@thinkhive/sdk 3.1.1 → 4.0.0
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/MIGRATION.md +83 -12
- package/README.md +279 -128
- package/dist/api/agents.d.ts +169 -0
- package/dist/api/agents.js +185 -0
- package/dist/api/apiKeys.d.ts +252 -0
- package/dist/api/apiKeys.js +298 -0
- package/dist/api/business-metrics.d.ts +188 -0
- package/dist/api/business-metrics.js +213 -0
- package/dist/api/calibration.d.ts +0 -62
- package/dist/api/calibration.js +5 -48
- package/dist/api/claims.js +10 -7
- package/dist/api/conversation-eval.d.ts +200 -0
- package/dist/api/conversation-eval.js +235 -0
- package/dist/api/deterministic-graders.d.ts +205 -0
- package/dist/api/deterministic-graders.js +191 -0
- package/dist/api/eval-health.d.ts +250 -0
- package/dist/api/eval-health.js +224 -0
- package/dist/api/human-review.d.ts +275 -0
- package/dist/api/human-review.js +236 -0
- package/dist/api/nondeterminism.d.ts +300 -0
- package/dist/api/nondeterminism.js +250 -0
- package/dist/api/quality-metrics.d.ts +303 -0
- package/dist/api/quality-metrics.js +198 -0
- package/dist/api/roi-analytics.d.ts +263 -0
- package/dist/api/roi-analytics.js +204 -0
- package/dist/api/runs.js +12 -6
- package/dist/api/transcript-patterns.d.ts +204 -0
- package/dist/api/transcript-patterns.js +227 -0
- package/dist/core/client.d.ts +83 -9
- package/dist/core/client.js +229 -34
- package/dist/core/config.d.ts +2 -3
- package/dist/core/config.js +3 -4
- package/dist/core/types.d.ts +57 -4
- package/dist/core/types.js +1 -1
- package/dist/index.d.ts +429 -76
- package/dist/index.js +262 -42
- package/package.json +2 -2
package/dist/index.d.ts
CHANGED
|
@@ -11,13 +11,35 @@
|
|
|
11
11
|
* @license MIT
|
|
12
12
|
*/
|
|
13
13
|
import { isInitialized, SDK_VERSION, DEFAULT_ENDPOINT } from './core/config';
|
|
14
|
-
import { ThinkHiveApiError, ThinkHiveValidationError } from './core/client';
|
|
14
|
+
import { ThinkHiveError, ThinkHiveApiError, ThinkHiveValidationError, PermissionDeniedError, AgentScopeError, RateLimitError, IpWhitelistError } from './core/client';
|
|
15
15
|
import { runs, createRunWithContext, toOpenAIMessages, fromOpenAIMessages } from './api/runs';
|
|
16
16
|
import { claims, isFact, isInference, isComputed, getHighConfidenceClaims, groupClaimsByType, groupClaimsByCategory } from './api/claims';
|
|
17
17
|
import { calibration, calculateBrierScore, calculateECE, isWellCalibrated, getCalibrationQuality } from './api/calibration';
|
|
18
|
+
import { humanReview } from './api/human-review';
|
|
19
|
+
import { nondeterminism, calculatePassAtK, calculatePassToK, requiredPassRateForPassAtK, isReliableEvaluation, getReliabilityRecommendation } from './api/nondeterminism';
|
|
20
|
+
import { evalHealth, hasHealthIssue, getSeverityLevel as getHealthSeverityLevel, isSaturated, getSaturationRecommendation } from './api/eval-health';
|
|
21
|
+
import { deterministicGraders, createRegexRule, createContainsRule, createLengthRule, createJsonSchemaRule, allRulesPassed, getFailedRules, calculateAverageScore } from './api/deterministic-graders';
|
|
22
|
+
import { conversationEval, aggregateWorst, aggregateAverage, aggregateWeighted, aggregateFinalTurn, aggregateMajority, getAggregator, getProblematicTurns, analyzeConversationTrend } from './api/conversation-eval';
|
|
23
|
+
import { transcriptPatterns, isHighRisk, getMatchesByCategory, getCriticalInsights, hasPiiExposure, hasFrustrationSignals, hasEscalationRequest, getCategoryDistribution, getRecommendations, needsAttention, sortMatchesBySeverity } from './api/transcript-patterns';
|
|
24
|
+
import { agents } from './api/agents';
|
|
25
|
+
import { apiKeys, hasPermission, isExpired as isApiKeyExpired, isValid as isApiKeyValid, getTimeUntilExpiry, canAccessAgent } from './api/apiKeys';
|
|
26
|
+
import { roiAnalytics, calculateRevenueAtRisk, calculateAutomationSavings, formatCurrency, getROIQuality } from './api/roi-analytics';
|
|
27
|
+
import { businessMetrics, isMetricReady, needsMoreTraces, awaitingExternalData, isMetricStale, getStatusMessage, getTraceProgress, formatMetricValue, getTrendEmoji } from './api/business-metrics';
