AnalyzeTracesTurnSnapshot[];
|
|
2166
2174
|
/** Total turns the actor took. */
|
|
2167
2175
|
turnCount: number;
|
|
@@ -2187,13 +2195,14 @@ interface TraceAnalystChatMessage {
|
|
|
2187
2195
|
[key: string]: unknown;
|
|
2188
2196
|
}
|
|
2189
2197
|
interface AnalyzeTracesTurnSnapshot {
|
|
2198
|
+
stage: AxAgentActorTurnCallbackArgs['stage'];
|
|
2190
2199
|
turn: number;
|
|
2191
2200
|
isError: boolean;
|
|
2192
2201
|
/** The JS code the actor produced for this turn. */
|
|
2193
2202
|
code: string;
|
|
2194
2203
|
/** The formatted action-log entry the actor sees on the next turn. */
|
|
2195
2204
|
output: string;
|
|
2196
|
-
/** Provider thought (when `
|
|
2205
|
+
/** Provider thought (when `executorOptions.showThoughts` is true and the
|
|
2197
2206
|
* provider returns it). */
|
|
2198
2207
|
thought?: string;
|
|
2199
2208
|
}
|
|
@@ -2202,18 +2211,18 @@ interface AnalyzeTracesOptions {
|
|
|
2202
2211
|
source: string | TraceAnalysisStore;
|
|
2203
2212
|
/** Caller-provided AxAIService. */
|
|
2204
2213
|
ai: AxAIService;
|
|
2205
|
-
/** Model id forwarded to actor
|
|
2214
|
+
/** Model id forwarded to the actor. */
|
|
2206
2215
|
model?: string;
|
|
2207
|
-
/**
|
|
2208
|
-
|
|
2216
|
+
/** Maximum model subqueries. 0 disables model fan-out. Default 4. */
|
|
2217
|
+
maxSubqueries?: number;
|
|
2209
2218
|
/** Maximum actor turns. Default 12. */
|
|
2210
2219
|
maxTurns?: number;
|
|
2211
|
-
/** Maximum parallel
|
|
2212
|
-
|
|
2220
|
+
/** Maximum parallel model subqueries. Default 2. */
|
|
2221
|
+
maxParallelSubqueries?: number;
|
|
2222
|
+
/** Cancels in-flight model and tool work. */
|
|
2223
|
+
signal?: AbortSignal;
|
|
2213
2224
|
/** Override the actor description. */
|
|
2214
2225
|
actorDescription?: string;
|
|
2215
|
-
/** Override the subagent description. */
|
|
2216
|
-
subagentDescription?: string;
|
|
2217
2226
|
/** Per-turn observability hook. */
|
|
2218
2227
|
onTurn?: (turn: AnalyzeTracesTurnSnapshot) => void | Promise<void>;
|
|
2219
2228
|
/** Override max runtime characters per turn. Default 6000. */
|
|
@@ -2221,8 +2230,7 @@ interface AnalyzeTracesOptions {
|
|
|
2221
2230
|
/** When set, every turn's snapshot is appended to this JSONL file
|
|
2222
2231
|
* immediately. If the analyst crashes mid-loop (provider 503,
|
|
2223
2232
|
* network error, validator reject) the partial reasoning is still
|
|
2224
|
-
* on disk
|
|
2225
|
-
* evidence. */
|
|
2233
|
+
* on disk for diagnosis and recovery. */
|
|
2226
2234
|
progressLogPath?: string;
|
|
2227
2235
|
}
|
|
2228
2236
|
|
|
@@ -2382,6 +2390,17 @@ interface OpenAutoPrResult {
|
|
|
2382
2390
|
*/
|
|
2383
2391
|
declare function openAutoPr<TArtifact, TScenario extends Scenario>(options: OpenAutoPrOptions<TArtifact, TScenario>): OpenAutoPrResult;
|
|
2384
2392
|
|
|
2393
|
+
/** Reject campaign designs whose denominator cannot be identified exactly. */
|
|
2394
|
+
declare function assertCampaignDesign<TScenario extends Scenario>(scenarios: readonly TScenario[], reps: number): void;
|
|
2395
|
+
/** Redacted but independently verifiable identity of one complete scenario. */
|
|
2396
|
+
declare function campaignScenarioIdentity<TScenario extends Scenario>(scenario: TScenario): CampaignScenarioIdentity & Pick<TScenario, 'id' | 'kind'>;
|
|
2397
|
+
/** Canonical split identity reconstructed from redacted scenario identities. */
|
|
2398
|
+
declare function campaignSplitDigestFromIdentities(scenarios: readonly CampaignScenarioIdentity[], reps: number): `sha256:${string}`;
|
|
2399
|
+
/** Canonical identity of the exact scenario payloads and replicate count. */
|
|
2400
|
+
declare function campaignSplitDigest<TScenario extends Scenario>(scenarios: readonly TScenario[], reps: number): `sha256:${string}`;
|
|
2401
|
+
/** Refuse a campaign whose retained task identities contradict its split digest. */
|
|
2402
|
+
declare function assertCampaignSplitIdentity(scenarios: readonly CampaignScenarioIdentity[], reps: number, splitDigest: string): void;
|
|
2403
|
+
|
|
2385
2404
|
interface ContinuousAgreement {
|
|
2386
2405
|
/** Cohen's κ_w with quadratic weights, computed on raw [0,1] scores. */
|
|
2387
2406
|
weightedKappa: number;
|
|
@@ -2975,6 +2994,7 @@ interface CampaignRunPlanCell {
|
|
|
2975
2994
|
}
|
|
2976
2995
|
interface CampaignRunPlan {
|
|
2977
2996
|
manifestHash: string;
|
|
2997
|
+
splitDigest: `sha256:${string}`;
|
|
2978
2998
|
totalCells: number;
|
|
2979
2999
|
cellsCached: number;
|
|
2980
3000
|
cellsToRun: number;
|
|
@@ -4084,17 +4104,6 @@ interface ParameterSweepProposerOptions {
|
|
|
4084
4104
|
/** Config/parameter-level proposer for FAPO's middle escalation level. */
|
|
4085
4105
|
declare function parameterSweepProposer(opts: ParameterSweepProposerOptions): SurfaceProposer;
|
|
4086
4106
|
|
|
4087
|
-
/** Fail when a code surface does not carry a complete immutable identity. */
|
|
4088
|
-
declare function assertCodeSurfaceIdentity(surface: unknown): asserts surface is CodeSurface;
|
|
4089
|
-
/** Canonical, location-independent identity of a finalized code candidate.
