refactor(web): evaluation configure schema update

web/app/components/evaluation/__tests__/index.spec.tsx

@@ -245,12 +245,10 @@ describe('Evaluation', () => {
       data: {
         evaluation_model: null,
         evaluation_model_provider: null,
-        metrics_config: {
-          default_metrics: [{
-            metric: 'context-precision',
-          }],
-          customized_metrics: null,
-        },
+        default_metrics: [{
+          metric: 'context-precision',
+        }],
+        customized_metrics: null,
         judgement_conditions: null,
       },
     })

web/app/components/evaluation/__tests__/store.spec.ts

@@ -132,18 +132,16 @@ describe('evaluation store', () => {
     const config: EvaluationConfig = {
       evaluation_model: 'gpt-4o-mini',
       evaluation_model_provider: 'openai',
-      metrics_config: {
-        default_metrics: [{
-          metric: 'faithfulness',
-          node_info_list: [
-            { node_id: 'node-1', title: 'Retriever', type: 'retriever' },
-          ],
-        }],
-        customized_metrics: {
-          evaluation_workflow_id: 'workflow-precision-review',
-          input_fields: {
-            'app.input.query': 'query',
-          },
+      default_metrics: [{
+        metric: 'faithfulness',
+        node_info_list: [
+          { node_id: 'node-1', title: 'Retriever', type: 'retriever' },
+        ],
+      }],
+      customized_metrics: {
+        evaluation_workflow_id: 'workflow-precision-review',
+        input_fields: {
+          'app.input.query': 'query',
         },
       },
       judgement_conditions: [{

web/app/components/evaluation/components/metric-selector/utils.ts

@@ -3,7 +3,7 @@ import type { MetricVisualTone } from './types'
 import type { EvaluationTargetType, NodeInfo } from '@/types/evaluation'
 
 export const toEvaluationTargetType = (resourceType: 'apps' | 'snippets'): EvaluationTargetType => {
-  return resourceType === 'snippets' ? 'snippets' : 'app'
+  return resourceType === 'snippets' ? 'snippets' : 'apps'
 }
 
 const humanizeMetricId = (metricId: string) => {

web/app/components/evaluation/store-utils.ts

@@ -16,7 +16,6 @@ import type {
   EvaluationDefaultMetric,
   EvaluationJudgementConditionGroup,
   EvaluationJudgementConditionItem,
-  EvaluationMetricsConfig,
   NodeInfo,
 } from '@/types/evaluation'
 import { getComparisonOperators, getDefaultOperator, getEvaluationMockConfig } from './mock'
@@ -71,7 +70,7 @@ const normalizeNodeInfoList = (value: NodeInfo[] | undefined): NodeInfo[] => {
 
 const normalizeDefaultMetrics = (
   resourceType: EvaluationResourceType,
-  value: EvaluationDefaultMetric[] | undefined,
+  value: EvaluationDefaultMetric[] | null | undefined,
 ): EvaluationMetric[] => {
   if (!value?.length)
     return []
@@ -292,9 +291,8 @@ export const buildStateFromEvaluationConfig = (
   resourceType: EvaluationResourceType,
   config: EvaluationConfig,
 ): EvaluationResourceState => {
-  const metricsConfig: EvaluationMetricsConfig = config.metrics_config ?? {}
-  const defaultMetrics = normalizeDefaultMetrics(resourceType, metricsConfig.default_metrics)
-  const customMetrics = normalizeCustomMetric(metricsConfig.customized_metrics)
+  const defaultMetrics = normalizeDefaultMetrics(resourceType, config.default_metrics)
+  const customMetrics = normalizeCustomMetric(config.customized_metrics)
 
   return {
     ...buildInitialState(resourceType),

web/service/use-evaluation.ts

@@ -29,30 +29,14 @@ const normalizeAvailableEvaluationWorkflowsParams = (params: AvailableEvaluation
   }
 }
 
-const toEvaluationTargetType = (resourceType: Exclude<EvaluationResourceType, 'datasets'>) => {
-  return resourceType === 'snippets' ? 'snippets' : 'app'
-}
-
 const getEvaluationConfigQueryOptions = (
   resourceType: EvaluationResourceType,
   resourceId: string,
 ) => {
-  if (resourceType === 'datasets') {
-    return consoleQuery.datasetEvaluation.config.queryOptions({
-      input: {
-        params: {
-          datasetId: resourceId,
-        },
-      },
-      enabled: !!resourceId,
-      refetchOnWindowFocus: false,
-    })
-  }
-
   return consoleQuery.evaluation.config.queryOptions({
     input: {
       params: {
-        targetType: toEvaluationTargetType(resourceType),
+        targetType: resourceType,
         targetId: resourceId,
       },
     },

web/types/evaluation.ts

@@ -1,9 +1,4 @@
-export type EvaluationTargetType = 'app' | 'snippets'
-
-export type EvaluationMetricsConfig = {
-  default_metrics?: EvaluationDefaultMetric[]
-  customized_metrics?: EvaluationCustomizedMetric | null
-}
+export type EvaluationTargetType = 'apps' | 'snippets' | 'datasets'
 
 export type EvaluationConditionValue = string | number | boolean | null
 
@@ -31,7 +26,8 @@ export type EvaluationJudgementConditions
 export type EvaluationConfig = {
   evaluation_model: string | null
   evaluation_model_provider: string | null
-  metrics_config: EvaluationMetricsConfig | null
+  default_metrics?: EvaluationDefaultMetric[] | null
+  customized_metrics?: EvaluationCustomizedMetric | null
   judgement_conditions: EvaluationJudgementConditions | null
 }
 
@@ -43,19 +39,23 @@ export type NodeInfo = {
 
 export type EvaluationDefaultMetric = {
   metric?: string
+  value_type?: string
   node_info_list?: NodeInfo[]
 }
 
 export type EvaluationCustomizedMetric = {
   evaluation_workflow_id?: string
   input_fields?: Record<string, string | null | undefined>
-  output_fields?: Array<Record<string, string | null | undefined>>
+  output_fields?: Array<{
+    variable?: string
+    value_type?: string
+  }>
 }
 
 export type EvaluationConfigData = {
   evaluation_model?: string
   evaluation_model_provider?: string
-  default_metrics?: EvaluationDefaultMetric[]
+  default_metrics?: EvaluationDefaultMetric[] | null
   customized_metrics?: EvaluationCustomizedMetric | null
   judgment_config?: EvaluationJudgementConditions | null
 }