Files
Assistant/cmd/agent-worker/task_evaluation_executor.go
m 7b09b3b2ec
CI / test (push) Successful in 45s
CI / test (pull_request) Successful in 46s
fix(cognitive): fail over task evaluation candidates
2026-08-23 20:52:09 +03:00

130 lines
6.9 KiB
Go

package main
import (
"bytes"
"context"
"encoding/json"
"fmt"
"net/http"
"strings"
"time"
"git.cin.su/m/Assistant/contracts/artifact"
"git.cin.su/m/Assistant/contracts/capability"
"git.cin.su/m/Assistant/contracts/work"
capabilitydomain "git.cin.su/m/Assistant/modules/capability/domain"
)
type taskEvaluationExecutor struct {
cognitiveURL string
callbackURL string
verifierURL string
client *http.Client
fallback localExecutor
}
func newTaskEvaluationExecutor(cognitiveURL, callbackURL, verifierURL string, client *http.Client) taskEvaluationExecutor {
if client == nil {
client = http.DefaultClient
}
return taskEvaluationExecutor{cognitiveURL: strings.TrimRight(cognitiveURL, "/"), callbackURL: strings.TrimRight(callbackURL, "/"), verifierURL: strings.TrimRight(verifierURL, "/"), client: client, fallback: newLocalExecutor()}
}
func (e taskEvaluationExecutor) Execute(ctx context.Context, request work.AgentExecutionRequest) (work.AgentExecutionResult, error) {
intent := request.Intent
if intent.TaskEvaluationExecution == nil || intent.TaskEvaluationProfile == nil || len(intent.TaskEvaluationCases) == 0 {
return e.fallback.Execute(ctx, request)
}
execution := intent.TaskEvaluationExecution
if execution.GraphID == "" || execution.NodeID == "" || execution.BindingRevision < 1 || execution.ActorID == "" || execution.PrincipalID == "" || execution.ScopeID == "" {
return e.fallback.Execute(ctx, request)
}
if e.cognitiveURL == "" || e.callbackURL == "" || e.verifierURL == "" {
return e.fallback.Execute(ctx, request)
}
payload := capability.TaskEvaluationExecutionRequest{Profile: *intent.TaskEvaluationProfile, CapabilityID: intent.TaskEvaluationProfile.CapabilityID, BindingID: intent.TaskEvaluationBindingID, BindingRevision: execution.BindingRevision, ActorID: execution.ActorID, PrincipalID: execution.PrincipalID, ScopeID: execution.ScopeID, Cases: intent.TaskEvaluationCases, MaxOutputBytes: 4096}
response, err := postJSON[capability.TaskEvaluationExecutionResponse](ctx, e.client, e.cognitiveURL+"/v1/task-evaluations/execute", payload)
if err != nil {
return work.AgentExecutionResult{}, err
}
if len(response.Responses) == 0 || response.ObservedAtUnixMillis < 1 {
return work.AgentExecutionResult{}, fmt.Errorf("task evaluation returned no observations")
}
if len(response.Observations) == len(intent.TaskEvaluationCases) && taskEvaluationHasCriticalFailure(response.Observations) {
evaluation, err := capabilitydomain.BuildTaskEvaluation(*intent.TaskEvaluationProfile, capability.TaskEvaluationID("evaluation:"+intent.DispatchID), response.BindingID, response.BindingRevision, intent.TaskEvaluationProfile.CapabilityID, response.Observations, artifact.ArtifactID(intent.TaskEvaluationCases[0].EvidenceArtifactRef), response.ObservedAtUnixMillis)
if err != nil {
return work.AgentExecutionResult{}, err
}
callback := work.TaskEvaluationEvidenceCallback{GraphID: execution.GraphID, Revision: execution.GraphRevision, NodeID: execution.NodeID, Evaluation: evaluation, Succeeded: true}
if _, err := postJSON[json.RawMessage](ctx, e.client, e.callbackURL+"/v1/work-graphs/task-evaluations/callback", callback); err != nil {
return work.AgentExecutionResult{}, err
}
