daishi-sdk 0.2.21 → 0.2.22

package/dist/generated.d.ts

@@ -19,7 +19,7 @@ export interface ErrorBody {
     /** On 401 and 403: where the developer docs are. */
     docs?: string;
 }
-/** A run record. Null fields fill in as the run launches, plays and ends. `resumes` lists every time a new server release picked the run up mid-match (oldest first, with the round the world was at): the world and the meter carried over, each seat’s own memory of the earlier rounds did not, so a resumed run is a different condition from one played in one release; the match archive carries the same list under `config.run.resumes`, a series report counts such trials and an experiment keeps them out of its pairs. */
+/** A run record. Null fields fill in as the run launches, plays and ends. `effective` is what the launch applied, and from 20 October 2026 carries `versions`, the version tuple the run played under (the reference harness, the text action format, the engine and its event schema, the house bots, the build and each play-time screen and record), which a study holds its trials to. `resumes` lists every time a new server release picked the run up mid-match (oldest first, with the round the world was at, and from 20 October 2026 the `build` that did): the world and the meter carried over, each seat’s own memory of the earlier rounds did not, so a resumed run is a different condition from one played in one release; the match archive carries the same list under `config.run.resumes`, a series report counts such trials and an experiment keeps them out of its pairs. */
 export interface Run {
     run_id: string;
     name: string;
@@ -112,6 +112,7 @@ export interface BatchReport {
     points: BatchReportPoint[];
     precision: BatchReportPrecision;
     rubric: string | null;
+    versions: ReadingTuple;
     cost: Cost;
     verdict: string;
 }
@@ -143,6 +144,7 @@ export interface Experiment {
     verdict: string | null;
     callback_url: string | null;
     replicates: StudyRerun | null;
+    equivalence_margin: number | null;
     note: string | null;
     note_at_ms: number | null;
     notes?: ExperimentNoteEntry[];
@@ -189,6 +191,7 @@ export interface ExperimentReport {
     what_differed: ExperimentReportWhatDiffered[];
     replication: Replication | null;
     reruns: RerunTally | null;
+    versions: ReadingTuple;
     cost: Cost;
     note: string | null;
     note_at_ms: number | null;
@@ -299,11 +302,13 @@ export interface ExperimentBody {
     callback_url?: string;
     /** The design document, as text (markdown or plain): what you predict, which result you will read and how, and what would change your view. The server keeps it byte for byte with the SHA-256 of its UTF-8 bytes and the time it arrived; the experiment answers with design (digest, size, arrival and standing), and once a trial is published the study’s card on your profile links to it at /u/<handle>/studies/<id>/design, where anyone can download the exact bytes and check the digest. Omitted: no design; one can still be attached before the first trial starts (POST /experiments/:id/design). */
     design?: string;
-    /** Re-run a published study: its experiment id (exp_..., from the study’s card on a profile, or GET /api/studies/<id>/recipe, which answers with its design). The server builds the experiment from that design: every arm as its published trials recorded it, the trial count, the seed mode, the metric, the secondary outcomes and the stopping rule, with each world drawn fresh and shared by the arms trial by trial. The body therefore carries none of the study’s settings, only name, max_spend_usd, callback_url and, if you wrote one, your own design document for the re-run; a setting of the study (arms, trials, seed_mode, metric, secondary_metrics, stopping, target_half_width, aa_check) beside replicates answers bad_replicates. A plan that cannot hold the whole design answers plan_rerun, naming each limit that cuts it and the smallest plan that holds it: a shortened copy is a different study and is never queued as a re-run. Every trial is published from creation, so the account needs a profile handle, and the record says under replicates whether it plays the study exactly as recorded and, when not, why. Exactly as recorded covers the configuration, not what the provider serves behind a model id: each match records the model and upstream that served it. */