|
|
28
|
+
import { qualityMetrics, passesQualityThreshold, isHallucinationRiskAcceptable, getQualityRecommendations, formatQualityScore, getGradeColor } from './api/quality-metrics';
|
|
18
29
|
import { linking, generateZendeskMarker, parseZendeskMarker, hasZendeskMarker, removeZendeskMarker, linkRunToTicket, linkRunToZendeskTicket, getBestLinkMethod, LINK_METHOD_CONFIDENCE } from './integrations/ticket-linking';
|
|
19
30
|
import { customerContext, captureCustomerContext, getContextAsOf, toContextSnapshot, calculateArrChange, calculateHealthTrend } from './integrations/customer-context';
|
|
20
|
-
export type { InitOptions, Framework, RunOptions, RunOutcome, ConversationMessage, CustomerContextSnapshot, TicketLinkingOptions, LinkMethod, TraceOptions, SpanData, BusinessContext, Claim, ClaimType, ClaimCategory, ConfidenceCalibration, EvidenceReference, AnalysisResult, CalibrationStatus, CalibrationBucket, PredictionType, RoiConfig, RoiSummary, ExplainabilityResult, ApiResponse, PaginatedResponse, RunResponse, } from './core/types';
|
|
31
|
+
export type { InitOptions, Framework, RunOptions, RunOutcome, ConversationMessage, CustomerContextSnapshot, TicketLinkingOptions, LinkMethod, TraceOptions, TraceCustomFlag, SpanData, BusinessContext, Claim, ClaimType, ClaimCategory, ConfidenceCalibration, EvidenceReference, AnalysisResult, CalibrationStatus, CalibrationBucket, PredictionType, RoiConfig, RoiSummary, ExplainabilityResult, ApiResponse, PaginatedResponse, RunResponse, } from './core/types';
|
|
32
|
+
export type { HumanReviewStatus, HumanReviewType, HumanReviewQueueItem, CalibrationSet, ReviewerCalibration, QueueStats, } from './api/human-review';
|
|
33
|
+
export type { NondeterminismRunType, NondeterminismRunStatus, NondeterminismRun, NondeterminismSample, TraceAnalysis, CriterionAnalysis, RunSummary, } from './api/nondeterminism';
|
|
34
|
+
export type { SaturationType, HealthStatus, RegressionSeverity, EvalHealthSnapshot, EvalRegression, HealthReport, } from './api/eval-health';
|
|
35
|
+
export type { RuleType, DeterministicEvalResult, RuleResult, RuleTypeInfo, RuleTemplate, } from './api/deterministic-graders';
|
|
36
|
+
export type { AggregateMethod, SessionTrace, TurnEvaluation, ConversationEvalResult, AggregationMethodInfo, } from './api/conversation-eval';
|
|
37
|
+
export type { PatternCategory, PatternOutcome, PatternSeverity, PatternMatch, PatternInsight, AnalysisResult as PatternAnalysisResult, PatternCategoryInfo, BuiltInPattern, } from './api/transcript-patterns';
|
|
38
|
+
export type { ApiKeyPermissions, ScopeType, Environment, CreateApiKeyOptions, ApiKey, CreateApiKeyResult, } from './api/apiKeys';
|
|
39
|
+
export type { IndustryConfig, CustomIndustryConfig, ROIMetrics, BusinessImpact, ROISummary, TrendDataPoint, Correlation, PatternCluster, CorrelationAnalysis, } from './api/roi-analytics';
|
|
40
|
+
export type { RetrievedContext, GroundTruthContext, GroundedSpan, UngroundedSpan, CitationMap, RAGEvaluation, RAGEvidence, HallucinationInstance, HallucinationReport, GroundednessResult, BatchEvaluationResult, BatchEvaluationSummary, } from './api/quality-metrics';
|
|
41
|
+
export type { MetricTrend, MetricStatus, CurrentMetricResponse, MetricHistoryPoint, MetricHistorySummary, MetricHistoryResponse, RecordMetricOptions, RecordMetricResponse, } from './api/business-metrics';
|
|
42
|
+
export type { Agent, CreateAgentOptions, UpdateAgentOptions, AgentConfig, UpdateAgentConfigOptions, DeleteAgentOptions, CascadeImpact, } from './api/agents';
|
|
21
43
|
/**
|
|
22
44
|
* Initialize ThinkHive SDK v3
|
|
23
45
|
*
|
|
@@ -75,52 +97,128 @@ export declare function traceChain<T>(options: {
|
|
|
75
97
|
input?: unknown;
|
|
76
98
|
}, fn: () => Promise<T>): Promise<T>;
|
|
77
99
|
/**
|
|
78
|
-
*
|
|
79
|
-
*
|
|
100
|
+
* Issues API client
|
|
101
|
+
* This is the recommended API for managing clustered failure patterns
|
|
102
|
+
*
|
|
103
|
+
* @since 3.1.0
|
|
80
104
|
*/
|
|
81
|
-
export declare const
|
|
105
|
+
export declare const issues: {
|
|
106
|
+
/**
|
|
107
|
+
* List issues for an agent
|
