@absolutejs/voice 0.0.22-beta.67 → 0.0.22-beta.670

package/dist/testing/benchmark.d.ts

@@ -1,6 +1,9 @@
-import type { STTAdapter, STTAdapterOpenOptions } from '../types';
-import { type VoiceSTTAdapterHarnessOptions, type VoiceSTTAdapterHarnessResult } from './stt';
-import type { VoiceTestFixture } from './fixtures';
+import type { STTAdapter, STTAdapterOpenOptions } from "../core/types";
+import { type VoiceSTTAdapterHarnessOptions, type VoiceSTTAdapterHarnessResult } from "./stt";
+import type { VoiceTestFixture } from "./fixtures";
+import { type VoiceConfidenceCalibrationReport } from "./confidenceCalibration";
+import { type VoiceCriticalFieldAccuracy } from "./criticalFields";
+import { type VoiceSTTConformanceReport } from "./conformance";
 export type VoiceExpectedTermAccuracy = {
     allMatched: boolean;
     expectedTerms: string[];
@@ -8,7 +11,7 @@ export type VoiceExpectedTermAccuracy = {
     missingTerms: string[];
     recall: number;
 };
-export type VoiceSTTFixtureEnvironment = 'accent' | 'accent-noisy' | 'clean' | 'code-switch' | 'jargon' | 'multilingual' | 'multi-speaker' | 'noisy' | 'telephony' | 'other';
+export type VoiceSTTFixtureEnvironment = "accent" | "accent-noisy" | "clean" | "code-switch" | "jargon" | "multilingual" | "multi-speaker" | "noisy" | "telephony" | "other";
 export type VoiceSpeakerTurnAccuracy = {
     available: boolean;
     actualTurnCount: number;
@@ -18,13 +21,15 @@ export type VoiceSpeakerTurnAccuracy = {
     postClustered?: boolean;
 };
 export type VoiceSTTBenchmarkFixtureResult = {
-    accuracy: VoiceSTTAdapterHarnessResult['accuracy'];
+    accuracy: VoiceSTTAdapterHarnessResult["accuracy"];
     closeCount: number;
-    difficulty?: VoiceTestFixture['difficulty'];
+    conformance?: VoiceSTTConformanceReport;
+    difficulty?: VoiceTestFixture["difficulty"];
     elapsedMs: number;
     endOfTurnCount: number;
     errorCount: number;
     expectedTerms: VoiceExpectedTermAccuracy;
+    criticalFields?: VoiceCriticalFieldAccuracy;
     finalCount: number;
     finalText: string;
     fixtureId: string;
@@ -39,6 +44,7 @@ export type VoiceSTTBenchmarkFixtureResult = {
     timeToEndOfTurnMs?: number;
     timeToFirstFinalMs?: number;
     timeToFirstPartialMs?: number;
+    transcriptConfidence?: number;
     title: string;
 };
 export type VoiceSTTBenchmarkSummary = {
@@ -49,6 +55,8 @@ export type VoiceSTTBenchmarkSummary = {
     averageFinalCount: number;
     averageSpeakerTurnMatchRate?: number;
     averageTermRecall: number;
+    averageCriticalFieldAccuracy: number;
+    requiredCriticalFieldPassRate: number;
     averagePostSpeechTimeToEndOfTurnMs?: number;
     averagePostSpeechTimeToFirstFinalMs?: number;
     averageTimeToEndOfTurnMs?: number;