+    /** Re-run a published study: its experiment id (exp_..., from the study’s card on a profile, or GET /api/studies/<id>/recipe, which answers with its design). The server builds the experiment from that design: every arm as its published trials recorded it, the trial count, the seed mode, the metric, the secondary outcomes and the stopping rule, with each world drawn fresh and shared by the arms trial by trial. The body therefore carries none of the study’s settings, only name, max_spend_usd, equivalence_margin, callback_url and, if you wrote one, your own design document for the re-run; a setting of the study (arms, trials, seed_mode, metric, secondary_metrics, stopping, target_half_width, aa_check) beside replicates answers bad_replicates. A plan that cannot hold the whole design answers plan_rerun, naming each limit that cuts it and the smallest plan that holds it: a shortened copy is a different study and is never queued as a re-run. Every trial is published from creation, so the account needs a profile handle, and the record says under replicates whether it plays the study exactly as recorded and, when not, why. Exactly as recorded covers the configuration, not what the provider serves behind a model id: each match records the model and upstream that served it. */
     replicates?: string;
+    /** With replicates only: the margin the re-run declares, before its first trial, for its reading beside the study it re-runs, on the metric’s analysis scale (index points for the Fitness Index, a fraction for a share, so 0.05 is 5 points, log units for output per turn, so 0.05 is about 5%), positive and at most 100. Each comparison then gets an equivalence check (TOST): whether the 90% Welch interval on the gap between the two studies lies inside plus or minus the margin, the one way a reading says there is no measured difference between them within it, across version tuples or not. Fixed at creation and kept in the witnessed creation record; it cannot be set or changed later, since a margin chosen after the data proves nothing. Without it a re-run is read with Welch’s interval alone. */
+    equivalence_margin?: number;
     /** With replicates only: the spend cap of each arm in USD, clamped to the plan’s range (default the plan’s own). The original’s caps were its owner’s and are not carried. An ordinary experiment sets a cap on each arm instead. */
     max_spend_usd?: number;
-    /** The declared read: how many trials a comparison may lose before it reads incomplete and renders no verdict. A trial is lost when an arm’s trial failed or was cancelled, lost a seat to an error, stopped short of its season at its spend cap, was picked up mid-match by a new server release, or measured nothing on the metric. No lost trial is ever replaced and no trial is added: the design plays the trials it declared and reads what it has. From 0 up to the trials less three, the pairs a verdict needs (bad_loss_tolerance otherwise). Default: a tenth of the trials, rounded down. The experiment and its report carry it as loss_tolerance and print it in rule_text; a comparison past it reads reason incomplete with verdict null, and every comparison carries lost and, with lost pairs, tipping. */
+    /** The declared read: how many trials a comparison may lose before it reads incomplete and renders no verdict. A trial is lost when an arm’s trial failed or was cancelled, lost a seat to an error, stopped short of its season at its spend cap, was picked up mid-match by a new server release, played under another version tuple than the study’s first trial, or measured nothing on the metric. No lost trial is ever replaced and no trial is added: the design plays the trials it declared and reads what it has. From 0 up to the trials less three, the pairs a verdict needs (bad_loss_tolerance otherwise). Default: a tenth of the trials, rounded down. The experiment and its report carry it as loss_tolerance and print it in rule_text; a comparison past it reads reason incomplete with verdict null, and every comparison carries lost and, with lost pairs, tipping. */
     loss_tolerance?: number;
     /** The study’s spend cap in USD over every arm and trial, declared now. Once the trials played so far have spent it, nothing still in line starts: what is left is cancelled under the cap (never as your doing) and the experiment ends with stop_reason spend_cap; a trial already playing runs to its own per-run cap. The trials it stops count as lost against loss_tolerance. At least 0.5 (bad_spend_cap otherwise). Default none: each arm’s own max_spend_usd still bounds each trial. */