|
108
|
+
*/
|
|
109
|
+
list(agentId: string, options?: {
|
|
110
|
+
status?: string;
|
|
111
|
+
startDate?: Date;
|
|
112
|
+
endDate?: Date;
|
|
113
|
+
limit?: number;
|
|
114
|
+
offset?: number;
|
|
115
|
+
}): Promise<any[]>;
|
|
116
|
+
/**
|
|
117
|
+
* Get a single issue by ID
|
|
118
|
+
*/
|
|
119
|
+
get(id: string): Promise<any>;
|
|
82
120
|
/**
|
|
83
|
-
*
|
|
84
|
-
* @deprecated Use runs.create() + claims.getRunAnalysis() instead
|
|
121
|
+
* Create a new issue
|
|
85
122
|
*/
|
|
86
|
-
|
|
87
|
-
|
|
88
|
-
|
|
89
|
-
|
|
90
|
-
|
|
91
|
-
|
|
92
|
-
|
|
123
|
+
create(data: {
|
|
124
|
+
agentId: string;
|
|
125
|
+
title: string;
|
|
126
|
+
description?: string;
|
|
127
|
+
type: string;
|
|
128
|
+
severity?: string;
|
|
129
|
+
pattern?: string;
|
|
130
|
+
exampleTraceIds?: string[];
|
|
131
|
+
}): Promise<any>;
|
|
132
|
+
/**
|
|
133
|
+
* Update an issue
|
|
134
|
+
*/
|
|
135
|
+
update(id: string, data: {
|
|
136
|
+
title?: string;
|
|
137
|
+
description?: string;
|
|
138
|
+
status?: string;
|
|
139
|
+
severity?: string;
|
|
140
|
+
assignedTo?: string;
|
|
141
|
+
resolutionNotes?: string;
|
|
142
|
+
}): Promise<any>;
|
|
143
|
+
/**
|
|
144
|
+
* Get fixes for an issue
|
|
145
|
+
*/
|
|
146
|
+
getFixes(issueId: string): Promise<any[]>;
|
|
147
|
+
};
|
|
148
|
+
/**
|
|
149
|
+
* Analyzer API client
|
|
150
|
+
* User-selected trace analysis with cost estimation and smart sampling
|
|
151
|
+
*
|
|
152
|
+
* Key improvements over v1 Explainer:
|
|
153
|
+
* - User-selected trace analysis (not automatic)
|
|
154
|
+
* - Cost estimation before execution
|
|
155
|
+
* - Smart sampling strategies
|
|
156
|
+
* - Root cause analysis by layer
|
|
157
|
+
* - Pattern aggregation across traces
|
|
158
|
+
*
|
|
159
|
+
* @since 3.1.0
|
|
160
|
+
*/
|
|
161
|
+
export declare const analyzer: {
|
|
93
162
|
/**
|
|
94
|
-
*
|
|
95
|
-
* @deprecated
|
|
163
|
+
* Analyze specific traces (user-selected)
|
|
96
164
|
*/
|
|
97
|
-
|
|
165
|
+
analyze(options: {
|
|
166
|
+
traceIds: string[];
|
|
167
|
+
tier?: "fast" | "standard" | "deep";
|
|
168
|
+
includeRootCause?: boolean;
|
|
169
|
+
includeLayers?: boolean;
|
|
170
|
+
}): Promise<any>;
|
|
98
171
|
/**
|
|
99
|
-
*
|
|
100
|
-
* @deprecated
|
|
172
|
+
* Analyze traces by time window
|
|
101
173
|
*/
|
|
102
|
-
|
|
103
|
-
|
|
174
|
+
analyzeWindow(options: {
|
|
175
|
+
agentId: string;
|
|
176
|
+
startDate: Date;
|
|
177
|
+
endDate: Date;
|
|
104
178
|
filters?: {
|
|
105
|
-
|
|
106
|
-
|
|
107
|
-
dateRange?: {
|
|
108
|
-
from: Date;
|
|
109
|
-
to: Date;
|
|
110
|
-
};
|
|
179
|
+
outcomes?: ("failure" | "error" | "success")[];
|
|
180
|
+
minSeverity?: "low" | "medium" | "high" | "critical";
|
|
111
181
|
};
|
|
112
|
-
|
|
182
|
+
sampling?: {
|
|
183
|
+
strategy: "all" | "failures_only" | "smart" | "random";
|
|
184
|
+
samplePercent?: number;
|
|
185
|
+
};
|
|
186
|
+
}): Promise<any>;
|
|
187
|
+
/**
|
|
188
|
+
* Estimate cost before running analysis
|
|
189
|
+
*/
|
|
190
|
+
estimateCost(options: {
|
|
191
|
+
traceIds?: string[];
|
|
192
|
+
agentId?: string;
|
|
193
|
+
startDate?: Date;
|
|
194
|
+
endDate?: Date;
|
|
195
|
+
tier: "fast" | "standard" | "deep";
|
|
113
196
|
}): Promise<{
|
|
114
|
-
|
|
115
|
-
|
|
116
|
-
|
|
117
|
-
|
|
197
|
+
estimatedTraces: number;
|
|
198
|
+
estimatedTokens: number;
|
|
199
|
+
estimatedCost: number;
|
|
200
|
+
estimatedCredits: number;
|
|
201
|
+
tier: string;
|
|
202
|
+
note: string;
|
|
118
203
|
}>;
|
|
204
|
+
/**
|
|
205
|
+
* Get analysis results for a specific trace
|
|
206
|
+
*/
|
|
207
|
+
get(traceId: string): Promise<any>;
|
|
208
|
+
/**
|
|
209
|
+
* Aggregate insights across multiple analyzed traces
|
|
210
|
+
*/
|
|
211
|
+
summarize(options: {
|
|
212
|
+
analysisIds?: string[];
|
|
213
|
+
agentId?: string;
|
|
214
|
+
startDate?: Date;
|
|
215
|
+
endDate?: Date;
|
|
216
|
+
}): Promise<any>;
|
|
119
217
|
};
|
|
120
218
|
export { isInitialized, SDK_VERSION, DEFAULT_ENDPOINT };
|
|
121
|
-
export { runs, claims, calibration, linking, customerContext, };
|
|
122
|
-
