[refactor] Drop duplicate EvaluationRun type, align step-reference shape

web/oss/src/components/EvalRunDetails/atoms/table/run.ts

@@ -1,3 +1,4 @@
+import type {EvaluationRun} from "@agenta/entities/evaluationRun"
 import {fetchWorkflowsBatch} from "@agenta/entities/workflow"
 import {atomFamily, selectAtom} from "jotai/utils"
 import {atomWithQuery} from "jotai-tanstack-query"
@@ -13,8 +14,6 @@ import {
 import {TERMINAL_STATUSES} from "../compare"
 import {effectiveProjectIdAtom} from "../run"
 
-import type {EvaluationRun} from "@/agenta-oss-common/lib/hooks/usePreviewEvaluations/types"
-
 export interface EvaluationRunQueryResult {
     rawRun: EvaluationRun
     camelRun: any

web/oss/src/components/EvalRunDetails/components/CompareRunsMenu.tsx

@@ -34,7 +34,7 @@ interface CompareRunsMenuProps {
 interface CandidateRun {
     id: string
     name: string
-    status?: string
+    status?: string | null
     description?: string | null
     createdAt?: string
     testsetNames: string[]

web/oss/src/components/EvaluationRunsTablePOC/types.ts

@@ -1,9 +1,9 @@
+import type {EvaluationRun} from "@agenta/entities/evaluationRun"
+
 import type {InfiniteTableRowBase} from "@/oss/components/InfiniteVirtualTable/types"
 import type {WindowingState} from "@/oss/components/InfiniteVirtualTable/types"
 import type {SnakeToCamelCaseKeys} from "@/oss/lib/Types"
 
-import type {EvaluationRun} from "@/agenta-oss-common/lib/hooks/usePreviewEvaluations/types"
-
 // The original `@/oss/state/evaluations/legacyAtoms` module no longer exists, and `legacy`
 // is only ever read through `any` casts — represented as an opaque record (per WP-4i).
 export type LegacyAutoEvaluation = Record<string, unknown>

web/oss/src/lib/hooks/usePreviewEvaluations/index.ts

@@ -127,7 +127,9 @@ interface PreviewEvaluationsQueryState {
     error: unknown
 }
 import {searchQueryAtom} from "./states/queryFilterAtoms"
-import {EnrichedEvaluationRun, EvaluationRun} from "./types"
+import type {EvaluationRun} from "@agenta/entities/evaluationRun"
+
+import {EnrichedEvaluationRun} from "./types"
 
 const SCENARIOS_ENDPOINT = "/evaluations/scenarios/"
 

web/oss/src/lib/hooks/usePreviewEvaluations/types.ts

@@ -1,3 +1,4 @@
+import type {EvaluationRun} from "@agenta/entities/evaluationRun"
 import type {Workflow} from "@agenta/entities/workflow"
 
 import {PreviewTestset, SnakeToCamelCaseKeys, WorkspaceMember} from "@/oss/lib/Types"
@@ -37,45 +38,6 @@ export type EvaluationRunDataStep =
       }
 
 export type IEvaluationRunDataStep = SnakeToCamelCaseKeys<EvaluationRunDataStep>
-export interface EvaluationRun {
-    /** Unique identifier for the evaluation run */
-    id: string
-    /** Display name for the run */
-    name: string
-    /** Optional description text for the run */
-    description: string
-    /** ISO timestamp of when the run was created */
-    created_at: string
-    /** ISO timestamp of when the run was last updated (present on backend run payloads) */
-    updated_at?: string
-    /** ID of the user who created the run */
-    created_by_id: string
-    /** Optional metadata object (arbitrary key-value pairs) */
-    meta: Record<string, any>
-    /** Flags associated with the run (internal use) */
-    flags: Record<string, any>
-    /** Current status of the run (e.g., "pending", "completed") */
-    status: string
-    data: {
-        /** Array of evaluation steps that define execution flow */
-        steps: EvaluationRunDataStep[]
-        /** Mappings define how to extract values from steps for display or evaluation */
-        mappings: {
-            /** Type of the mapping, determines what the value represents */
-            kind: "input" | "ground_truth" | "application" | "evaluator"
-            /** Display name for the mapped value */
-            name: string
-            /** Path reference to the data inside a step */
-            step: {
-                /** The step key this mapping belongs to */
-                key: string
-                /** Path within the step data (e.g., 'country' or 'data.outputs.metric') */
-                path: string
-            }
-        }[]
-    }
-}
-
 export interface EnrichedEvaluationRun extends SnakeToCamelCaseKeys<EvaluationRun> {
     /** All distinct testsets referenced in this run */
     testsets: PreviewTestset[]

web/packages/agenta-entities/src/evaluationRun/core/schema.ts

@@ -47,6 +47,7 @@ export const evaluationRunStepReferenceSchema = z.object({
     slug: z.string().nullable().optional(),
     version: z.coerce.number().nullable().optional(),
 })
+export type EvaluationRunStepReference = z.infer<typeof evaluationRunStepReferenceSchema>
 
 export const evaluationRunDataStepSchema = z.object({
     key: z.string(),

web/packages/agenta-entities/src/evaluationRun/etl/resolveMappings.ts

@@ -39,6 +39,7 @@
  */
 
 import type {EvaluationResult} from "../core"
+import type {EvaluationRunStepReference} from "../core/schema"
 
 import type {HydratedScenarioRow, HydratableScenario} from "./hydrateScenariosTransform"
 
@@ -54,7 +55,8 @@ export interface RunStep {
      */
     type: string
     origin?: string | null
-    references?: Record<string, {id: string; slug?: string; version?: string} | null> | null
+    /** Shape comes from the validated core schema; do not redeclare it here. */
+    references?: Record<string, EvaluationRunStepReference | null> | null
     inputs?: {key: string}[] | null
 }
 
@@ -108,7 +110,7 @@ export interface ColumnGroup {
     /** Stable cache key for this group — useful for grouping in renderers. */
     key: string
     /** The step.references that drove the grouping (preserved for downstream code). */
-    refs: Record<string, {id?: string; slug?: string} | null | undefined> | null
+    refs: Record<string, EvaluationRunStepReference | null | undefined> | null
 }
 
 export interface ResolvedColumn {