daishi-sdk 0.2.18 → 0.2.20

package/dist/generated.d.ts

@@ -5,7 +5,7 @@ export type JsonObject = {
 /** The API version these types describe. */
 export declare const API_VERSION = "1.0.0";
 /** Every error code the API documents. */
-export type ErrorCode = 'token_required' | 'scope_required' | 'plan_api' | 'rate_limited' | 'handover' | 'bad_request' | 'bad_scenario' | 'bad_replicates' | 'profile_required' | 'bad_season_ticks' | 'bad_frame' | 'bad_env' | 'too_many_scenarios' | 'unknown_scenario' | 'bad_skill' | 'too_many_skills' | 'unknown_skill' | 'bad_roster' | 'bad_model' | 'bad_name' | 'bad_provider' | 'provider_key_required' | 'openrouter_key_required' | 'openrouter_balance_low' | 'billing_frozen' | 'email_unverified' | 'bad_trials' | 'bad_spend_cap' | 'plan_seats' | 'plan_ticks' | 'plan_series' | 'plan_quota' | 'too_many_runs' | 'run_history_full' | 'seed_not_allowed' | 'unknown_run' | 'unknown_batch' | 'unknown_experiment' | 'experiment_trial' | 'experiment_not_running' | 'experiment_running' | 'bad_cancel_reason' | 'cancel_reason_fixed' | 'bad_replaced_by' | 'experiment_arms' | 'bad_experiment' | 'experiment_scenario' | 'bad_game' | 'bad_arena' | 'experiment_world_only' | 'experiment_same_arms' | 'experiment_env_seed' | 'experiment_stopping' | 'bad_metric' | 'bad_design' | 'design_final' | 'design_launching' | 'design_late' | 'no_design' | 'no_creation_record' | 'experiment_effort_roster' | 'experiment_no_probe' | 'experiment_too_few_seats' | 'plan_experiment_arms' | 'plan_rerun' | 'unknown_log' | 'plan_invites' | 'invite_world' | 'invite_series' | 'too_many_open_worlds' | 'no_invite' | 'invite_fixed' | 'bad_invite' | 'never_played' | 'run_playing' | 'study_playing' | 'bad_note' | 'note_too_long' | 'unknown_set' | 'unknown_setup' | 'bad_set' | 'set_full' | 'series_trial' | 'in_another_set' | 'guest_seats' | 'set_empty' | 'run_launching' | 'internal_error';
+export type ErrorCode = 'token_required' | 'scope_required' | 'plan_api' | 'rate_limited' | 'handover' | 'bad_request' | 'bad_scenario' | 'bad_replicates' | 'profile_required' | 'bad_season_ticks' | 'bad_frame' | 'bad_env' | 'too_many_scenarios' | 'unknown_scenario' | 'bad_skill' | 'too_many_skills' | 'unknown_skill' | 'bad_roster' | 'bad_model' | 'bad_name' | 'bad_provider' | 'provider_key_required' | 'openrouter_key_required' | 'openrouter_balance_low' | 'billing_frozen' | 'email_unverified' | 'bad_trials' | 'bad_spend_cap' | 'bad_loss_tolerance' | 'plan_seats' | 'plan_ticks' | 'plan_series' | 'plan_quota' | 'too_many_runs' | 'run_history_full' | 'seed_not_allowed' | 'unknown_run' | 'unknown_batch' | 'unknown_experiment' | 'experiment_trial' | 'experiment_not_running' | 'experiment_running' | 'bad_cancel_reason' | 'cancel_reason_fixed' | 'bad_replaced_by' | 'experiment_arms' | 'bad_experiment' | 'experiment_scenario' | 'bad_game' | 'bad_arena' | 'experiment_world_only' | 'experiment_same_arms' | 'experiment_env_seed' | 'experiment_stopping' | 'bad_metric' | 'bad_design' | 'design_final' | 'design_launching' | 'design_late' | 'no_design' | 'no_creation_record' | 'experiment_effort_roster' | 'experiment_no_probe' | 'experiment_too_few_seats' | 'plan_experiment_arms' | 'plan_rerun' | 'unknown_log' | 'plan_invites' | 'invite_world' | 'invite_series' | 'too_many_open_worlds' | 'no_invite' | 'invite_fixed' | 'bad_invite' | 'never_played' | 'run_playing' | 'study_playing' | 'bad_note' | 'note_too_long' | 'unknown_set' | 'unknown_setup' | 'bad_set' | 'set_full' | 'series_trial' | 'in_another_set' | 'guest_seats' | 'set_empty' | 'run_launching' | 'internal_error';
 export type Method = 'GET' | 'POST' | 'PUT' | 'DELETE';
 export interface RequestParts {