export { createRunWithContext, toOpenAIMessages, fromOpenAIMessages, isFact, isInference, isComputed, getHighConfidenceClaims, groupClaimsByType, groupClaimsByCategory, calculateBrierScore, calculateECE, isWellCalibrated, getCalibrationQuality, generateZendeskMarker, parseZendeskMarker, hasZendeskMarker, removeZendeskMarker, linkRunToTicket, linkRunToZendeskTicket, getBestLinkMethod, LINK_METHOD_CONFIDENCE, captureCustomerContext, getContextAsOf, toContextSnapshot, calculateArrChange, calculateHealthTrend, };
|
|
123
|
-
export { ThinkHiveApiError, ThinkHiveValidationError };
|
|
219
|
+
export { runs, claims, calibration, linking, customerContext, humanReview, nondeterminism, evalHealth, deterministicGraders, conversationEval, transcriptPatterns, apiKeys, agents, roiAnalytics, qualityMetrics, businessMetrics, };
|
|
220
|
+
export { createRunWithContext, toOpenAIMessages, fromOpenAIMessages, isFact, isInference, isComputed, getHighConfidenceClaims, groupClaimsByType, groupClaimsByCategory, calculateBrierScore, calculateECE, isWellCalibrated, getCalibrationQuality, generateZendeskMarker, parseZendeskMarker, hasZendeskMarker, removeZendeskMarker, linkRunToTicket, linkRunToZendeskTicket, getBestLinkMethod, LINK_METHOD_CONFIDENCE, captureCustomerContext, getContextAsOf, toContextSnapshot, calculateArrChange, calculateHealthTrend, calculatePassAtK, calculatePassToK, requiredPassRateForPassAtK, isReliableEvaluation, getReliabilityRecommendation, hasHealthIssue, getHealthSeverityLevel, isSaturated, getSaturationRecommendation, createRegexRule, createContainsRule, createLengthRule, createJsonSchemaRule, allRulesPassed, getFailedRules, calculateAverageScore, aggregateWorst, aggregateAverage, aggregateWeighted, aggregateFinalTurn, aggregateMajority, getAggregator, getProblematicTurns, analyzeConversationTrend, isHighRisk, getMatchesByCategory, getCriticalInsights, hasPiiExposure, hasFrustrationSignals, hasEscalationRequest, getCategoryDistribution, getRecommendations, needsAttention, sortMatchesBySeverity, hasPermission, isApiKeyExpired, isApiKeyValid, getTimeUntilExpiry, canAccessAgent, calculateRevenueAtRisk, calculateAutomationSavings, formatCurrency, getROIQuality, passesQualityThreshold, isHallucinationRiskAcceptable, getQualityRecommendations, formatQualityScore, getGradeColor, isMetricReady, needsMoreTraces, awaitingExternalData, isMetricStale, getStatusMessage, getTraceProgress, formatMetricValue, getTrendEmoji, };
|
|
221
|
+
export { ThinkHiveError, ThinkHiveApiError, ThinkHiveValidationError, PermissionDeniedError, AgentScopeError, RateLimitError, IpWhitelistError, };
|
|
124
222
|
declare const _default: {
|
|
125
223
|
init: typeof init;
|
|
126
224
|
getTracer: typeof getTracer;
|
|
@@ -209,11 +307,6 @@ declare const _default: {
|
|
|
209
307
|
calibration: {
|
|
210
308
|
status(agentId: string, predictionType: import("./core/types").PredictionType): Promise<import("./core/types").CalibrationStatus>;
|
|
211
309
|
allMetrics(agentId: string): Promise<import("./api/calibration").CalibrationMetrics[]>;
|
|
212
|
-
recordOutcome(input: import("./api/calibration").RecordOutcomeInput): Promise<{
|
|
213
|
-
recorded: boolean;
|
|
214
|
-
brierContribution: number;
|
|
215
|
-
message: string;
|
|
216
|
-
}>;
|
|
217
310
|
retrain(agentId: string, options?: {
|
|
218
311
|
predictionTypes?: import("./core/types").PredictionType[];
|
|
219
312
|
minSamples?: number;
|
|
@@ -226,15 +319,6 @@ declare const _default: {
|
|
|
226
319
|
}>;
|
|
227
320
|
newMetrics: import("./api/calibration").CalibrationMetrics[];
|
|
228
321
|
}>;
|
|
229
|
-
reliabilityDiagram(agentId: string, predictionType: import("./core/types").PredictionType): Promise<{
|
|
230
|
-
agentId: string;
|
|
231
|
-
predictionType: import("./core/types").PredictionType;
|
|
232
|
-
buckets: import("./core/types").CalibrationBucket[];
|
|
233
|
-
perfectCalibrationLine: Array<{
|
|
234
|
-
x: number;
|
|
235
|
-
y: number;
|
|
236
|
-
}>;
|
|
237
|
-
}>;
|
|
238
322
|
};
|
|
239
323
|
linking: {
|
|
240
324
|
create(input: import("./integrations/ticket-linking").CreateLinkInput): Promise<import("./integrations/ticket-linking").LinkResponse>;