|
|
4090
|
-
* Commit metadata is excluded: two commits with the same base, final tree,
|
|
4091
|
-
* and patch bytes are the same executable candidate. */
|
|
4092
|
-
declare function codeSurfaceIdentityMaterial(surface: CodeSurface): string;
|
|
4093
|
-
/** Full SHA-256 content identity for a prompt or finalized code surface. */
|
|
4094
|
-
declare function surfaceContentHash(surface: MutableSurface): `sha256:${string}`;
|
|
4095
|
-
/** Short loop key derived from the same content identity as provenance. */
|
|
4096
|
-
declare function surfaceHash(surface: MutableSurface): string;
|
|
4097
|
-
|
|
4098
4107
|
/**
|
|
4099
4108
|
* `runOptimization` — the improvement loop body. Runs N generations: the
|
|
4100
4109
|
* `SurfaceProposer` proposes K candidate surfaces per generation, each
|
|
@@ -4219,8 +4228,8 @@ declare function runOptimization<TScenario extends Scenario, TArtifact>(opts: Ru
|
|
|
4219
4228
|
*
|
|
4220
4229
|
* Hard-refuses unsafe configurations:
|
|
4221
4230
|
* - `tracing: 'off'` when a proposer is wired (improvement is unattributable)
|
|
4222
|
-
* - `autoOnPromote: 'config'` —
|
|
4223
|
-
*
|
|
4231
|
+
* - `autoOnPromote: 'config'` — live mutation is unsupported without
|
|
4232
|
+
* isolated deployment, rollback, and independent validation.
|
|
4224
4233
|
*/
|
|
4225
4234
|
|
|
4226
4235
|
type RunImprovementLoopOptions<TScenario extends Scenario, TArtifact> = RunOptimizationOptions<TScenario, TArtifact> & {
|
|
@@ -4234,15 +4243,11 @@ type RunImprovementLoopOptions<TScenario extends Scenario, TArtifact> = RunOptim
|
|
|
4234
4243
|
/** What to do when the gate ships:
|
|
4235
4244
|
* - `'pr'`: open a PR via `openAutoPr`
|
|
4236
4245
|
* - `'none'`: just report — caller decides what to do with the winner
|
|
4237
|
-
*
|
|
4238
|
-
* deferred to Pass B behind safety stack. */
|
|
4246
|
+
* Live-runtime self-mutation is intentionally unsupported. */
|
|
4239
4247
|
autoOnPromote: 'pr' | 'none';
|
|
4240
4248
|
/** GH owner / repo for the auto-PR. Required when autoOnPromote === 'pr'. */
|
|
4241
4249
|
ghOwner?: string;
|
|
4242
4250
|
ghRepo?: string;
|
|
4243
|
-
/** Optional render override — substrate writes a diff-shaped surface; pass
|
|
4244
|
-
* a function to format the promoted surface differently. */
|
|
4245
|
-
renderPromotedDiff?: (winnerSurface: MutableSurface, baselineSurface: MutableSurface) => string;
|
|
4246
4251
|
/** Placebo control. When supplied AND the winner differs from baseline, the
|
|
4247
4252
|
* loop scores a THIRD holdout arm: the winner surface with its content
|
|
4248
4253
|
* footprint-matched-blanked by this function (typically via `neutralizeText`).
|
|
@@ -4255,6 +4260,8 @@ type RunImprovementLoopOptions<TScenario extends Scenario, TArtifact> = RunOptim
|
|
|
4255
4260
|
interface RunImprovementLoopResult<TArtifact, TScenario extends Scenario> extends RunOptimizationResult<TArtifact, TScenario> {
|
|
4256
4261
|
baselineOnHoldout: CampaignResult<TArtifact, TScenario>;
|
|
4257
4262
|
winnerOnHoldout: CampaignResult<TArtifact, TScenario>;
|
|
4263
|
+
neutralizedOnHoldout?: CampaignResult<TArtifact, TScenario>;
|
|
4264
|
+
neutralizedSurface?: MutableSurface;
|
|
4258
4265
|
gateResult: Awaited<ReturnType<Gate<TArtifact, TScenario>['decide']>>;
|
|
4259
4266
|
/** Unified baseline→winner surface diff. Computed UNCONDITIONALLY (not only
|
|
4260
4267
|
* when `autoOnPromote === 'pr'`) so the diff that the gate decided on is
|
|
@@ -4267,10 +4274,6 @@ interface RunImprovementLoopResult<TArtifact, TScenario extends Scenario> extend
|
|
|
4267
4274
|
* Gated-promotion shell over `runOptimization`: scores the winner against the baseline on a holdout set, runs the release gate, and optionally opens a PR.
|
|
4268
4275
|
*/
|
|
4269
4276
|
declare function runImprovementLoop<TScenario extends Scenario, TArtifact>(opts: RunImprovementLoopOptions<TScenario, TArtifact>): Promise<RunImprovementLoopResult<TArtifact, TScenario>>;
|
|
4270
|
-
/**
|
|
4271
|
-
* Default surface diff renderer: produces a unified baseline/winner text diff for prompt surfaces or a worktree-ref summary for code surfaces.
|
|
4272
|
-
*/
|
|
4273
|
-
declare function defaultRenderDiff(winnerSurface: MutableSurface, baselineSurface: MutableSurface): string;
|
|
4274
4277
|
|
|
4275
4278
|
/**
|
|
4276
4279
|
* `compareProposers` — a head-to-head lift benchmark across surface proposers
|
|
@@ -5912,8 +5915,8 @@ type CanonicalRawAnalystFinding = z.infer<typeof CanonicalRawAnalystFindingSchem
|
|
|
5912
5915
|
* Analyst-kind factory — the typed way to define trace analysts.
|
|
5913
5916
|
*
|
|
5914
5917
|
* A "kind" is a specialized analyst whose actor prompt, tool subset,
|
|
5915
|
-
* and Ax
|
|
5916
|
-
* classification, knowledge gap discovery, knowledge poisoning,
|
|
5918
|
+
* and bounded Ax subqueries target one failure-mode lens (failure-mode
|
|
5919
|
+
* classification, knowledge gap discovery, knowledge poisoning,
|
|
5917
5920
|
* self-improvement, ...). Kinds emit findings in the typed
|
|
5918
5921
|
* `CanonicalRawAnalystFinding` shape via a JSON-array Ax output; the factory
|
|
5919
5922
|
* validates each row with Zod and lifts it into `AnalystFinding[]`.
|
|
@@ -5924,12 +5927,11 @@ type CanonicalRawAnalystFinding = z.infer<typeof CanonicalRawAnalystFindingSchem
|
|
|
5924
5927
|
* - Each kind picks a narrow tool subset from `ANALYST_TOOL_GROUPS`.
|
|
5925
5928
|
* A kind that never needs full-trace dumps can drop `viewTrace` /
|
|
5926
5929
|
* `viewSpans` and stay cheap.
|
|
5927
|
-
* - Each kind declares its
|
|
5928
|
-
*
|
|
5929
|
-
* (poisoning) usually stay at 0 since they have a tighter brief.
|
|
5930
|
+
* - Each kind declares its subquery + parallelism budget. Discovery-heavy
|
|
5931
|
+
* kinds can fan out more bounded semantic questions than narrow lenses.