     query?: Record<string, string | number | boolean | undefined | null>;
@@ -130,6 +130,8 @@ export interface Experiment {
     pilot: boolean;
     name_commit_min_hex: number;
     creation_sha256: string | null;
+    loss_tolerance: number | null;
+    max_total_spend_usd: number | null;
     seed_source: ExperimentSeedSource;
     metric: 'fitness_index' | 'rank' | 'survived' | 'index_survival' | 'index_economy' | 'index_social' | 'index_adaptation' | 'index_competitiveness' | 'aggression_per_100_rounds' | 'trades_per_100_rounds' | 'messages_per_100_rounds' | 'map_coverage' | 'witness_sensitivity' | 'compliance_rate' | 'compliance_latency' | 'injection_resistance' | 'output_tokens_per_turn' | 'said_evaluated' | 'aggression_between_seats_per_100_rounds' | 'trades_between_seats_per_100_rounds' | 'betrayals_per_100_rounds' | 'ring_share';
     secondary_metrics: Array<'fitness_index' | 'rank' | 'survived' | 'index_survival' | 'index_economy' | 'index_social' | 'index_adaptation' | 'index_competitiveness' | 'aggression_per_100_rounds' | 'trades_per_100_rounds' | 'messages_per_100_rounds' | 'map_coverage' | 'witness_sensitivity' | 'compliance_rate' | 'compliance_latency' | 'injection_resistance' | 'output_tokens_per_turn' | 'said_evaluated' | 'aggression_between_seats_per_100_rounds' | 'trades_between_seats_per_100_rounds' | 'betrayals_per_100_rounds' | 'ring_share'>;
@@ -170,13 +172,19 @@ export interface ExperimentReport {
     analysis_scale: 'metric' | 'negative_rank' | 'log';
     rule: StoppingRule;
     rule_text: string;
+    loss_tolerance: number | null;
+    max_total_spend_usd: number | null;
+    focal: ExperimentReportFocal;
     arms: ExperimentReportArm[];
     pairs: ExperimentReportPair[];
     comparisons: ExperimentReportComparison[];
     secondary: ExperimentReportSecondary[];
     detectable_gap: ExperimentReportDetectableGap;
     decision: ExperimentReportDecision;
+    completeness: ExperimentReportCompleteness;
+    interference: ExperimentReportInterference;
     verdict: string;
+    limitations: ExperimentReportLimitation[];
     plain: string;
     what_differed: ExperimentReportWhatDiffered[];
     cost: Cost;
@@ -293,6 +301,10 @@ export interface ExperimentBody {
     replicates?: string;
     /** 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. */
+    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;
 }
 export interface ExperimentArm {
     /** How the report names this arm. Default arm 1, arm 2, ...; labels must be distinct. */
@@ -810,6 +822,11 @@ export interface ExperimentDesign {
     standing: 'declared' | 'named' | 'after';
     name_prefix: string | null;
 }
+export interface ExperimentReportFocal {
+    seats: number[] | null;
+    background: number[];
+    text: string;
+}
 export interface ExperimentReportArm {
     index: number;
     label: string;
@@ -819,10 +836,12 @@ export interface ExperimentReportArm {
     differs_from_control: string[];
     trials: ExperimentReportArmTrials;
     score: ExperimentReportArmScore | null;
+    background: ExperimentReportArmBackground | null;
     cost_usd: number;
     agent_turns: number;
     usage: JsonObject;
     awareness: JsonObject;
+    health: ExperimentReportArmHealth;
     effort?: ExperimentReportArmEffort;
     contact?: ExperimentReportArmContact;
 }
@@ -840,6 +859,34 @@ export interface ExperimentReportArmScore {
     sd: number | null;
     ci: Interval | null;
 }
+export interface ExperimentReportArmBackground {
+    seats: number;
+    mean: number;
+    iqm: number;
+    median: number;
+    sd: number | null;
+    ci: Interval | null;
+}
+export interface ExperimentReportArmHealth {
+    measured_seats: number;