     max_total_spend_usd?: number;
@@ -643,6 +648,7 @@ export interface RunExperiment {
 export interface RunResume {
     at_ms: number;
     tick: number;
+    build?: string | null;
 }
 export interface RunLive {
     phase: string;
@@ -740,6 +746,7 @@ export interface BatchReportPoint {
     match_id: string | null;
     frame: 'monitored' | 'neutral' | null;
     resumed_at: number[];
+    other_version: boolean;
     agents: BatchReportPointAgent[] | null;
 }
 export interface BatchReportPointAgent {
@@ -760,6 +767,33 @@ export interface BatchReportPrecision {
     target: number;
     trials_for_target: number | null;
 }
+export interface ReadingTuple {
+    play: PlayVersions | null;
+    rubric: string | null;
+    analysis: string;
+    extractor: string | null;
+    other_trials: number;
+    text: string;
+}
+export interface PlayVersions {
+    harness: string;
+    actions: string;
+    engine: string;
+    events: number;
+    anchors: string;
+    build: string | null;
+    screens: PlayScreens;
+}
+export interface PlayScreens {
+    awareness_scan: string;
+    awareness_probe: string;
+    interactions: string;
+    ledger: string;
+    probes: string;
+    exploration: string;
+    seat_counts: string;
+    payloads: string;
+}
 export interface Cost {
     total_usd: number;
     agent_turns: number;
@@ -1050,6 +1084,8 @@ export interface Replication {
     original: ReplicationOriginal;
     state: 'read' | 'playing' | 'incomplete' | 'cancelled' | 'unpublished' | 'original_open';
     comparisons: StudyDifference[];
+    versions: ReplicationVersions | null;
+    equivalence_margin: number | null;
     text: string;
 }
 export interface ReplicationOriginal {
@@ -1069,6 +1105,7 @@ export interface StudyDifference {
     difference: number | null;
     interval: WelchInterval | null;
     apart: boolean | null;
+    equivalence: EquivalenceCheck | null;
     text: string;
 }
 export interface StudyUnits {
@@ -1082,6 +1119,18 @@ export interface WelchInterval {
     level: number;
     df: number;
 }
+export interface EquivalenceCheck {
+    margin: number;
+    lo: number;
+    hi: number;
+    level: number;
+    p: number;
+    within: boolean;
+}
+export interface ReplicationVersions {
+    same: boolean;
+    differences: string[];
+}
 export interface RerunTally {
     reruns: RerunEntry[];
     counted: number;
@@ -1101,6 +1150,7 @@ export interface RerunEntry {
     exact: boolean;
     state: 'read' | 'playing' | 'incomplete' | 'cancelled' | 'unpublished' | 'original_open';
     apart: boolean | null;
+    versions_differ: boolean | null;
     comparisons: StudyDifference[];
 }
 /** One setup of a run set: how many runs are filed under it and the series batches among them, the harness versions their seats registered under (a set may span versions; nothing is pooled across them), when the earliest run ever filed under it started its match (a run moved out does not raise it), its design, and its counts summed over its runs (null while none carries one; `uncounted` runs whose match could not be read add nothing). */

package/dist/index.d.ts

@@ -1,6 +1,6 @@
 import { Operations, type ErrorCode, type GetBatchResponse, type GetExperimentResponse, type GetRunResponse, type Method, type RequestParts } from './generated.js';
 export * from './generated.js';
-export declare const VERSION = "0.2.21";
+export declare const VERSION = "0.2.22";
 export declare const DEFAULT_BASE_URL = "https://daishi.ai/api/v1";
 export interface DaishiOptions {
     /** A Studio access token (`dsk_...`). Default: the `DAISHI_TOKEN` environment variable. */

package/dist/index.js

@@ -1,6 +1,6 @@
 import { Operations } from './generated.js';
 export * from './generated.js';
-export const VERSION = '0.2.21';
+export const VERSION = '0.2.22';
 export const DEFAULT_BASE_URL = 'https://daishi.ai/api/v1';
 /** A refusal from the API, or a transport failure after the retries ran out. */
 export class DaishiError extends Error {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "daishi-sdk",
-  "version": "0.2.21",
+  "version": "0.2.22",
   "description": "Typed client for the Daishi Studio API: scenarios, runs, series batches and paired experiments, with retries and waiting helpers.",
   "keywords": [
     "daishi",