|
|
5930
5932
|
*
|
|
5931
5933
|
* Optimizer hook: kinds may declare `goldens` — labeled examples used
|
|
5932
|
-
* by `
|
|
5934
|
+
* by `AxBootstrapFewShot` / `AxGEPA` to fit the actor
|
|
5933
5935
|
* description programmatically. Stored on the kind, not the registry,
|
|
5934
5936
|
* because the right metric is kind-specific.
|
|
5935
5937
|
*/
|
|
@@ -5949,20 +5951,18 @@ interface TraceAnalystKindSpec {
|
|
|
5949
5951
|
version: string;
|
|
5950
5952
|
/** Actor system prompt. Must instruct the LLM to emit `findings` per the schema. */
|
|
5951
5953
|
actorDescription: string;
|
|
5952
|
-
/** Responder system prompt; falls back to a minimal "format the findings" instruction. */
|
|
5953
|
-
responderDescription?: string;
|
|
5954
5954
|
/** Tool functions the actor may call. Pick narrow subsets via `ANALYST_TOOL_GROUPS`. */
|
|
5955
5955
|
buildTools: (store: TraceAnalysisStore) => AxFunction[];
|
|
5956
|
-
/**
|
|
5957
|
-
|
|
5958
|
-
|
|
5959
|
-
|
|
5956
|
+
/** Bounded semantic subqueries. `maxCalls: 0` disables model fan-out. */
|
|
5957
|
+
subqueries?: {
|
|
5958
|
+
maxCalls: number;
|
|
5959
|
+
maxParallel?: number;
|
|
5960
5960
|
};
|
|
5961
5961
|
/** Actor turn cap. Default 12. */
|
|
5962
5962
|
maxTurns?: number;
|
|
5963
5963
|
/** Runtime char cap. Default 6000. */
|
|
5964
5964
|
maxRuntimeChars?: number;
|
|
5965
|
-
/** Maximum output tokens for every actor
|
|
5965
|
+
/** Maximum output tokens for every actor and subquery model call. Default 4096. */
|
|
5966
5966
|
maxOutputTokens?: number;
|
|
5967
5967
|
/** Cost classification surfaced in `registry.list()` and budget enforcement. */
|
|
5968
5968
|
cost: AnalystCost;
|
|
@@ -6019,7 +6019,7 @@ interface TraceAnalystProposerOptions {
|
|
|
6019
6019
|
applyMaxTokens?: number;
|
|
6020
6020
|
/** Ax provider name. Default 'openai' — works for any OpenAI-compatible base
|
|
6021
6021
|
* via `apiURL`. Use 'deepseek' to hit DeepSeek's native provider. */
|
|
6022
|
-
provider?:
|
|
6022
|
+
provider?: AxAIArgs<unknown>['name'];
|
|
6023
6023
|
/** Which analyst kinds to run. Default = the full shipped suite. */
|
|
6024
6024
|
kinds?: readonly TraceAnalystKindSpec[];
|
|
6025
6025
|
/** Resolve the OTLP traces (JSONL string) the analyst should read for THIS
|
|
@@ -6654,9 +6654,8 @@ interface HostedClient {
|
|
|
6654
6654
|
* reads). The hosted `/v1/ingest/traces` endpoint receives the FULL loop,
|
|
6655
6655
|
* not just the `cost.*` spans `runCampaign` already emits per cell.
|
|
6656
6656
|
*
|
|
6657
|
-
* The record is built from the
|
|
6658
|
-
*
|
|
6659
|
-
* could drift from what the gate actually saw.
|
|
6657
|
+
* The record is built from the loop result and its settled cost receipts — no
|
|
6658
|
+
* second usage collector can contradict what the measured cells recorded.
|
|
6660
6659
|
*/
|
|
6661
6660
|
|
|
6662
6661
|
interface LoopProvenanceCandidate {
|
|
@@ -6666,6 +6665,8 @@ interface LoopProvenanceCandidate {
|
|
|
6666
6665
|
surfaceHash: string;
|
|
6667
6666
|
/** Full sha256 content hash — byte-identical-verifiable. */
|
|
6668
6667
|
contentHash: string;
|
|
6668
|
+
/** Exact scored rows that produced this candidate's search result. */
|
|
6669
|
+
campaignDigest: `sha256:${string}`;
|
|
6669
6670
|
/** Proposer label, when the proposer returned a `ProposedCandidate`. */
|
|
6670
6671
|
label?: string;
|
|
6671
6672
|
/** Proposer rationale — the "because Z". When the proposer returned a bare
|
|
@@ -6699,13 +6700,33 @@ interface LoopProvenanceBackend {
|
|
|
6699
6700
|
totalOutputTokens: number;
|
|
6700
6701
|
totalCostUsd: number;
|
|
6701
6702
|
}
|
|
6703
|
+
interface LoopProvenanceEvidence {
|
|
6704
|
+
search: {
|
|
6705
|
+
splitDigest: `sha256:${string}`;
|
|
6706
|
+
baselineCampaignDigest: `sha256:${string}`;
|
|
6707
|
+
};
|
|
6708
|
+
holdout: {
|
|
6709
|
+
splitDigest: `sha256:${string}`;
|
|
6710
|
+
baselineCampaignDigest: `sha256:${string}`;
|
|
6711
|
+
winnerCampaignDigest: `sha256:${string}`;
|
|
6712
|
+
neutralized?: {
|
|
6713
|
+
contentHash: `sha256:${string}`;
|
|
6714
|
+
campaignDigest: `sha256:${string}`;
|
|
6715
|
+
composite: number;
|
|
6716
|
+
lift: number;
|
|
6717
|
+
};
|
|
6718
|
+
};
|
|
6719
|
+
costReceiptsDigest: `sha256:${string}`;
|
|
6720
|
+
}
|
|
6702
6721
|
/**
|
|
6703
6722
|
* The durable provenance record. Aligns to the hosted `EvalRunEvent` path but
|
|
6704
6723
|
* ADDS the rationale + the explicit baseline→candidate diff (both omitted from
|
|
6705
6724
|
* the bare hosted event) + backend provenance.