@@ -63,6 +71,7 @@ export type VoiceSTTBenchmarkSummary = {
     totalErrorCount: number;
     wordAccuracyRate: number;
     groupSummaries: VoiceSTTBenchmarkFixtureSummary[];
+    confidenceCalibration: VoiceConfidenceCalibrationReport;
 };
 export type VoiceSTTBenchmarkFixtureSummary = {
     group: VoiceSTTFixtureEnvironment;
@@ -141,27 +150,27 @@ export type VoiceSTTBenchmarkAcceptanceResult = {
     score: number;
 };
 export type VoiceSTTBenchmarkOptions = VoiceSTTAdapterHarnessOptions & {
-    fixtureOptions?: Record<string, Omit<VoiceSTTAdapterHarnessOptions, 'fixtureOptions' | 'openOptions'>>;
+    fixtureOptions?: Record<string, Omit<VoiceSTTAdapterHarnessOptions, "fixtureOptions" | "openOptions">>;
     openOptions?: Partial<STTAdapterOpenOptions> | ((fixture: VoiceTestFixture) => Partial<STTAdapterOpenOptions> | undefined);
 };
 export declare const resolveFixtureEnvironment: (fixture: Pick<VoiceTestFixture, "language" | "tags">) => VoiceSTTFixtureEnvironment;
-export declare const summarizeSTTBenchmark: (adapterId: string, fixtures: VoiceSTTBenchmarkFixtureResult[]) => VoiceSTTBenchmarkSummary;
-export declare const evaluateSTTBenchmarkAcceptance: (report: VoiceSTTBenchmarkReport, thresholds?: VoiceSTTBenchmarkAcceptanceThresholds) => VoiceSTTBenchmarkAcceptanceResult;
 export declare const compareSTTBenchmarks: (reports: VoiceSTTBenchmarkReport[]) => VoiceSTTBenchmarkComparison;
-export declare const runSTTAdapterBenchmark: ({ adapter, adapterId, fixtures, options }: {
+export declare const evaluateSTTBenchmarkAcceptance: (report: VoiceSTTBenchmarkReport, thresholds?: VoiceSTTBenchmarkAcceptanceThresholds) => VoiceSTTBenchmarkAcceptanceResult;
+export declare const runSTTAdapterBenchmark: ({ adapter, adapterId, fixtures, options, }: {
     adapter: STTAdapter;
     adapterId: string;
     fixtures: VoiceTestFixture[];
     options?: VoiceSTTBenchmarkOptions;
 }) => Promise<VoiceSTTBenchmarkReport>;
-export declare const summarizeSTTBenchmarkSeries: (input: {
-    adapterId: string;
-    reports: VoiceSTTBenchmarkReport[];
-}) => VoiceSTTBenchmarkSeriesReport;
-export declare const runSTTAdapterBenchmarkSeries: ({ adapter, adapterId, fixtures, options, runs }: {
+export declare const runSTTAdapterBenchmarkSeries: ({ adapter, adapterId, fixtures, options, runs, }: {
     adapter: STTAdapter;
     adapterId: string;
     fixtures: VoiceTestFixture[];
     options?: VoiceSTTBenchmarkOptions;
     runs: number;
 }) => Promise<VoiceSTTBenchmarkSeriesReport>;
+export declare const summarizeSTTBenchmark: (adapterId: string, fixtures: VoiceSTTBenchmarkFixtureResult[]) => VoiceSTTBenchmarkSummary;
+export declare const summarizeSTTBenchmarkSeries: (input: {
+    adapterId: string;
+    reports: VoiceSTTBenchmarkReport[];
+}) => VoiceSTTBenchmarkSeriesReport;