|
|
@@ -288,44 +372,302 @@ declare const _default: {
|
|
|
288
372
|
getLatestSnapshot(customerId: string): Promise<import("./integrations/customer-context").CustomerMetricsSnapshot | null>;
|
|
289
373
|
getSnapshotAsOf(customerId: string, timestamp: string | Date): Promise<import("./integrations/customer-context").CustomerMetricsSnapshot | null>;
|
|
290
374
|
};
|
|
291
|
-
|
|
375
|
+
humanReview: {
|
|
376
|
+
getQueue(options?: import("./api/human-review").ListQueueOptions): Promise<import("./api/human-review").HumanReviewQueueItem[]>;
|
|
377
|
+
addToQueue(options: import("./api/human-review").AddToQueueOptions): Promise<import("./api/human-review").HumanReviewQueueItem>;
|
|
378
|
+
getItem(itemId: string): Promise<import("./api/human-review").HumanReviewQueueItem>;
|
|
379
|
+
claim(itemId: string): Promise<import("./api/human-review").HumanReviewQueueItem>;
|
|
380
|
+
release(itemId: string): Promise<import("./api/human-review").HumanReviewQueueItem>;
|
|
381
|
+
skip(itemId: string): Promise<import("./api/human-review").HumanReviewQueueItem>;
|
|
382
|
+
submit(itemId: string, review: import("./api/human-review").SubmitReviewOptions): Promise<import("./api/human-review").HumanReviewQueueItem>;
|
|
383
|
+
getStats(agentId?: string): Promise<import("./api/human-review").QueueStats>;
|
|
384
|
+
getNextItem(agentId?: string): Promise<import("./api/human-review").HumanReviewQueueItem | null>;
|
|
385
|
+
getReviewTypes(): Promise<Array<{
|
|
386
|
+
id: string;
|
|
387
|
+
name: string;
|
|
388
|
+
description: string;
|
|
389
|
+
autoTrigger: boolean;
|
|
390
|
+
}>>;
|
|
391
|
+
getCalibrationSets(agentId?: string): Promise<import("./api/human-review").CalibrationSet[]>;
|
|
392
|
+
createCalibrationSet(options: import("./api/human-review").CreateCalibrationSetOptions): Promise<import("./api/human-review").CalibrationSet>;
|
|
393
|
+
getCalibrationSet(setId: string): Promise<import("./api/human-review").CalibrationSet>;
|
|
394
|
+
getCertifiedReviewers(calibrationSetId: string): Promise<import("./api/human-review").ReviewerCalibration[]>;
|
|
395
|
+
getReviewerCalibrations(userId: string): Promise<import("./api/human-review").ReviewerCalibration[]>;
|
|
396
|
+
};
|
|
397
|
+
nondeterminism: {
|
|
398
|
+
createRun(options: import("./api/nondeterminism").CreateRunOptions): Promise<import("./api/nondeterminism").NondeterminismRun>;
|
|
399
|
+
getRuns(options?: import("./api/nondeterminism").ListRunsOptions): Promise<import("./api/nondeterminism").NondeterminismRun[]>;
|
|
400
|
+
getRun(runId: string): Promise<import("./api/nondeterminism").NondeterminismRun>;
|
|
401
|
+
startRun(runId: string): Promise<void>;
|
|
402
|
+
completeRun(runId: string): Promise<void>;
|
|
403
|
+
recordSample(options: import("./api/nondeterminism").RecordSampleOptions): Promise<import("./api/nondeterminism").NondeterminismSample>;
|
|
404
|
+
getSamples(runId: string): Promise<import("./api/nondeterminism").NondeterminismSample[]>;
|
|
405
|
+
getRunSummary(runId: string): Promise<import("./api/nondeterminism").RunSummary>;
|
|
406
|
+
analyzeRun(runId: string): Promise<import("./api/nondeterminism").RunSummary>;
|
|
407
|
+
getInfo(): Promise<import("./api/nondeterminism").PassAtKInfo>;
|
|
408
|
+
};
|
|
409
|
+
evalHealth: {
|
|
410
|
+
getReport(agentId: string): Promise<import("./api/eval-health").HealthReport>;
|
|
411
|
+
getSnapshots(options: import("./api/eval-health").GetSnapshotsOptions): Promise<import("./api/eval-health").EvalHealthSnapshot[]>;
|
|
412
|
+
getLatestSnapshot(agentId: string, criterionId?: string): Promise<import("./api/eval-health").EvalHealthSnapshot | null>;
|
|
413
|
+
recordSnapshot(options: import("./api/eval-health").CreateSnapshotOptions): Promise<import("./api/eval-health").EvalHealthSnapshot>;
|
|
414
|
+
getRegressions(agentId: string): Promise<import("./api/eval-health").EvalRegression[]>;
|
|
415
|
+
recordRegression(options: import("./api/eval-health").CreateRegressionOptions): Promise<import("./api/eval-health").EvalRegression>;
|
|
416
|
+