|
|
6706
6725
|
*/
|
|
6707
6726
|
interface LoopProvenanceRecord {
|
|
6708
|
-
schema: 'tangle.loop-provenance
|
|
6727
|
+
schema: 'tangle.loop-provenance';
|
|
6728
|
+
/** SHA-256 over the canonical record with this field omitted. */
|
|
6729
|
+
recordDigest: `sha256:${string}`;
|
|
6709
6730
|
runId: string;
|
|
6710
6731
|
runDir: string;
|
|
6711
6732
|
timestamp: string;
|
|
@@ -6719,6 +6740,8 @@ interface LoopProvenanceRecord {
|
|
|
6719
6740
|
diff: string;
|
|
6720
6741
|
/** Every candidate across every generation, with its rationale and structured cause. */
|
|
6721
6742
|
candidates: LoopProvenanceCandidate[];
|
|
6743
|
+
/** Exact campaign, split, surface, and receipt identities behind every summary. */
|
|
6744
|
+
evidence: LoopProvenanceEvidence;
|
|
6722
6745
|
/** Baseline composite on the search split that generated the candidates. */
|
|
6723
6746
|
baselineSearchComposite: number;
|
|
6724
6747
|
/** The gate verdict — decision + reasons + contributing gates + delta. */
|
|
@@ -6729,6 +6752,7 @@ interface LoopProvenanceRecord {
|
|
|
6729
6752
|
contributingGates: Array<{
|
|
6730
6753
|
name: string;
|
|
6731
6754
|
passed: boolean;
|
|
6755
|
+
detail: unknown;
|
|
6732
6756
|
}>;
|
|
6733
6757
|
};
|
|
6734
6758
|
/** baseline-on-holdout composite mean. */
|
|
@@ -6750,9 +6774,8 @@ interface BuildLoopProvenanceArgs<TArtifact, TScenario extends Scenario> {
|
|
|
6750
6774
|
winnerSurface: MutableSurface;
|
|
6751
6775
|
winnerLabel?: string;
|
|
6752
6776
|
winnerRationale?: string;
|
|
6753
|
-
|
|
6754
|
-
|
|
6755
|
-
baselineSearchComposite: number;
|
|
6777
|
+
/** Exact baseline campaign on the search split. */
|
|
6778
|
+
baselineSearchCampaign: CampaignResult<TArtifact, TScenario>;
|
|
6756
6779
|
/** Per-generation candidate records straight off the loop result. */
|
|
6757
6780
|
generations: Array<{
|
|
6758
6781
|
generationIndex: number;
|
|
@@ -6763,18 +6786,37 @@ interface BuildLoopProvenanceArgs<TArtifact, TScenario extends Scenario> {
|
|
|
6763
6786
|
surfaces: Array<{
|
|
6764
6787
|
surfaceHash: string;
|
|
6765
6788
|
surface: MutableSurface;
|
|
6789
|
+
campaign: CampaignResult<TArtifact, TScenario>;
|
|
6766
6790
|
}>;
|
|
6767
6791
|
}>;
|
|
6768
6792
|
gate: GateResult;
|
|
6769
6793
|
baselineOnHoldout: CampaignResult<TArtifact, TScenario>;
|
|
6770
6794
|
winnerOnHoldout: CampaignResult<TArtifact, TScenario>;
|
|
6771
|
-
|
|
6772
|
-
|
|
6795
|
+
neutralizedSurface?: MutableSurface;
|
|
6796
|
+
neutralizedOnHoldout?: CampaignResult<TArtifact, TScenario>;
|
|
6797
|
+
/** Settled run-wide receipts — agent calls are the source for backend provenance. */
|
|
6798
|
+
costReceipts: ReadonlyArray<CostReceipt>;
|
|
6799
|
+
totalCostUsd: number;
|
|
6800
|
+
totalDurationMs: number;
|
|
6801
|
+
}
|
|
6802
|
+
interface LoopProvenanceArgsFromResult<TArtifact, TScenario extends Scenario> {
|
|
6803
|
+
runId: string;
|
|
6804
|
+
runDir: string;
|
|
6805
|
+
timestamp: string;
|
|
6806
|
+
baselineSurface: MutableSurface;
|
|
6807
|
+
result: RunImprovementLoopResult<TArtifact, TScenario>;
|
|
6808
|
+
costReceipts: ReadonlyArray<CostReceipt>;
|
|
6773
6809
|
totalCostUsd: number;
|
|
6774
6810
|
totalDurationMs: number;
|
|
6775
6811
|
}
|
|
6812
|
+
/** One translation from a completed improvement loop into durable evidence. */
|
|
6813
|
+
declare function loopProvenanceArgsFromResult<TArtifact, TScenario extends Scenario>(input: LoopProvenanceArgsFromResult<TArtifact, TScenario>): BuildLoopProvenanceArgs<TArtifact, TScenario>;
|
|
6776
6814
|
/** Build the durable provenance record from a completed loop result. */
|
|
6777
6815
|
declare function buildLoopProvenanceRecord<TArtifact, TScenario extends Scenario>(args: BuildLoopProvenanceArgs<TArtifact, TScenario>): LoopProvenanceRecord;
|
|
6816
|
+
/** Digest the exact campaign fields that can affect a measured comparison. */
|
|
6817
|
+
declare function campaignMeasurementDigest<TArtifact, TScenario extends Scenario>(campaign: CampaignResult<TArtifact, TScenario>): `sha256:${string}`;
|
|
6818
|
+
/** Recompute and validate the self-addressed durable record. */
|
|
6819
|
+
declare function verifyLoopProvenanceRecord(record: LoopProvenanceRecord): LoopProvenanceRecord;
|
|
6778
6820
|
/**
|
|
6779
6821
|
* Build the loop's OTLP-ingestable spans from a provenance record. One root
|
|
6780
6822
|
* span per loop (`tangle.runId`), one span per generation, one span per
|
|
@@ -7282,6 +7324,19 @@ interface SingleRunLock {
|
|
|
7282
7324
|
*/
|
|
7283
7325
|
declare function acquireSingleRunLock(opts: SingleRunLockOptions): SingleRunLock;
|
|
7284
7326
|
|
|
7327
|
+
/** Validate the immutable identity shape; the owning executor verifies the Git objects and patch. */
|
|
7328
|
+
declare function assertCodeSurfaceIdentity(surface: unknown): asserts surface is CodeSurface;
|
|
7329
|
+
/** Canonical, location-independent identity of a finalized code candidate.
|
|
7330
|
+
* Commit metadata is excluded: two commits with the same base, final tree,
|
|
7331
|
+
* and patch bytes are the same executable candidate. */
|
|
7332
|
+
declare function codeSurfaceIdentityMaterial(surface: CodeSurface): string;
|
|
7333
|
+
/** Full SHA-256 content identity for a prompt or finalized code surface. */
|
|
7334
|
+
declare function surfaceContentHash(surface: MutableSurface): `sha256:${string}`;
|
|
7335
|
+
/** Short loop key derived from the same content identity as provenance. */
|
|
7336
|
+
declare function surfaceHash(surface: MutableSurface): string;
|
|
7337
|
+
/** Canonical customer-visible description of the exact before/after surfaces. */
|
|
7338
|
+
declare function renderSurfaceDiff(winnerSurface: MutableSurface, baselineSurface: MutableSurface): string;
|
|
7339
|
+
|
|
7285
7340
|
/**
|
|
7286
7341
|
* Transient-transport-failure classification for dispatch retry policies.