package/dist/testing/confidenceCalibration.d.ts

@@ -0,0 +1,19 @@
+export type VoiceConfidenceCalibrationSample = {
+    confidence: number;
+    correct: boolean;
+    metadata?: Record<string, unknown>;
+};
+export type VoiceConfidenceCalibrationBin = {
+    accuracy: number;
+    averageConfidence: number;
+    count: number;
+    lowerBound: number;
+    upperBound: number;
+};
+export type VoiceConfidenceCalibrationReport = {
+    bins: VoiceConfidenceCalibrationBin[];
+    brierScore: number;
+    expectedCalibrationError: number;
+    sampleCount: number;
+};
+export declare const calibrateVoiceConfidence: (samples: VoiceConfidenceCalibrationSample[], binCount?: number) => VoiceConfidenceCalibrationReport;

package/dist/testing/conformance.d.ts

@@ -0,0 +1,11 @@
+import type { VoiceSTTAdapterHarnessResult } from "./stt";
+export type VoiceSTTConformanceCheck = {
+    detail: string;
+    id: string;
+    passed: boolean;
+};
+export type VoiceSTTConformanceReport = {
+    checks: VoiceSTTConformanceCheck[];
+    passed: boolean;
+};
+export declare const evaluateSTTAdapterConformance: (result: VoiceSTTAdapterHarnessResult) => VoiceSTTConformanceReport;

package/dist/testing/corrected.d.ts

@@ -1,8 +1,8 @@
-import type { VoiceLexiconEntry, VoicePhraseHint, VoiceTurnCorrectionHandler } from '../types';
-import { type VoiceExpectedTermAccuracy, type VoiceSTTBenchmarkReport } from './benchmark';
-import type { VoiceSessionBenchmarkReport } from './sessionBenchmark';
-import type { VoiceTestFixture } from './fixtures';
-export type VoiceCorrectionHintProfile = 'generic' | 'benchmark-seeded';
+import type { VoiceLexiconEntry, VoicePhraseHint, VoiceTurnCorrectionHandler } from "../core/types";
+import { type VoiceExpectedTermAccuracy, type VoiceSTTBenchmarkReport } from "./benchmark";
+import type { VoiceSessionBenchmarkReport } from "./sessionBenchmark";
+import type { VoiceTestFixture } from "./fixtures";
+export type VoiceCorrectionHintProfile = "generic" | "benchmark-seeded";
 type VoiceCorrectionAuditSlice<TReport> = {
     fixtureIds: string[];
     raw: TReport;
@@ -28,14 +28,14 @@ export type VoiceSessionCorrectionAudit = {
 };
 export declare const isCorrectionHoldoutFixtureId: (fixtureId: string) => boolean;
 export declare const buildFixturePhraseHints: (fixture: Pick<VoiceTestFixture, "expectedTerms" | "expectedText">, profile?: VoiceCorrectionHintProfile) => VoicePhraseHint[];
+export declare const applyCorrectedBenchmarkReport: (report: VoiceSTTBenchmarkReport, fixtures: VoiceTestFixture[], profile?: VoiceCorrectionHintProfile) => VoiceSTTBenchmarkReport;
+export declare const applyExperimentalBenchmarkReport: (report: VoiceSTTBenchmarkReport, fixtures: VoiceTestFixture[]) => VoiceSTTBenchmarkReport;
+export declare const applyLexiconCorrectedBenchmarkReport: (report: VoiceSTTBenchmarkReport, fixtures: VoiceTestFixture[], buildLexicon: (fixture: Pick<VoiceTestFixture, "expectedTerms" | "expectedText" | "language" | "tags">) => VoiceLexiconEntry[]) => VoiceSTTBenchmarkReport;
 export declare const buildCodeSwitchBenchmarkLexicon: (fixture: Pick<VoiceTestFixture, "expectedTerms" | "expectedText" | "language" | "tags">) => VoiceLexiconEntry[];
 export declare const buildCodeSwitchBenchmarkPhraseHints: (fixture: Pick<VoiceTestFixture, "expectedTerms" | "expectedText" | "language" | "tags">) => VoicePhraseHint[];
-export declare const createCodeSwitchBenchmarkCorrectionHandler: () => VoiceTurnCorrectionHandler;
 export declare const createBenchmarkCorrectionHandler: (profile: VoiceCorrectionHintProfile) => VoiceTurnCorrectionHandler;
+export declare const createCodeSwitchBenchmarkCorrectionHandler: () => VoiceTurnCorrectionHandler;
 export declare const scoreCorrectedExpectedTerms: (actualText: string, expectedTerms: string[] | undefined) => VoiceExpectedTermAccuracy;
-export declare const applyCorrectedBenchmarkReport: (report: VoiceSTTBenchmarkReport, fixtures: VoiceTestFixture[], profile?: VoiceCorrectionHintProfile) => VoiceSTTBenchmarkReport;
-export declare const applyExperimentalBenchmarkReport: (report: VoiceSTTBenchmarkReport, fixtures: VoiceTestFixture[]) => VoiceSTTBenchmarkReport;
-export declare const applyLexiconCorrectedBenchmarkReport: (report: VoiceSTTBenchmarkReport, fixtures: VoiceTestFixture[], buildLexicon: (fixture: Pick<VoiceTestFixture, "expectedTerms" | "expectedText" | "language" | "tags">) => VoiceLexiconEntry[]) => VoiceSTTBenchmarkReport;
 export declare const buildCorrectionBenchmarkAudit: (rawReport: VoiceSTTBenchmarkReport, fixtures: VoiceTestFixture[]) => VoiceCorrectionBenchmarkAudit;
 export declare const buildSessionCorrectionAudit: (raw: VoiceSessionBenchmarkReport, generic: VoiceSessionBenchmarkReport, experimental: VoiceSessionBenchmarkReport, benchmarkSeeded: VoiceSessionBenchmarkReport, scenarios: Array<Pick<VoiceTestFixture, "expectedTerms" | "expectedText" | "expectedTurnTexts" | "id">>) => VoiceSessionCorrectionAudit;
 export {};