resolveRegression(regressionId: string, options: import("./api/eval-health").ResolveRegressionOptions): Promise<void>;
|
|
417
|
+
acknowledgeRegression(regressionId: string): Promise<void>;
|
|
418
|
+
};
|
|
419
|
+
deterministicGraders: {
|
|
420
|
+
evaluate(options: import("./api/deterministic-graders").EvaluateOptions): Promise<import("./api/deterministic-graders").DeterministicEvalResult>;
|
|
421
|
+
bulkEvaluate(options: import("./api/deterministic-graders").BulkEvaluateOptions): Promise<import("./api/deterministic-graders").BulkEvaluateResult>;
|
|
422
|
+
getRuleTypes(): Promise<import("./api/deterministic-graders").RuleTypeInfo[]>;
|
|
423
|
+
getTemplates(): Promise<import("./api/deterministic-graders").RuleTemplate[]>;
|
|
424
|
+
};
|
|
425
|
+
conversationEval: {
|
|
426
|
+
getSessionTraces(sessionId: string): Promise<import("./api/conversation-eval").SessionTrace[]>;
|
|
427
|
+
evaluate(options: import("./api/conversation-eval").EvaluateConversationOptions): Promise<import("./api/conversation-eval").ConversationEvalResult>;
|
|
428
|
+
getAggregationMethods(): Promise<import("./api/conversation-eval").AggregationMethodInfo[]>;
|
|
429
|
+
};
|
|
430
|
+
transcriptPatterns: {
|
|
431
|
+
analyze(traceId: string): Promise<import("./api/transcript-patterns").AnalysisResult>;
|
|
432
|
+
bulkAnalyze(traceIds: string[]): Promise<import("./api/transcript-patterns").BulkAnalysisResult>;
|
|
433
|
+
getCategories(): Promise<import("./api/transcript-patterns").PatternCategoryInfo[]>;
|
|
434
|
+
getBuiltInPatterns(): Promise<import("./api/transcript-patterns").BuiltInPattern[]>;
|
|
435
|
+
};
|
|
436
|
+
apiKeys: {
|
|
437
|
+
create: typeof import("./api/apiKeys").create;
|
|
438
|
+
list: typeof import("./api/apiKeys").list;
|
|
439
|
+
revoke: typeof import("./api/apiKeys").revoke;
|
|
440
|
+
rotate: typeof import("./api/apiKeys").rotate;
|
|
441
|
+
test: typeof import("./api/apiKeys").test;
|
|
442
|
+
hasPermission: typeof hasPermission;
|
|
443
|
+
isExpired: typeof isApiKeyExpired;
|
|
444
|
+
isValid: typeof isApiKeyValid;
|
|
445
|
+
getTimeUntilExpiry: typeof getTimeUntilExpiry;
|
|
446
|
+
canAccessAgent: typeof canAccessAgent;
|
|
447
|
+
};
|
|
448
|
+
agents: {
|
|
449
|
+
list(options?: {
|
|
450
|
+
companyId?: string;
|
|
451
|
+
departmentId?: string;
|
|
452
|
+
}): Promise<import("./api/agents").Agent[]>;
|
|
453
|
+
get(id: string): Promise<import("./api/agents").Agent>;
|
|
454
|
+
create(options: import("./api/agents").CreateAgentOptions): Promise<import("./api/agents").Agent>;
|
|
455
|
+
update(id: string, options: import("./api/agents").UpdateAgentOptions): Promise<import("./api/agents").Agent>;
|
|
456
|
+
delete(id: string, options?: import("./api/agents").DeleteAgentOptions): Promise<void | import("./api/agents").CascadeImpact>;
|
|
457
|
+
getConfig(agentId: string): Promise<import("./api/agents").AgentConfig | null>;
|
|
458
|
+
updateConfig(agentId: string, config: import("./api/agents").UpdateAgentConfigOptions): Promise<import("./api/agents").AgentConfig>;
|
|
459
|
+
};
|
|
460
|
+
issues: {
|
|
292
461
|
/**
|
|
293
|
-
*
|
|
294
|
-
* @deprecated Use runs.create() + claims.getRunAnalysis() instead
|
|
462
|
+
* List issues for an agent
|
|
295
463
|
*/
|
|
296
|
-
|
|
297
|
-
|
|
298
|
-
|
|
299
|
-
|
|
300
|
-
|
|
301
|
-
|
|
302
|
-
}): Promise<
|
|
464
|
+
list(agentId: string, options?: {
|
|
465
|
+
status?: string;
|
|
466
|
+
startDate?: Date;
|
|
467
|
+
endDate?: Date;
|
|
468
|
+
limit?: number;
|
|
469
|
+
offset?: number;
|
|
470
|
+
}): Promise<any[]>;
|
|
303
471
|
/**
|
|
304
|
-
* Get a
|
|
305
|
-
* @deprecated
|
|
472
|
+
* Get a single issue by ID
|
|
306
473
|
*/
|
|
307
|
-
get(
|
|
474
|
+
get(id: string): Promise<any>;
|
|
308
475
|
/**
|
|
309
|
-
*
|
|
310
|
-
* @deprecated
|
|
476
|
+
* Create a new issue
|
|
311
477
|
*/
|
|
312
|
-
|
|
313
|
-
|
|
478
|
+
create(data: {
|
|
479
|
+
agentId: string;
|
|
480
|
+
title: string;
|
|
481
|
+
description?: string;
|
|
482
|
+
type: string;
|
|
483
|
+
severity?: string;
|
|
484
|
+
pattern?: string;
|
|
485
|
+
exampleTraceIds?: string[];
|
|
486
|
+