|
|
7287
7342
|
*
|
|
@@ -7396,4 +7451,4 @@ declare function verifyCodeSurface(surface: CodeSurface, worktreeDir?: string):
|
|
|
7396
7451
|
* identity against the checkout at `worktreeRef`. */
|
|
7397
7452
|
declare function resolveWorktreePath(surface: CodeSurface, worktreeDir?: string): string;
|
|
7398
7453
|
|
|
7399
|
-
export { type AcceptedEdit, type AceProposerOptions, type AnalystArtifact, type AnalystScenario, type AnalyzeCrossSurfaceInteractionsInput, type ApplySkillPatchResult, type AxisEvidence, type AxisVerdict, type BuildAnalystSurfaceDispatchOptions, type BuildEvidenceVectorOptions, type BuildLoopProvenanceArgs, type CampaignAggregates, type CampaignArtifactWriter, type CampaignBreakdown, type CampaignCellResult, type CampaignCostMeter, type CampaignResult, type CampaignRunPlan, type CampaignRunPlanCell, type CampaignStorage, type CampaignTokenUsage, type CampaignTraceWriter, type CodeSurface, type CodeSurfaceVerification, type CompareProposersOptions, type CompositeProposerOptions, type CostLedgerHandle, type CrossSurfaceAdditionDecision, type CrossSurfaceAdditionRejectionReason, type CrossSurfaceAttemptCompleteness, type CrossSurfaceBestSingleSelection, type CrossSurfaceBootstrapPolicy, type CrossSurfaceCandidate, type CrossSurfaceCandidateComparison, type CrossSurfaceCandidateEvidence, type CrossSurfaceCandidateOutcome, type CrossSurfaceCandidateSummary, type CrossSurfaceComponent, type CrossSurfaceComponentEvidence, type CrossSurfaceCompositionStep, type CrossSurfaceDistribution, type CrossSurfaceEligibility, type CrossSurfaceEvidenceBreakdown, type CrossSurfaceIneligibilityReason, type CrossSurfaceInteractionAwareSelection, type CrossSurfaceInteractionEffect, type CrossSurfaceInteractionPath, type CrossSurfaceInteractionReport, type CrossSurfaceInteractionTask, type CrossSurfaceNaiveStackSelection, type CrossSurfacePairCompatibility, type CrossSurfacePairEvidence, type CrossSurfacePairIncompatibilityReason, type CrossSurfacePairwiseEntry, type CrossSurfaceRankedSingle, type CrossSurfaceRelativeCost, type CrossSurfaceSelectionPolicy, type CrossSurfaceSelections, type CrossSurfaceTaskRow, DEFAULT_POLICY_EDIT_HISTORY_LIMITS, type DefaultProductionGateOptions, type DimensionRegression, type DiscriminationScore, type DispatchContext, type DispatchFn, type EmitLoopProvenanceArgs, type EmitLoopProvenanceResult, type EvalFixture, type EvalFixtureFile, type EvalFixtureLoadOptions, type EvalFixtureRunPlan, type EvalFixtureScenario, type EvalFixtureValidationMode, type EvidenceVector, type EvolutionaryProposerOptions, type FailureModeRecallJudgeOptions, type FapoAttributionSignals, type FapoEntryConfig, type FapoFailureCluster, type FapoOptimizationLevel, type FapoProposerOptions, type FapoReviewInput, type FapoReviewIssue, type FapoReviewResult, type FapoScopeContract, FileSearchLedger, FsLabeledScenarioStore, type FsLabeledScenarioStoreOptions, type Gate, type GateContext, type GateDecision, type GateResult, type GenerationCandidate, type GenerationRecord, type GepaProposerConstraints, type GepaProposerOptions, type GitWorktreeAdapterOptions, type Governor, type GovernorContext, type GovernorOp, type HaloProposerOptions, type HeldOutGateOptions, type HeldoutSignificance, type HeldoutSignificanceOptions, type HeuristicGovernorOptions, type JsonPolicyEditTargetSurface, type JsonPrimitive, type JsonValue, type JudgeAggregate, type JudgeConfig, type JudgeDimension, type JudgeScore, type LabelTrust, type LabeledScenarioRecord, type LabeledScenarioSampleArgs, type LabeledScenarioSource, type LabeledScenarioStore, LabeledScenarioStoreError, type LabeledScenarioWrite, Lineage, type LineageEdge, type LineageGraph, type LineageNode, type LineageNodeInput, type LineageStore, type LlmJudgeDimension, type LlmJudgeOptions, type LlmPolicyEditProposerOptions, type LoadEvalFixtureScenariosOptions, type LoopProvenanceBackend, type LoopProvenanceCandidate, type LoopProvenanceRecord, type MemoryCurationProposerOptions, type MutableSurface, type Mutator, type NeutralizationGateOptions, type ObjectiveSource, type OpenAutoPrOptions, type OpenAutoPrResult, type OpenSearchLedgerOptions, type OptimizationProposer, type OptimizerConfig, type OptimizerEntryConfig, POLICY_EDIT_CANDIDATE_RECORD_SCHEMA, type PairedHoldout, type ParameterCandidate, type ParameterChange, type ParameterSweepProposerOptions, type ParetoParent, type ParetoSignificanceGateOptions, type PlanCampaignRunOptions, type PlanEvalFixtureRunOptions, type PlaybackContext, type PlaybackDriver, type PlaybackStep, type PolicyEditAuthorScenarioRow, type PolicyEditCandidateRecord, type PolicyEditCandidateSummary, type PolicyEditFindingInput, type PolicyEditFindingSource, type PolicyEditHistoryCandidateContext, type PolicyEditHistoryGenerationContext, type PolicyEditHistoryProjectionOptions, type PolicyEditObjective, type PolicyEditOutcomeContext, type PolicyEditProposerOptions, type PowerPreflight, type PowerPreflightOptions, type ProfileDispatchFn, ProfileMatrixError, type ProfileSummary, type PromotionObjective, type PromotionPolicy, type ProposeContext, type ProposePatchesArgs, type ProposedCandidate, type ProposerComparison, type ProposerEntry, type ProposerPairwise, type ProposerScore, type RedactionStatus, type ReferenceEquivalenceJudgeOptions, type ReferenceEquivalenceScenario, type RejectedEdit, type RolloutArgumentDiff, type RolloutArgumentDiffOptions, type RolloutCall, type RunCampaignOptions, type RunEvalOptions, type RunImprovementLoopOptions, type RunImprovementLoopResult, type RunLineageLoopOptions, type RunLineageLoopResult, type RunLineageLoopSeed, type RunLineageOptions, type RunLineageResult, type RunLineageSeed, type RunLineageStepResult, type