+    calls: number;
+    error_turns: number;
+    provider_retries: ExperimentReportArmHealthProviderRetries | null;
+    world_rejections: ExperimentReportArmHealthWorldRejections | null;
+}
+export interface ExperimentReportArmHealthProviderRetries {
+    seats: number;
+    calls: number;
+    retries: number;
+    wait_ms: number;
+    causes: JsonObject;
+}
+export interface ExperimentReportArmHealthWorldRejections {
+    seats: number;
+    turn_budget: number;
+    pace: number;
+    invalid: number;
+}
 export interface ExperimentReportArmEffort {
     measured_seats: number;
     short_seats: number;
@@ -887,6 +934,8 @@ export interface ExperimentReportComparison {
     detail: string;
     exact: ExactCheck | null;
     by_world: WorldReading | null;
+    lost: number;
+    tipping: ExperimentReportComparisonTipping | null;
 }
 export interface ExactCheck {
     unit: 'pair' | 'world';
@@ -899,6 +948,19 @@ export interface WorldReading {
     repeated: number;
     interval: Interval | null;
 }
+export interface ExperimentReportComparisonTipping {
+    lost: number;
+    pairs: number;
+    against: 'arm_better' | 'control_better' | 'undecided';
+    to_undo: number | null;
+    to_decide: ExperimentReportComparisonTippingToDecide | null;
+    widest_seen: number;
+    text: string;
+}
+export interface ExperimentReportComparisonTippingToDecide {
+    arm: number | null;
+    control: number | null;
+}
 export interface ExperimentReportSecondary {
     metric: 'fitness_index' | 'rank' | 'survived' | 'index_survival' | 'index_economy' | 'index_social' | 'index_adaptation' | 'index_competitiveness' | 'aggression_per_100_rounds' | 'trades_per_100_rounds' | 'messages_per_100_rounds' | 'map_coverage' | 'witness_sensitivity' | 'compliance_rate' | 'compliance_latency' | 'injection_resistance' | 'output_tokens_per_turn' | 'said_evaluated' | 'aggression_between_seats_per_100_rounds' | 'trades_between_seats_per_100_rounds' | 'betrayals_per_100_rounds' | 'ring_share';
     analysis_scale: 'metric' | 'negative_rank' | 'log';
@@ -945,6 +1007,36 @@ export interface ExperimentReportDecision {
     reason: string;
     summary: string;
 }
+export interface ExperimentReportCompleteness {
+    state: 'complete' | 'incomplete' | 'pending';
+    why: 'losses' | 'spend_cap' | 'cancelled' | null;
+    trials: number;
+    loss_tolerance: number | null;
+    lost: number;
+    spend: ExperimentReportCompletenessSpend;
+    text: string;
+}
+export interface ExperimentReportCompletenessSpend {
+    cap_usd: number | null;
+    spent_usd: number;
+    reached: boolean;
+}
+export interface ExperimentReportInterference {
+    per_arm: ExperimentReportInterferencePerArm[];
+    reading: 'alike' | 'uneven' | 'unread';
+    text: string;
+}
+export interface ExperimentReportInterferencePerArm {
+    arm: number;
+    label: string;
+    calls: number;
+    retries: number;
+    per_100_calls: number | null;
+}
+export interface ExperimentReportLimitation {
+    kind: 'construct' | 'access' | 'sample' | 'selection' | 'version' | 'measurement' | 'scope';
+    text: string;
+}
 export interface ExperimentReportWhatDiffered {
     arm: number;
     label: string;

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 = "0.2.18";
+export declare const VERSION = "0.2.20";
 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 = '0.2.18';
+export const VERSION = '0.2.20';
 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": "0.2.18",
+  "version": "0.2.20",
   "description": "Typed client for the Daishi Studio API: scenarios, runs, series batches and paired experiments, with retries and waiting helpers.",
   "keywords": [
     "daishi",