}): Promise<any>;
|
|
487
|
+
/**
|
|
488
|
+
* Update an issue
|
|
489
|
+
*/
|
|
490
|
+
update(id: string, data: {
|
|
491
|
+
title?: string;
|
|
492
|
+
description?: string;
|
|
493
|
+
status?: string;
|
|
494
|
+
severity?: string;
|
|
495
|
+
assignedTo?: string;
|
|
496
|
+
resolutionNotes?: string;
|
|
497
|
+
}): Promise<any>;
|
|
498
|
+
/**
|
|
499
|
+
* Get fixes for an issue
|
|
500
|
+
*/
|
|
501
|
+
getFixes(issueId: string): Promise<any[]>;
|
|
502
|
+
};
|
|
503
|
+
analyzer: {
|
|
504
|
+
/**
|
|
505
|
+
* Analyze specific traces (user-selected)
|
|
506
|
+
*/
|
|
507
|
+
analyze(options: {
|
|
508
|
+
traceIds: string[];
|
|
509
|
+
tier?: "fast" | "standard" | "deep";
|
|
510
|
+
includeRootCause?: boolean;
|
|
511
|
+
includeLayers?: boolean;
|
|
512
|
+
}): Promise<any>;
|
|
513
|
+
/**
|
|
514
|
+
* Analyze traces by time window
|
|
515
|
+
*/
|
|
516
|
+
analyzeWindow(options: {
|
|
517
|
+
agentId: string;
|
|
518
|
+
startDate: Date;
|
|
519
|
+
endDate: Date;
|
|
314
520
|
filters?: {
|
|
315
|
-
|
|
316
|
-
|
|
317
|
-
dateRange?: {
|
|
318
|
-
from: Date;
|
|
319
|
-
to: Date;
|
|
320
|
-
};
|
|
521
|
+
outcomes?: ("failure" | "error" | "success")[];
|
|
522
|
+
minSeverity?: "low" | "medium" | "high" | "critical";
|
|
321
523
|
};
|
|
322
|
-
|
|
524
|
+
sampling?: {
|
|
525
|
+
strategy: "all" | "failures_only" | "smart" | "random";
|
|
526
|
+
samplePercent?: number;
|
|
527
|
+
};
|
|
528
|
+
}): Promise<any>;
|
|
529
|
+
/**
|
|
530
|
+
* Estimate cost before running analysis
|
|
531
|
+
*/
|
|
532
|
+
estimateCost(options: {
|
|
533
|
+
traceIds?: string[];
|
|
534
|
+
agentId?: string;
|
|
535
|
+
startDate?: Date;
|
|
536
|
+
endDate?: Date;
|
|
537
|
+
tier: "fast" | "standard" | "deep";
|
|
538
|
+
}): Promise<{
|
|
539
|
+
estimatedTraces: number;
|
|
540
|
+
estimatedTokens: number;
|
|
541
|
+
estimatedCost: number;
|
|
542
|
+
estimatedCredits: number;
|
|
543
|
+
tier: string;
|
|
544
|
+
note: string;
|
|
545
|
+
}>;
|
|
546
|
+
/**
|
|
547
|
+
* Get analysis results for a specific trace
|
|
548
|
+
*/
|
|
549
|
+
get(traceId: string): Promise<any>;
|
|
550
|
+
/**
|
|
551
|
+
* Aggregate insights across multiple analyzed traces
|
|
552
|
+
*/
|
|
553
|
+
summarize(options: {
|
|
554
|
+
analysisIds?: string[];
|
|
555
|
+
agentId?: string;
|
|
556
|
+
startDate?: Date;
|
|
557
|
+
endDate?: Date;
|
|
558
|
+
}): Promise<any>;
|
|
559
|
+
};
|
|
560
|
+
roiAnalytics: {
|
|
561
|
+
summary(options?: {
|
|
562
|
+
startDate?: string | Date;
|
|
563
|
+
endDate?: string | Date;
|
|
564
|
+
agentId?: string;
|
|
565
|
+
}): Promise<import("./api/roi-analytics").ROISummary>;
|
|
566
|
+
byAgent(agentId: string, options?: {
|
|
567
|
+
startDate?: string | Date;
|
|
568
|
+
endDate?: string | Date;
|
|
323
569
|
}): Promise<{
|
|
324
|
-
|
|
325
|
-
|
|
570
|
+
agent: {
|
|
571
|
+
id: string;
|
|
572
|
+
name: string;
|
|
573
|
+
industry: string;
|
|
574
|
+
};
|
|
575
|
+
industryConfig: Partial<import("./api/roi-analytics").IndustryConfig>;
|
|
576
|
+
roi: import("./api/roi-analytics").ROIMetrics;
|
|
577
|
+
recentImpacts: Array<{
|
|
578
|
+
impactScore: number;
|
|
579
|
+
revenueRisk: number;
|
|
580
|
+
roiCategory: string;
|
|
581
|
+
totalFinancialImpact: number;
|
|
326
582
|
}>;
|
|
327
|
-
total: number;
|
|
328
583
|
}>;
|
|
584
|
+
trends(options?: {
|
|
585
|
+
startDate?: string | Date;
|
|
586
|
+
endDate?: string | Date;
|
|
587
|
+
agentId?: string;
|
|
588
|
+
}): Promise<import("./api/roi-analytics").TrendDataPoint[]>;
|
|
589
|
+
calculate(options: {
|
|
590
|
+
traceId?: string;
|
|
591
|
+
userMessage?: string;
|
|
592
|
+
agentResponse?: string;
|
|
593
|
+
industryConfig?: import("./api/roi-analytics").CustomIndustryConfig;
|
|
594
|
+
}): Promise<import("./api/roi-analytics").BusinessImpact>;
|
|
595
|
+
industries(): Promise<import("./api/roi-analytics").IndustryConfig[]>;
|
|
596
|
+
correlations(options?: {
|
|
597
|
+
startDate?: string | Date;
|
|
598
|
+
endDate?: string | Date;
|
|
599
|
+
agentId?: string;
|
|
600
|
+
}): Promise<import("./api/roi-analytics").CorrelationAnalysis>;