RunOptimizationOptions, type RunOptimizationResult, type RunProfileMatrixOptions, type RunProfileMatrixResult, type RunSkillOptOptions, type RunSkillOptResult, SEARCH_LEDGER_SCHEMA, type Scenario, type ScenarioAggregate, type ScenarioRollup, type ScenarioSignal, type ScoreboardRenderOptions, type ScoreboardRow, type ScoreboardSummary, type ScoredRollout, type ScoredSurfaceOutcome, type SearchAccountingAudit, type SearchArtifactRef, type SearchAttemptAccounting, type SearchCandidateDecidedEvent, type SearchCandidateLineage, type SearchCandidateRegisteredEvent, type SearchCandidateSlot, type SearchCandidateSlotClosedEvent, type SearchCandidateSurface, type SearchCompletedEvent, type SearchCostAccounting, type SearchFailureReason, type SearchLedger, type SearchLedgerAppendResult, SearchLedgerConflictError, type SearchLedgerEntry, SearchLedgerError, type SearchLedgerEvent, type SearchLedgerHash, SearchLedgerIntegrityError, type SearchLedgerReplay, type SearchModelIdentity, type SearchOperationKind, type SearchOperationRecordedEvent, type SearchPlan, type SearchPlannedEvent, type SearchPlannedOperation, type SearchPlannedTask, type SearchSourceRef, type SearchSurfaceEffect, type SearchSurfaceEvidence, type SearchSurfaceKind, type SearchTaskAttemptedEvent, type SearchTaskOutcome, type SearchTokenAccounting, type SelectPolicyEditAuthorRowsOptions, type SequentialDecideFn, type SequentialDecideOptions, type SequentialDecision, type SequentialObservation, type SequentialPairedGate, type SequentialPairedGateOptions, type SerializedJsonBudget, type SessionScript, type SingleRunLock, type SingleRunLockOptions, type SkillOptEpochRecord, type SkillOptEvidence, type SkillOptProposer, type SkillOptProposerOptions, type SkillPatch, type SkillPatchOp, SkillPatchParseError, type SkillPatchRejection, type SurfaceProposer, type SurfaceScore, type TraceAnalystProposerOptions, type TraceSpan, type TransientFailureOptions, type UngroundedLiteralReport, type UserStory, type UserStoryVerdict, type Worktree, type WorktreeAdapter, WorktreeAdapterError, aceProposer, acquireSingleRunLock, analyzeCrossSurfaceInteractions, applySkillPatch, assertCodeSurfaceIdentity, assertPolicyEditAuthorContextBudget, buildAnalystSurfaceDispatch, buildEvidenceVector, buildLoopProvenanceRecord, callbackGovernor, campaignBreakdown, campaignMeanComposite, classifyUngroundedLiterals, codeSurfaceIdentityMaterial, compareProposers, composeGate, compositeProposer, countSentenceEdits, createReferenceEquivalenceJudge, createRunCostLedger, defaultProductionGate,
|
|
7454
|
+
export { type AcceptedEdit, type AceProposerOptions, type AnalystArtifact, type AnalystScenario, type AnalyzeCrossSurfaceInteractionsInput, type ApplySkillPatchResult, type AxisEvidence, type AxisVerdict, type BuildAnalystSurfaceDispatchOptions, type BuildEvidenceVectorOptions, type BuildLoopProvenanceArgs, type CampaignAggregates, type CampaignArtifactWriter, type CampaignBreakdown, type CampaignCellResult, type CampaignCostMeter, type CampaignResult, type CampaignRunPlan, type CampaignRunPlanCell, type CampaignScenarioIdentity, type CampaignStorage, type CampaignTokenUsage, type CampaignTraceWriter, type CodeSurface, type CodeSurfaceVerification, type CompareProposersOptions, type CompositeProposerOptions, type CostLedgerHandle, type CrossSurfaceAdditionDecision, type CrossSurfaceAdditionRejectionReason, type CrossSurfaceAttemptCompleteness, type CrossSurfaceBestSingleSelection, type CrossSurfaceBootstrapPolicy, type CrossSurfaceCandidate, type CrossSurfaceCandidateComparison, type CrossSurfaceCandidateEvidence, type CrossSurfaceCandidateOutcome, type CrossSurfaceCandidateSummary, type CrossSurfaceComponent, type CrossSurfaceComponentEvidence, type CrossSurfaceCompositionStep, type CrossSurfaceDistribution, type CrossSurfaceEligibility, type CrossSurfaceEvidenceBreakdown, type CrossSurfaceIneligibilityReason, type CrossSurfaceInteractionAwareSelection, type CrossSurfaceInteractionEffect, type CrossSurfaceInteractionPath, type CrossSurfaceInteractionReport, type CrossSurfaceInteractionTask, type CrossSurfaceNaiveStackSelection, type CrossSurfacePairCompatibility, type CrossSurfacePairEvidence, type CrossSurfacePairIncompatibilityReason, type CrossSurfacePairwiseEntry, type CrossSurfaceRankedSingle, type CrossSurfaceRelativeCost, type CrossSurfaceSelectionPolicy, type CrossSurfaceSelections, type CrossSurfaceTaskRow, DEFAULT_POLICY_EDIT_HISTORY_LIMITS, type DefaultProductionGateOptions, type DimensionRegression, type DiscriminationScore, type DispatchContext, type DispatchFn, type EmitLoopProvenanceArgs, type EmitLoopProvenanceResult, type EvalFixture, type EvalFixtureFile, type EvalFixtureLoadOptions, type EvalFixtureRunPlan, type EvalFixtureScenario, type EvalFixtureValidationMode, type EvidenceVector, type EvolutionaryProposerOptions, type FailureModeRecallJudgeOptions, type FapoAttributionSignals, type FapoEntryConfig, type FapoFailureCluster, type FapoOptimizationLevel, type FapoProposerOptions, type FapoReviewInput, type FapoReviewIssue, type FapoReviewResult, type FapoScopeContract, FileSearchLedger, FsLabeledScenarioStore, type FsLabeledScenarioStoreOptions, type Gate, type GateContext, type GateDecision, type GateResult, type GenerationCandidate, type GenerationRecord, type GepaProposerConstraints, type GepaProposerOptions, type GitWorktreeAdapterOptions, type Governor, type GovernorContext, type GovernorOp, type HaloProposerOptions, type HeldOutGateOptions, type HeldoutSignificance, type HeldoutSignificanceOptions, type HeuristicGovernorOptions, type JsonPolicyEditTargetSurface, type JsonPrimitive, type JsonValue, type JudgeAggregate, type JudgeConfig, type JudgeDimension, type JudgeScore, type