@tangle-network/agent-eval 0.181.0 → 0.183.0

package/dist/{types-lPkDQNqJ.d.ts → types-BlIJ5URG.d.ts}

@@ -47,7 +47,7 @@ interface WorkerLogSource {
  * The difference between "the artifact is missing" and "this store never
  * records that fact" is the difference between a run that spent $0 and a
  * harness that does not price inference — and the second harness is where a
- * loops-shaped assumption becomes a fabricated zero. A reader declares its
+ * store-specific assumption becomes a fabricated zero. A reader declares its
  * limits once; the analyzer reports `unavailable` for everything downstream.
  *
  * `null` on a field means the source DOES carry that fact.
@@ -72,8 +72,8 @@ declare const NO_SOURCE_LIMITS: SourceLimits;
  * dependent metrics into `unavailable` rather than 0.
  *
  * This is the whole input contract. Any store that can produce these strings
- * (an on-disk loops run, an object-store archive, a database, a test fixture)
- * is a valid source; `loopsSupervisorRunReader` is ONE implementation.
+ * (a Runtime run, an object-store archive, a database, or a test fixture) is a
+ * valid source; the Runtime reader is one implementation.
  */
 interface SupervisorRunSources {
   /** Stable identity of the run being analyzed (a directory, a run id, a URL). */
@@ -86,13 +86,13 @@ interface SupervisorRunSources {
   /**
    * Supervision journal — spawned / settled / cancelled / metered events (JSONL).
    * Recursive readers put `role: 'supervisor' | 'worker'` on spawned rows;
-   * settled `verdict` may be a legacy string or `{ valid, score, ... }`.
+   * settled `verdict` may be a string or `{ valid, score, ... }`.
    */
   readonly journal: string | null;
   /**
    * Source-specific reason `journal` is null. The analyzer uses it verbatim as
-   * the `unavailable` reason on every journal-dependent metric, so a non-loops
-   * layout names its own journal file instead of inheriting the loops paths.
+   * the `unavailable` reason on every journal-dependent metric, so another
+   * layout names its own journal file instead of inheriting a path assumption.
    */
   readonly journalMissingReason?: string;
   /** Per-brain-call tap (JSONL): finish_reason, completion tokens, requested max tokens. */
@@ -108,16 +108,13 @@ interface SupervisorRunSources {
   /** Why `workers` is null (only set when it is). */
   readonly workersMissingReason: string | null;
   /**
-   * Run result document (JSON). For a Runtime run dir this is `result.json`,
-   * the `SupervisedResult` that `supervise()` returned, verbatim; its `kind`
-   * is the run's status. For a loops run dir it is the legacy result document.
+   * Runtime `result.json`, the `SupervisedResult` returned by `supervise()`.
    */
   readonly result: string | null;
   /**
    * Runtime's terminal failure record (`failure.json`), written when
    * `supervise()` threw before a result landed. `null` = the store was read and
-   * holds no such record; `undefined` = the store has no failure document at
-   * all (the loops layout), which the analyzer reports as its own absence.
+   * holds no such record; `undefined` = the store has no failure document.
    */
   readonly failure?: string | null;
   /**
@@ -150,9 +147,8 @@ interface SupervisorRunSources {
   /** What this store structurally cannot record. See `SourceLimits`. */
   readonly limits: SourceLimits;
   /**
-   * Where the ROOT invocation's transcript lives. Undefined lets the rollout
-   * minter fall back to the loops layout (`<supRunDir>/journal.jsonl`); any
-   * other store must say, or the row points at a path that never existed.
+   * Where the ROOT invocation's transcript lives. Any store that does not
+   * retain one must set this to `null`, or omit the path entirely.
    */
   readonly rootTranscriptRef?: string | null;
   /**
@@ -310,8 +306,7 @@ interface SpendMeasurement {
  * The run's total inference spend, measured two ways.
  *
  * `closeRecord` is the spend the store recorded as settled when the run
- * closed (loops `state.json` `result.spentUsd`; Runtime `result.json`
- * `spentTotal.usd`) — the billing-shaped answer. `journalDerived` is the
+ * closed (`result.json` `spentTotal.usd`) — the billing-shaped answer. `journalDerived` is the
  * spend execution observably consumed (journal `metered` + `settled` rows) —
  * the execution-accounting answer. Neither is canonical for the other's
  * question. The two cover different records at different moments, so
@@ -382,11 +377,7 @@ type SupervisorStatusSource =
 /** Runtime's `result.json` `kind`: the `SupervisedResult` discriminant, read verbatim. */
 'runtime-result' |
 /** Runtime's `failure.json`: `supervise()` threw before a result landed. */
-'runtime-failure' |
-/** Control-plane-era loops `state.json` `status`. */
-'legacy-state' |
-/** Control-plane-era loops `result.json` `sup_status`. */
-'legacy-result';
+'runtime-failure';
 /** The error a run directory recorded. */
 interface TerminalFailure {
   /**
@@ -411,16 +402,16 @@ interface TerminalFailure {
 }
 interface OutcomeMetrics {
   /**
-   * The run's terminal status, exactly as its record spells it: Runtime's
-   * `winner` / `no-winner`, `failed` for a Runtime failure record, or the
-   * legacy loops status. `supStatusSource` names the record it came from.
+   * The run's terminal status, exactly as its Runtime record spells it:
+   * `winner` / `no-winner`, or `failed` for a Runtime failure record.
+   * `supStatusSource` names the record it came from.
    */
   readonly supStatus: Measured<string>;
   readonly supStatusSource: Measured<SupervisorStatusSource>;
   /**
    * Runtime's `reason` on a `no-winner` result (`all-children-down`,
    * `budget-exhausted`, `aborted`, `driver-failed`). `null` = the terminal
-   * record carries no reason (a winner, a failure record, a legacy document).
+   * record carries no reason (a winner or a failure record).
    */
   readonly supReason: Measured<string | null>;
   /**
@@ -530,4 +521,4 @@ interface SupervisorRunTreeGap {
 }
 //#endregion
 export { WallDistribution as A, SupervisorRunSources as C, SupervisorStatusSource as D, SupervisorRunTreeGapCode as E, isUnavailable as M, showMeasured as N, TerminalFailure as O, unavailable as P, SupervisorRunRollup as S, SupervisorRunTreeGap as T, SpendMeasurements as _, NamedResourceReceipt as a, SupervisorRunReader as b, PatchStats as c, RoleSpend as d, RollupCellRow as f, SpendMeasurement as g, SourceLimits as h, NO_SOURCE_LIMITS as i, WorkerLogSource as j, Unavailable as k, PerWorkerRow as l, SUPERVISOR_RUN_SCHEMA as m, EconomicsMetrics as n, OrchestrationMetrics as o, SUPERVISOR_RUN_ROLLUP_SCHEMA as p, Measured as r, OutcomeMetrics as s, DecisionMetrics as t, ResourceSpendRecord as u, SteerBreakdown as v, SupervisorRunTree as w, SupervisorRunReport as x, SupervisorRunNodeRole as y };
-//# sourceMappingURL=types-lPkDQNqJ.d.ts.map
+//# sourceMappingURL=types-BlIJ5URG.d.ts.map