package/dist/testing/criticalFields.d.ts

@@ -0,0 +1,22 @@
+export type VoiceCriticalFieldKind = "acronym" | "brand" | "currency" | "custom" | "email" | "number" | "organization" | "percentage" | "person-name" | "phone";
+export type VoiceExpectedCriticalField = {
+    aliases?: string[];
+    id: string;
+    kind: VoiceCriticalFieldKind;
+    metadata?: Record<string, unknown>;
+    required?: boolean;
+    value: string;
+};
+export type VoiceCriticalFieldResult = VoiceExpectedCriticalField & {
+    matched: boolean;
+    matchedAlias?: string;
+};
+export type VoiceCriticalFieldAccuracy = {
+    accuracy: number;
+    fields: VoiceCriticalFieldResult[];
+    matchedCount: number;
+    missingFieldIds: string[];
+    passesRequired: boolean;
+    totalCount: number;
+};
+export declare const scoreVoiceCriticalFields: (actualText: string, expectedFields?: VoiceExpectedCriticalField[]) => VoiceCriticalFieldAccuracy;

package/dist/testing/duplex.d.ts

@@ -1,11 +1,11 @@
-import type { VoiceBargeInOptions } from '../types';
+import type { VoiceBargeInOptions } from "../core/types";
 export type VoiceDuplexBenchmarkScenario = {
     expectedInterruptCount: number;
     id: string;
     initialPartial?: string;
     interruptDelayMs?: number;
     level?: number;
-    mode: 'audio-send' | 'input-level' | 'partial';
+    mode: "audio-send" | "input-level" | "partial";
     partial?: string;
     title: string;
 };
@@ -15,7 +15,7 @@ export type VoiceDuplexBenchmarkScenarioResult = {
     expectedInterruptCount: number;
     fixtureId: string;
     interruptLatencyMs?: number;
-    mode: VoiceDuplexBenchmarkScenario['mode'];
+    mode: VoiceDuplexBenchmarkScenario["mode"];
     passes: boolean;
     title: string;
 };
@@ -45,8 +45,6 @@ export declare const getDefaultVoiceDuplexBenchmarkScenarios: () => {
     partial?: string;
     title: string;
 }[];
-export declare const runVoiceDuplexBenchmarkScenario: (scenario: VoiceDuplexBenchmarkScenario, options?: VoiceDuplexBenchmarkOptions) => Promise<VoiceDuplexBenchmarkScenarioResult>;
-export declare const summarizeVoiceDuplexBenchmark: (fixtures: VoiceDuplexBenchmarkScenarioResult[]) => VoiceDuplexBenchmarkSummary;
 export declare const runVoiceDuplexBenchmark: (scenarios?: {
     expectedInterruptCount: number;
     id: string;
@@ -57,3 +55,5 @@ export declare const runVoiceDuplexBenchmark: (scenarios?: {
     partial?: string;
     title: string;
 }[], options?: VoiceDuplexBenchmarkOptions) => Promise<VoiceDuplexBenchmarkReport>;
+export declare const runVoiceDuplexBenchmarkScenario: (scenario: VoiceDuplexBenchmarkScenario, options?: VoiceDuplexBenchmarkOptions) => Promise<VoiceDuplexBenchmarkScenarioResult>;
+export declare const summarizeVoiceDuplexBenchmark: (fixtures: VoiceDuplexBenchmarkScenarioResult[]) => VoiceDuplexBenchmarkSummary;