|
|
601
|
+
};
|
|
602
|
+
qualityMetrics: {
|
|
603
|
+
getRagScores(traceId: string): Promise<{
|
|
604
|
+
traceId: string;
|
|
605
|
+
evaluation: import("./api/quality-metrics").RAGEvaluation;
|
|
606
|
+
evidence: import("./api/quality-metrics").RAGEvidence;
|
|
607
|
+
}>;
|
|
608
|
+
getHallucinationReport(traceId: string): Promise<{
|
|
609
|
+
traceId: string;
|
|
610
|
+
report: import("./api/quality-metrics").HallucinationReport;
|
|
611
|
+
}>;
|
|
612
|
+
evaluateRag(input: {
|
|
613
|
+
query: string;
|
|
614
|
+
response: string;
|
|
615
|
+
retrievedContexts: import("./api/quality-metrics").RetrievedContext[];
|
|
616
|
+
groundTruthContexts?: import("./api/quality-metrics").GroundTruthContext[];
|
|
617
|
+
citations?: string[];
|
|
618
|
+
}): Promise<{
|
|
619
|
+
evaluation: import("./api/quality-metrics").RAGEvaluation;
|
|
620
|
+
evidence: import("./api/quality-metrics").RAGEvidence;
|
|
621
|
+
}>;
|
|
622
|
+
detectHallucinations(input: {
|
|
623
|
+
response: string;
|
|
624
|
+
contexts: Array<{
|
|
625
|
+
content: string;
|
|
626
|
+
metadata?: Record<string, unknown>;
|
|
627
|
+
}>;
|
|
628
|
+
query?: string;
|
|
629
|
+
previousResponses?: string[];
|
|
630
|
+
}): Promise<{
|
|
631
|
+
report: import("./api/quality-metrics").HallucinationReport;
|
|
632
|
+
}>;
|
|
633
|
+
getGroundedness(traceId: string): Promise<{
|
|
634
|
+
traceId: string;
|
|
635
|
+
groundedness: import("./api/quality-metrics").GroundednessResult;
|
|
636
|
+
spans: {
|
|
637
|
+
grounded: Array<{
|
|
638
|
+
text: string;
|
|
639
|
+
confidence: number;
|
|
640
|
+
sourceIndex: number;
|
|
641
|
+
}>;
|
|
642
|
+
ungrounded: Array<{
|
|
643
|
+
text: string;
|
|
644
|
+
confidence: number;
|
|
645
|
+
}>;
|
|
646
|
+
};
|
|
647
|
+
summary: {
|
|
648
|
+
totalSpans: number;
|
|
649
|
+
groundedSpans: number;
|
|
650
|
+
ungroundedSpans: number;
|
|
651
|
+
groundednessRatio: number;
|
|
652
|
+
};
|
|
653
|
+
}>;
|
|
654
|
+
evaluateBatch(options: {
|
|
655
|
+
traceIds: string[];
|
|
656
|
+
includeDetails?: boolean;
|
|
657
|
+
}): Promise<{
|
|
658
|
+
summary: import("./api/quality-metrics").BatchEvaluationSummary;
|
|
659
|
+
results: import("./api/quality-metrics").BatchEvaluationResult[];
|
|
660
|
+
}>;
|
|
661
|
+
};
|
|
662
|
+
businessMetrics: {
|
|
663
|
+
current(agentId: string, metricName?: string): Promise<import("./api/business-metrics").CurrentMetricResponse>;
|
|
664
|
+
history(agentId: string, metricName: string, options?: {
|
|
665
|
+
startDate?: string | Date;
|
|
666
|
+
endDate?: string | Date;
|
|
667
|
+
granularity?: "hourly" | "daily" | "weekly" | "monthly";
|
|
668
|
+
}): Promise<import("./api/business-metrics").MetricHistoryResponse>;
|
|
669
|
+
record(agentId: string, options: import("./api/business-metrics").RecordMetricOptions): Promise<import("./api/business-metrics").RecordMetricResponse>;
|
|
670
|
+
recordBatch(agentId: string, metrics: import("./api/business-metrics").RecordMetricOptions[]): Promise<import("./api/business-metrics").RecordMetricResponse[]>;
|
|
329
671
|
};
|
|
330
672
|
generateZendeskMarker: typeof generateZendeskMarker;
|
|
331
673
|
parseZendeskMarker: typeof parseZendeskMarker;
|
|
@@ -335,5 +677,16 @@ declare const _default: {
|
|
|
335
677
|
isInference: typeof isInference;
|
|
336
678
|
calculateBrierScore: typeof calculateBrierScore;
|
|
337
679
|
isWellCalibrated: typeof isWellCalibrated;
|
|
680
|
+
calculatePassAtK: typeof calculatePassAtK;
|
|
681
|
+
isReliableEvaluation: typeof isReliableEvaluation;
|
|
682
|
+
isHighRisk: typeof isHighRisk;
|
|
683
|
+
needsAttention: typeof needsAttention;
|
|
684
|
+
calculateRevenueAtRisk: typeof calculateRevenueAtRisk;
|
|
685
|
+
formatCurrency: typeof formatCurrency;
|
|
686
|
+
passesQualityThreshold: typeof passesQualityThreshold;
|
|
687
|
+
isHallucinationRiskAcceptable: typeof isHallucinationRiskAcceptable;
|
|
688
|
+
isMetricReady: typeof isMetricReady;
|
|
689
|
+
needsMoreTraces: typeof needsMoreTraces;
|
|
690
|
+
getStatusMessage: typeof getStatusMessage;
|
|
338
691
|
};
|
|
339
692
|
export default _default;
|