package/dist/types-BlIJ5URG.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types-BlIJ5URG.d.ts","names":[],"sources":["../src/supervisor-run/types.ts"],"mappings":";;;;UAkCiB;WACN;;;KAIC,SAAS,KAAK,IAAI;iBAEd,YAAY,iBAAiB;iBAI7B,cAAc,aAAa,KAAK;;iBAKhC,aAAa,GAAG;;KAWpB;;;;;UAMK;;;;;WAKN;;WAEA;;WAEA;;WAEA;;WAEA;;WAEA;;WAEA;;WAEA;WACA;WACA;WACA;;;;;;;;;;;;;UAcM;;WAEN;;WAEA;;WAEA;;WAEA;;WAEA;;;cAIE,kBAAkB;;;;;;;;;;UAiBd;;WAEN;WACA;;WAEA;;WAEA;;;;;;WAMA;;;;;;WAMA;;WAEA;;WAEA;;WAEA;;WAEA;;WAEA,kBAAkB;;WAElB;;;;WAIA;;;;;;WAMA;;;;;;;WAOA;;WAEA;;WAEA;;WAEA;;;;;WAKA;IACP;IACA;IACA;IACA;;IAEA;IACA;;WAEO;;WAEA,QAAQ;;;;;WAKR;;;;;WAKA;;;;;;UAOM;;WAEN;EACT,QAAQ,QAAQ;;cAOL;cACA;UAEI;;WAEN;WACA;;WAEA;;WAEA;;UAGM;WACN,gBAAgB;WAChB,gBAAgB;WAChB,kBAAkB;;WAElB,QAAQ;WACR,iBAAiB;WACjB,gBAAgB,kBAAkB;;WAElC,kBAAkB;;;;;;WAMlB,OAAO;WACP,WAAW;WACX,gBAAgB;;WAEhB,UAAU;;WAEV,gBAAgB;;WAEhB,iBAAiB;WACjB,oBAAoB;WACpB,kBAAkB;;;;;;;;;WASlB,sBAAsB;;WAEtB,QAAQ;WACR,SAAS;;WAET,mBAAmB;;UAGb;WACN,iBAAiB,SAAS;WAC1B,iBAAiB,SAAS;;;;;;;WAO1B,UAAU;;WAEV,UAAU;;WAEV,WAAW;;WAEX,oBAAoB;;WAEpB,wBAAwB;;WAExB,eAAe;WACf,qBAAqB;;UAGf;WACN,UAAU;WACV,WAAW;;;;;;WAMX,WAAW;WACX,YAAY;WACZ,KAAK;WACL;;UAGM;;WAEN;WACA;;WAEA,MAAM;;WAEN;;WAEA;;WAEA;;WAEA;;WAEA;WACA;;WAEA;WACA;WACA;WACA;WACA;;WAEA;;;;;;;KAQC,mBAAmB;;UAGd;WACN,KAAK;;WAEL;;;;;;;WAOA;;WAEA;;WAEA;;;;;;;;;;;;;UAcM;WACN,gBAAgB;WAChB,aAAa;;;UAIP;WACN,MAAM;WACN,MAAM;WACN,QAAQ;WACR,OAAO;;;UAID;WACN;WACA;WACA;WACA,WAAW,kBAAkB;;UAGvB;;WAEN,kBAAkB,kBAAkB;;WAEpC,OAAO;;;;;;;;WAQP,kBAAkB;;WAElB,SAAS;;WAET,OAAO;;;;;;;;WAQP,UAAU;;;;;;WAMV;WACA,yBAAyB;WACzB,0BAA0B,SAAS;WACnC,WAAW,kBAAkB;;UAGvB;WACN;WACA;WACA;WACA;;;KAIC;;;;;;UAOK;;;;;WAKN;;WAEA;;WAEA;;WAEA;;;;;;;;WAQA;;UAGM;;;;;;WAMN,WAAW;WACX,iBAAiB,SAAS;;;;;;WAM1B,WAAW;;;;;;;WAOX,SAAS,SAAS;WAClB,YAAY;WACZ,WAAW;WACX,eAAe;WACf,YAAY;WACZ,aAAa;WACb,YAAY;WACZ,YAAY;WACZ,UAAU;WACV,OAAO,SAAS;;WAEhB;;UAGM;WACN,eAAe;;WAEf;WACA;WACA;WACA,cAAc;WACd,yBAAyB;WACzB;WACA,eAAe;WACf,UAAU;WACV,WAAW;WACX,SAAS;;WAET;;WAEA;;UAGM;;WAEN,kBAAkB,kBAAkB;WACpC;WACA;WACA,QAAQ;WACR,OAAO;WACP,aAAa;WACb,SAAS;WACT,UAAU;WACV,KAAK;;UAGC;WACN,eAAe;WACf;WACA,aAAa;WACb,iBAAiB;WACjB;WACA,WAAW;WACX,mBAAmB;WACnB,iBAAiB;WACjB,aAAa;WACb,qBAAqB;WACrB,eAAe;;;;;WAKf,UAAU;;;;;;;WAOV;aACE;eAA2B,OAAO;eAA2B;;aAC7D;eAAwB,OAAO;eAA2B;;;WAE5D,eAAe;WACf,kBAAkB;;;;;;;;UASZ;WACN;WACA,gBAAgB;;WAEhB,eAAe;;;KAId;UASK;WACN,MAAM;WACN;WACA;WACA"}

