All files / api/src/middlewares/optimizer hyperparameters.ts

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
 
 
1x
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
import type { UserDataStorageConnector } from '@api/types/connector';
import type { AppContext } from '@api/types/hono';
import { emitSSEEvent } from '@api/utils/sse-event-manager';
import type {
  SkillOptimizationArm,
  SkillOptimizationEvaluationResult,
} from '@shared/types/data';
 
export async function updatePulledArm(
  c: AppContext,
  userDataStorageConnector: UserDataStorageConnector,
  arm: SkillOptimizationArm,
  evaluationResults: SkillOptimizationEvaluationResult[],
) {
  // Extract evaluation_id and score pairs for atomic update
  const evaluationScores = evaluationResults.map((result) => ({
    evaluation_id: result.evaluation_id,
    score: result.score,
  }));
 
  await updateArmStats(c, userDataStorageConnector, arm, evaluationScores);
}
 
export async function updateArmStats(
  c: AppContext,
  userDataStorageConnector: UserDataStorageConnector,
  arm: SkillOptimizationArm,
  evaluationResults: Array<{ evaluation_id: string; score: number }>,
) {
  // Atomically update arm stats for each evaluation AND increment cluster/skill counters
  // PostgreSQL calculates new stats (n, mean, n2, total_reward) per evaluation internally
  // This ensures proper tracking of each evaluation method's performance
  const result = await userDataStorageConnector.updateArmAndIncrementCounters(
    c,
    arm.id,
    evaluationResults,
  );
 
  // Emit SSE event with all updated data (arm, cluster, skill)
  emitSSEEvent('skill-optimization:arm-updated', {
    arm: result.arm,
    cluster: result.cluster,
    skill: result.skill,
  });
}