return work.AgentExecutionResult{State: work.AgentExecutionCompleted, Summary: "task evaluation recorded as UNKNOWN after bounded provider recovery", OutcomeRef: string(evaluation.ID), ArtifactRefs: []string{string(evaluation.ArtifactID)}, ObservedAtUnixMillis: time.Now().UnixMilli()}, nil
}
if len(intent.TaskEvaluationCases) == 0 {
return work.AgentExecutionResult{}, fmt.Errorf("task evaluation cases are missing")
}
caseByID := make(map[string]capability.TaskEvaluationCase, len(intent.TaskEvaluationCases))
for _, item := range intent.TaskEvaluationCases {
caseByID[item.CaseID] = item
}
observations := make([]capability.TaskCaseObservation, 0, len(response.Responses))
for _, item := range response.Responses {
caseValue, ok := caseByID[item.CaseID]
if !ok || caseValue.Verification.Mode == "" {
return e.fallback.Execute(ctx, request)
}
verified, verifyErr := postJSON[capability.TaskCaseVerificationResponse](ctx, e.client, e.verifierURL+"/v1/task-evaluations/verify", capability.TaskCaseVerificationRequest{CaseID: item.CaseID, ResponseText: item.ResponseText, LatencyMicros: item.LatencyMicros, EvidenceArtifactRef: item.EvidenceArtifactRef, Verification: caseValue.Verification})
if verifyErr != nil {
return e.fallback.Execute(ctx, request)
}
observations = append(observations, capability.TaskCaseObservation{CaseID: verified.CaseID, Result: verified.Result, QualityScoreMillis: verified.QualityScoreMillis, LatencyMicros: verified.LatencyMicros, CriticalFailure: verified.CriticalFailure, EvidenceArtifactRef: verified.EvidenceArtifactRef, VerificationEvidenceRef: verified.VerificationEvidenceRef, VerifierIdentity: verified.VerifierIdentity})
}
evaluation, err := capabilitydomain.BuildTaskEvaluation(*intent.TaskEvaluationProfile, capability.TaskEvaluationID("evaluation:"+intent.DispatchID), response.BindingID, response.BindingRevision, intent.TaskEvaluationProfile.CapabilityID, observations, artifact.ArtifactID(intent.TaskEvaluationCases[0].EvidenceArtifactRef), response.ObservedAtUnixMillis)
if err != nil {
return work.AgentExecutionResult{}, err
}
callback := work.TaskEvaluationEvidenceCallback{GraphID: execution.GraphID, Revision: execution.GraphRevision, NodeID: execution.NodeID, Evaluation: evaluation, Succeeded: true}
if _, err := postJSON[json.RawMessage](ctx, e.client, e.callbackURL+"/v1/work-graphs/task-evaluations/callback", callback); err != nil {
return work.AgentExecutionResult{}, err
}
return work.AgentExecutionResult{State: work.AgentExecutionCompleted, Summary: "task evaluation executed, aggregated and accepted by Work Graph callback", OutcomeRef: string(evaluation.ID), ArtifactRefs: []string{string(evaluation.ArtifactID)}, ObservedAtUnixMillis: time.Now().UnixMilli()}, nil
}
func taskEvaluationHasCriticalFailure(observations []capability.TaskCaseObservation) bool {
for _, observation := range observations {
if observation.CriticalFailure {
return true
}
}
return false
}
func postJSON[T any](ctx context.Context, client *http.Client, url string, payload any) (T, error) {
var result T
body, err := json.Marshal(payload)
if err != nil {
return result, err
}
req, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(body))
if err != nil {
return result, err
}
req.Header.Set("Content-Type", "application/json")
resp, err := client.Do(req)
if err != nil {
return result, err
}
defer resp.Body.Close()
if resp.StatusCode < 200 || resp.StatusCode >= 300 {
return result, fmt.Errorf("POST %s: status %s", url, resp.Status)
}
if resp.StatusCode == http.StatusNoContent {
return result, nil
}
if err := json.NewDecoder(resp.Body).Decode(&result); err != nil {
return result, err
}
return result, nil
}