package/dist/wire/index.js

@@ -1,2 +1,2 @@
-import { A as WIRE_VERSION, C as RUBRIC_VERSION_SCHEME, D as TracesIngestRequestSchema, E as TraceEventSchema, O as TracesIngestResponseSchema, S as ListRubricsResponseSchema, T as RubricSchema, _ as FeedbackIngestResponseSchema, a as runRpcBatch, b as JudgeRequestSchema, c as WireError, d as handleListRubrics, f as handleTracesIngest, g as ErrorResponseSchema, h as listBuiltinRubrics, i as dispatchRpc, j as hashRubric, k as VersionResponseSchema, l as handleFeedbackIngest, m as getBuiltinRubric, n as startServer, o as runRpcOnce, p as handleVersion, r as startServerAsync, s as buildOpenApi, t as createApp, u as handleJudge, v as FeedbackTrajectorySchema, w as RubricDimensionSchema, x as JudgeResultSchema, y as HealthResponseSchema } from "../server-D_cjseFN.js";
+import { A as WIRE_VERSION, C as RUBRIC_VERSION_SCHEME, D as TracesIngestRequestSchema, E as TraceEventSchema, O as TracesIngestResponseSchema, S as ListRubricsResponseSchema, T as RubricSchema, _ as FeedbackIngestResponseSchema, a as runRpcBatch, b as JudgeRequestSchema, c as WireError, d as handleListRubrics, f as handleTracesIngest, g as ErrorResponseSchema, h as listBuiltinRubrics, i as dispatchRpc, j as hashRubric, k as VersionResponseSchema, l as handleFeedbackIngest, m as getBuiltinRubric, n as startServer, o as runRpcOnce, p as handleVersion, r as startServerAsync, s as buildOpenApi, t as createApp, u as handleJudge, v as FeedbackTrajectorySchema, w as RubricDimensionSchema, x as JudgeResultSchema, y as HealthResponseSchema } from "../server-BZQR1f0C.js";
 export { ErrorResponseSchema, FeedbackIngestResponseSchema, FeedbackTrajectorySchema, HealthResponseSchema, JudgeRequestSchema, JudgeResultSchema, ListRubricsResponseSchema, RUBRIC_VERSION_SCHEME, RubricDimensionSchema, RubricSchema, TraceEventSchema, TracesIngestRequestSchema, TracesIngestResponseSchema, VersionResponseSchema, WIRE_VERSION, WireError, buildOpenApi, createApp, dispatchRpc, getBuiltinRubric, handleFeedbackIngest, handleJudge, handleListRubrics, handleTracesIngest, handleVersion, hashRubric, listBuiltinRubrics, runRpcBatch, runRpcOnce, startServer, startServerAsync };