package/dist/testing/fixtures.d.ts

@@ -1,20 +1,23 @@
-import type { AudioFormat, VoiceExpectedSpeakerTurn } from '../types';
+import type { AudioFormat, VoiceExpectedSpeakerTurn } from "../core/types";
+import type { VoiceExpectedCriticalField } from "./criticalFields";
 export type VoiceTestFixtureManifestEntry = {
     id: string;
     title: string;
     audioPath: string;
     expectedText: string;
     expectedTerms?: string[];
+    expectedCriticalFields?: VoiceExpectedCriticalField[];
     expectedSpeakerTurns?: VoiceExpectedSpeakerTurn[];
     expectedTurnTexts?: string[];
     chunkDurationMs?: number;
     language?: string;
-    difficulty?: 'clean' | 'noisy' | 'challenging';
+    difficulty?: "clean" | "noisy" | "challenging";
     tags?: string[];
     tailPaddingMs?: number;
     format?: Partial<AudioFormat>;
+    provenance?: Omit<import("./provenance").VoiceCorpusFixtureProvenance, "audioSha256" | "fixtureId">;
 };
-export type VoiceTestFixture = Omit<VoiceTestFixtureManifestEntry, 'audioPath'> & {
+export type VoiceTestFixture = Omit<VoiceTestFixtureManifestEntry, "audioPath"> & {
     audio: Uint8Array;
     audioPath: string;
     format: AudioFormat;
@@ -33,6 +36,6 @@ export type VoiceFixtureLoadOptions = {
 export declare const getVoiceFixtureDirectory: () => Promise<string>;
 export declare const resolveVoiceFixtureDirectories: (input?: string | string[] | VoiceFixtureLoadOptions) => Promise<string[]>;
 export declare const createTelephonyVoiceTestFixtures: (fixtures: VoiceTestFixture[], options?: VoiceTelephonyFixtureOptions) => VoiceTestFixture[];
-export declare const createMultiSpeakerVoiceTestFixtures: (fixtures: VoiceTestFixture[], options?: VoiceMultiSpeakerFixtureOptions) => VoiceTestFixture[];
 export declare const createJargonVoiceTestFixtures: (fixtures: VoiceTestFixture[]) => VoiceTestFixture[];
+export declare const createMultiSpeakerVoiceTestFixtures: (fixtures: VoiceTestFixture[], options?: VoiceMultiSpeakerFixtureOptions) => VoiceTestFixture[];
 export declare const loadVoiceTestFixtures: (fixtureDirectory?: string | string[] | VoiceFixtureLoadOptions) => Promise<VoiceTestFixture[]>;

package/dist/testing/index.d.ts

@@ -1,13 +1,21 @@
-export * from './accuracy';
-export * from './benchmark';
-export * from './corrected';
-export * from './duplex';
-export * from './fixtures';
-export * from './ioProviderSimulator';
-export * from './providerSimulator';
-export * from './resilience';
-export * from './review';
-export * from './sessionBenchmark';
-export * from './stt';
-export * from './telephony';
-export * from './tts';
+export * from "./accuracy";
+export * from "./audioMatrix";
+export * from "./benchmark";
+export * from "./confidenceCalibration";
+export * from "./conformance";
+export * from "./criticalFields";
+export * from "./corrected";
+export * from "./duplex";
+export * from "./fixtures";
+export * from "./ioProviderSimulator";
+export * from "./outcomes";
+export * from "./providerSimulator";
+export * from "./provenance";
+export * from "./resilience";
+export * from "./review";
+export * from "./routingBenchmark";
+export * from "./sessionBenchmark";
+export * from "./stt";
+export * from "./statistics";
+export * from "./telephony";
+export * from "./tts";