LabelTrust, type LabeledScenarioRecord, type LabeledScenarioSampleArgs, type LabeledScenarioSource, type LabeledScenarioStore, LabeledScenarioStoreError, type LabeledScenarioWrite, Lineage, type LineageEdge, type LineageGraph, type LineageNode, type LineageNodeInput, type LineageStore, type LlmJudgeDimension, type LlmJudgeOptions, type LlmPolicyEditProposerOptions, type LoadEvalFixtureScenariosOptions, type LoopProvenanceArgsFromResult, type LoopProvenanceBackend, type LoopProvenanceCandidate, type LoopProvenanceEvidence, type LoopProvenanceRecord, type MemoryCurationProposerOptions, type MutableSurface, type Mutator, type NeutralizationGateOptions, type ObjectiveSource, type OpenAutoPrOptions, type OpenAutoPrResult, type OpenSearchLedgerOptions, type OptimizationProposer, type OptimizerConfig, type OptimizerEntryConfig, POLICY_EDIT_CANDIDATE_RECORD_SCHEMA, type PairedHoldout, type ParameterCandidate, type ParameterChange, type ParameterSweepProposerOptions, type ParetoParent, type ParetoSignificanceGateOptions, type PlanCampaignRunOptions, type PlanEvalFixtureRunOptions, type PlaybackContext, type PlaybackDriver, type PlaybackStep, type PolicyEditAuthorScenarioRow, type PolicyEditCandidateRecord, type PolicyEditCandidateSummary, type PolicyEditFindingInput, type PolicyEditFindingSource, type PolicyEditHistoryCandidateContext, type PolicyEditHistoryGenerationContext, type PolicyEditHistoryProjectionOptions, type PolicyEditObjective, type PolicyEditOutcomeContext, type PolicyEditProposerOptions, type PowerPreflight, type PowerPreflightOptions, type ProfileDispatchFn, ProfileMatrixError, type ProfileSummary, type PromotionObjective, type PromotionPolicy, type ProposeContext, type ProposePatchesArgs, type ProposedCandidate, type ProposerComparison, type ProposerEntry, type ProposerPairwise, type ProposerScore, type RedactionStatus, type ReferenceEquivalenceJudgeOptions, type ReferenceEquivalenceScenario, type RejectedEdit, type RolloutArgumentDiff, type RolloutArgumentDiffOptions, type RolloutCall, type RunCampaignOptions, type RunEvalOptions, type RunImprovementLoopOptions, type RunImprovementLoopResult, type RunLineageLoopOptions, type RunLineageLoopResult, type RunLineageLoopSeed, type RunLineageOptions, type RunLineageResult, type RunLineageSeed, type RunLineageStepResult, type RunOptimizationOptions, type RunOptimizationResult, type RunProfileMatrixOptions, type RunProfileMatrixResult, type RunSkillOptOptions, type RunSkillOptResult, SEARCH_LEDGER_SCHEMA, type Scenario, type ScenarioAggregate, type ScenarioRollup, type ScenarioSignal, type ScoreboardRenderOptions, type ScoreboardRow, type ScoreboardSummary, type ScoredRollout, type ScoredSurfaceOutcome, type SearchAccountingAudit, type SearchArtifactRef, type SearchAttemptAccounting, type SearchCandidateDecidedEvent, type SearchCandidateLineage, type SearchCandidateRegisteredEvent, type SearchCandidateSlot, type SearchCandidateSlotClosedEvent, type SearchCandidateSurface, type SearchCompletedEvent, type SearchCostAccounting, type SearchFailureReason, type SearchLedger, type SearchLedgerAppendResult, SearchLedgerConflictError, type SearchLedgerEntry, SearchLedgerError, type SearchLedgerEvent, type SearchLedgerHash, SearchLedgerIntegrityError, type SearchLedgerReplay, type SearchModelIdentity, type SearchOperationKind, type SearchOperationRecordedEvent, type SearchPlan, type SearchPlannedEvent, type SearchPlannedOperation, type SearchPlannedTask, type SearchSourceRef, type SearchSurfaceEffect, type SearchSurfaceEvidence, type SearchSurfaceKind, type SearchTaskAttemptedEvent, type SearchTaskOutcome, type SearchTokenAccounting, type SelectPolicyEditAuthorRowsOptions, type SequentialDecideFn, type SequentialDecideOptions, type SequentialDecision, type SequentialObservation, type SequentialPairedGate, type SequentialPairedGateOptions, type SerializedJsonBudget, type SessionScript, type SingleRunLock, type SingleRunLockOptions, type SkillOptEpochRecord, type SkillOptEvidence, type SkillOptProposer, type SkillOptProposerOptions, type SkillPatch, type SkillPatchOp, SkillPatchParseError, type SkillPatchRejection, type SurfaceProposer, type SurfaceScore, type TraceAnalystProposerOptions, type TraceSpan, type TransientFailureOptions, type UngroundedLiteralReport, type UserStory, type UserStoryVerdict, type Worktree, type WorktreeAdapter, WorktreeAdapterError, aceProposer, acquireSingleRunLock, analyzeCrossSurfaceInteractions, applySkillPatch, assertCampaignDesign, assertCampaignSplitIdentity, assertCodeSurfaceIdentity, assertPolicyEditAuthorContextBudget, buildAnalystSurfaceDispatch, buildEvidenceVector, buildLoopProvenanceRecord, callbackGovernor, campaignBreakdown, campaignMeanComposite, campaignMeasurementDigest, campaignScenarioIdentity, campaignSplitDigest, campaignSplitDigestFromIdentities, classifyUngroundedLiterals, codeSurfaceIdentityMaterial, compareProposers, composeGate, compositeProposer, countSentenceEdits, createReferenceEquivalenceJudge, createRunCostLedger, defaultProductionGate, detectScale, dimensionRegressions, discoverEvalFixtures, emitLoopProvenance, evolutionaryProposer, extractFapoAttributionSignals, extractH2Sections, failureModeRecallJudge, fapoEscalationEntry, fapoProposer, fsCampaignStorage, fsLineageStore, gepaParetoEntry, gepaProposer, gepaReflectionEntry, gitWorktreeAdapter, haloProposer, heldOutGate, heldoutSignificance, heuristicGovernor, inMemoryCampaignStorage, isProposedCandidate, isTransientTransportFailure, labelTrustRank, lineageNodeId, llmJudge, llmPolicyEditProposer, loadEvalFixture, loadEvalFixtureScenarios, loopProvenanceArgsFromResult, loopProvenanceSpans, makePlaybackDispatch, memLineageStore, memoryCurationProposer, neutralizationGate, neutralizeText, openAutoPr, openSearchLedger, pairHoldout, parameterSweepProposer, paretoPolicy, paretoSignificanceGate, parseSkillPatchResponse, patchEditCount, planCampaignRun, planEvalFixtureRun, policyEditProposer, powerPreflight, projectPolicyEditHistory, provenanceRecordPath, provenanceSpansPath, renderScoreboardMarkdown, renderSurfaceDiff, resolveRunDir, resolveWorktreePath, rolloutArgumentDiff, runCampaign, runEval, runImprovementLoop, runLineage, runLineageLoop, runOptimization, runProfileMatrix, runSkillOpt, scoreDiscrimination, scoreUserStory, scoreboardSummary, selectDiscriminative, selectPolicyEditAuthorRows, sequentialDecide, sequentialPairedGate, skillOptEntry, skillOptProposer, surfaceContentHash, surfaceHash, tangleTracesRoot, traceAnalystProposer, userStoryScoreboard, validatePolicyEditCandidateRecord, validateSearchLedgerEvent, verifyCodeSurface, verifyLoopProvenanceRecord };
|
package/dist/campaign/index.js
CHANGED
|
@@ -66,19 +66,19 @@ import {
|
|
|
66
66
|
userStoryScoreboard,
|
|
67
67
|
validateSearchLedgerEvent,
|
|
68
68
|
verifyCodeSurface
|
|
69
|
-
} from "../chunk-
|
|
69
|
+
} from "../chunk-JN2FCO5W.js";
|
|
70
70
|
import {
|
|
71
71
|
assertCodeSurfaceIdentity,
|
|
72
72
|
buildEvidenceVector,
|
|
73
73
|
buildLoopProvenanceRecord,
|
|
74
74
|
campaignBreakdown,
|
|
75
75
|
campaignMeanComposite,
|
|
76
|
+
campaignMeasurementDigest,
|
|
76
77
|
codeSurfaceIdentityMaterial,
|
|
77
78
|
composeGate,
|
|
78
79
|
countSentenceEdits,
|
|
79
80
|
createReferenceEquivalenceJudge,
|
|
80
81
|
defaultProductionGate,
|
|
81
|
-
defaultRenderDiff,
|
|
82
82
|
detectScale,
|
|
83
83
|
dimensionRegressions,
|
|
84
84
|
emitLoopProvenance,
|
|
@@ -90,6 +90,7 @@ import {
|
|
|
90
90
|
isProposedCandidate,
|
|
91
91
|
labelTrustRank,
|
|
92
92
|
llmJudge,
|
|
93
|
+
loopProvenanceArgsFromResult,
|
|
93
94
|
loopProvenanceSpans,
|
|
94
95
|
openAutoPr,
|
|
95
96
|
pairHoldout,
|
|
@@ -98,16 +99,23 @@ import {
|
|
|
98
99
|
powerPreflight,
|
|
99
100
|
provenanceRecordPath,
|
|
100
101
|
provenanceSpansPath,
|
|
102
|
+
renderSurfaceDiff,
|
|
101
103
|
runEval,
|
|
102
104
|
runImprovementLoop,
|
|
103
105
|
runOptimization,
|
|
104
106
|
surfaceContentHash,
|
|
105
|
-
surfaceHash
|
|
106
|
-
|
|
107
|
+
surfaceHash,
|
|
108
|
+
verifyLoopProvenanceRecord
|
|
109
|
+
} from "../chunk-32BZXMSO.js";
|
|
107
110
|
import {
|
|
108
111
|
SearchLedgerConflictError,
|
|
109
112
|
SearchLedgerError,
|
|
110
113
|
SearchLedgerIntegrityError,
|
|
114
|
+
assertCampaignDesign,
|
|
115
|
+
assertCampaignSplitIdentity,
|
|
116
|
+
campaignScenarioIdentity,
|
|
117
|
+
campaignSplitDigest,
|
|
118
|
+
campaignSplitDigestFromIdentities,
|
|
111
119
|
createRunCostLedger,
|
|
112
120
|
fsCampaignStorage,
|
|
113
121
|
inMemoryCampaignStorage,
|
|
@@ -115,20 +123,20 @@ import {
|
|
|
115
123
|
resolveRunDir,
|
|
116
124
|
runCampaign,
|
|
117
125
|
tangleTracesRoot
|
|
118
|
-
} from "../chunk-
|
|
126
|
+
} from "../chunk-XDIRG3TO.js";
|
|
119
127
|
import "../chunk-3YYRZDON.js";
|
|
120
128
|
import {
|
|
121
129
|
POLICY_EDIT_CANDIDATE_RECORD_SCHEMA,
|
|
122
130
|
validatePolicyEditCandidateRecord
|
|
123
|
-
} from "../chunk-
|
|
131
|
+
} from "../chunk-S5TT5R3L.js";
|
|
124
132
|
import "../chunk-ARU2PZFM.js";
|
|
125
133
|
import "../chunk-NJC7U437.js";
|
|
126
134
|
import "../chunk-PJQFMIOX.js";
|
|
127
135
|
import "../chunk-JHOJHHU7.js";
|
|
128
136
|
import "../chunk-VI2UW6B6.js";
|
|
129
|
-
import "../chunk-
|
|
137
|
+
import "../chunk-WW2A73HW.js";
|
|
130
138
|
import "../chunk-GGE4NNQT.js";
|
|
131
|
-
import "../chunk-
|
|
139
|
+
import "../chunk-PXD6ZFNY.js";
|
|
132
140
|
import "../chunk-PC4UYEBM.js";
|
|
133
141
|
import "../chunk-S3UZOQ5Y.js";
|
|
134
142
|
import "../chunk-MA6HLL3S.js";
|
|
@@ -155,6 +163,8 @@ export {
|
|
|
155
163
|
acquireSingleRunLock,
|
|
156
164
|
analyzeCrossSurfaceInteractions,
|
|
157
165
|
applySkillPatch,
|
|
166
|
+
assertCampaignDesign,
|
|
167
|
+
assertCampaignSplitIdentity,
|
|
158
168
|
assertCodeSurfaceIdentity,
|
|
159
169
|
assertPolicyEditAuthorContextBudget,
|
|
160
170
|
buildAnalystSurfaceDispatch,
|
|
@@ -163,6 +173,10 @@ export {
|
|
|
163
173
|
callbackGovernor,
|
|
164
174
|
campaignBreakdown,
|
|
165
175
|
campaignMeanComposite,
|
|
176
|
+
campaignMeasurementDigest,
|
|
177
|
+
campaignScenarioIdentity,
|
|
178
|
+
campaignSplitDigest,
|
|
179
|
+
campaignSplitDigestFromIdentities,
|
|
166
180
|
classifyUngroundedLiterals,
|
|
167
181
|
codeSurfaceIdentityMaterial,
|
|
168
182
|
compareProposers,
|
|
@@ -172,7 +186,6 @@ export {
|
|
|
172
186
|
createReferenceEquivalenceJudge,
|
|
173
187
|
createRunCostLedger,
|
|
174
188
|
defaultProductionGate,
|
|
175
|
-
defaultRenderDiff,
|
|
176
189
|
detectScale,
|
|
177
190
|
dimensionRegressions,
|
|
178
191
|
discoverEvalFixtures,
|
|
@@ -202,6 +215,7 @@ export {
|
|
|
202
215
|
llmPolicyEditProposer,
|
|
203
216
|
loadEvalFixture,
|
|
204
217
|
loadEvalFixtureScenarios,
|
|
218
|
+
loopProvenanceArgsFromResult,
|
|
205
219
|
loopProvenanceSpans,
|
|
206
220
|
makePlaybackDispatch,
|
|
207
221
|
memLineageStore,
|
|
@@ -224,6 +238,7 @@ export {
|
|
|
224
238
|
provenanceRecordPath,
|
|
225
239
|
provenanceSpansPath,
|
|
226
240
|
renderScoreboardMarkdown,
|
|
241
|
+
renderSurfaceDiff,
|
|
227
242
|
resolveRunDir,
|
|
228
243
|
resolveWorktreePath,
|
|
229
244
|
rolloutArgumentDiff,
|
|
@@ -251,6 +266,7 @@ export {
|
|
|
251
266
|
userStoryScoreboard,
|
|
252
267
|
validatePolicyEditCandidateRecord,
|
|
253
268
|
validateSearchLedgerEvent,
|
|
254
|
-
verifyCodeSurface
|
|
269
|
+
verifyCodeSurface,
|
|
270
|
+
verifyLoopProvenanceRecord
|
|
255
271
|
};
|
|
256
272
|
//# sourceMappingURL=index.js.map
|