package saas import ( "context" "crypto/sha256" "encoding/hex" "encoding/json" "errors" "fmt" "strings" "time" "data-terminal-cloud/internal/audit" "data-terminal-cloud/internal/database" "data-terminal-cloud/internal/identity" "data-terminal-cloud/internal/outbox" "github.com/jackc/pgx/v5" ) var ErrExportUnavailable = errors.New("export download is not available") type exportURLSigner interface { PresignDownload(context.Context, string, time.Duration, string) (string, error) } type ExportTask struct { ID string `json:"id"` SubjectID string `json:"subject_id"` Scope map[string]any `json:"scope"` Status string `json:"status"` ObjectSize *int64 `json:"object_size,omitempty"` ContentDigest string `json:"content_digest,omitempty"` ExpiresAt *time.Time `json:"expires_at,omitempty"` Attempts int `json:"attempts"` LastError string `json:"last_error,omitempty"` CreatedAt time.Time `json:"created_at"` UpdatedAt time.Time `json:"updated_at"` CompletedAt *time.Time `json:"completed_at,omitempty"` DownloadURL string `json:"download_url,omitempty"` DownloadUntil *time.Time `json:"download_until,omitempty"` objectRef string } func (s *Service) ConfigureExportSigning(signer exportURLSigner, downloadTTL time.Duration) { s.exportSigner = signer s.exportDownloadTTL = downloadTTL } func (s *Service) CreateExportTask(ctx context.Context, principal identity.Principal, subjectPublicID, idempotencyKey, requestID string, scope map[string]any) (ExportTask, bool, error) { if scope == nil { scope = map[string]any{} } requestBytes, err := json.Marshal(scope) if err != nil { return ExportTask{}, false, err } digestValue := sha256.Sum256(requestBytes) digest := hex.EncodeToString(digestValue[:]) created := false var publicID string err = database.Transact(ctx, s.pool, func(tx pgx.Tx) error { access, _, err := resolveSubject(ctx, tx, principal, subjectPublicID, identity.CapabilityPrivacyManage, true) if err != nil { return err } scopeJSON, _ := json.Marshal(scope) var internalID string err = tx.QueryRow(ctx, `INSERT INTO export_task(tenant_id,subject_id,requested_by,scope,idempotency_key,request_digest) VALUES($1::uuid,$2::uuid,$3::uuid,$4::jsonb,$5,$6) ON CONFLICT (tenant_id,requested_by,idempotency_key) DO NOTHING RETURNING id::text,public_id::text`, access.tenantID, access.subjectID, access.userID, scopeJSON, idempotencyKey, digest).Scan(&internalID, &publicID) if errors.Is(err, pgx.ErrNoRows) { var storedDigest string if err := tx.QueryRow(ctx, `SELECT public_id::text,request_digest FROM export_task WHERE tenant_id=$1::uuid AND requested_by=$2::uuid AND idempotency_key=$3 FOR UPDATE`, access.tenantID, access.userID, idempotencyKey).Scan(&publicID, &storedDigest); err != nil { return err } if storedDigest != digest { return ErrIdempotencyConflict } return nil } if err != nil { return fmt.Errorf("create export task: %w", err) } created = true dedupe := "export-generate:" + internalID payload, _ := json.Marshal(map[string]any{"export_task_id": internalID}) if _, err := outbox.NewStore(s.pool).Enqueue(ctx, tx, outbox.NewEvent{TenantID: &access.tenantID, AggregateType: "export_task", AggregateID: internalID, EventType: "export.generate_requested", DeduplicationKey: &dedupe, Payload: payload}); err != nil { return err } return s.audit.Record(ctx, tx, audit.Event{TenantID: strPtr(access.tenantID), ActorType: "user", ActorPublicID: strPtr(access.userPublicID), Action: "export.request", ResourceType: "export_task", ResourceID: strPtr(publicID), RequestID: requestID, Result: "success", Detail: map[string]any{"subject_id": subjectPublicID, "scope": scope}}) }) if err != nil { return ExportTask{}, false, err } result, err := s.GetExportTask(ctx, principal, publicID) return result, created, err } func (s *Service) ListExportTasks(ctx context.Context, principal identity.Principal, subjectPublicID string) ([]ExportTask, error) { access, _, err := resolveSubject(ctx, s.pool, principal, subjectPublicID, identity.CapabilityPrivacyRead, false) if err != nil { return nil, err } rows, err := s.pool.Query(ctx, exportTaskSelectSQL+` WHERE et.tenant_id=$1::uuid AND et.subject_id=$2::uuid ORDER BY et.created_at DESC,et.id DESC`, access.tenantID, access.subjectID) if err != nil { return nil, err } defer rows.Close() items := make([]ExportTask, 0) for rows.Next() { item, err := scanExportTask(rows) if err != nil { return nil, err } items = append(items, item) } return items, rows.Err() } func (s *Service) GetExportTask(ctx context.Context, principal identity.Principal, publicID string) (ExportTask, error) { row := s.pool.QueryRow(ctx, exportTaskSelectSQL+` JOIN household h ON h.id=sp.household_id AND h.tenant_id=sp.tenant_id AND h.status<>'deleted' JOIN household_member hm ON hm.household_id=h.id AND hm.tenant_id=h.tenant_id AND hm.status='active' JOIN user_account ua ON ua.id=hm.user_id AND ua.tenant_id=h.tenant_id AND ua.status='active' WHERE et.public_id=$1::uuid AND sp.status IN ('active','suspended') AND ua.oidc_issuer=$2 AND ua.oidc_subject=$3`, publicID, principal.Issuer, principal.Subject) item, err := scanExportTask(row) if errors.Is(err, pgx.ErrNoRows) { return ExportTask{}, ErrAccessDenied } if err != nil { return ExportTask{}, err } var role string if err := s.pool.QueryRow(ctx, `SELECT hm.role FROM export_task et JOIN subject_profile sp ON sp.id=et.subject_id AND sp.status IN ('active','suspended') JOIN household_member hm ON hm.household_id=sp.household_id AND hm.status='active' JOIN user_account ua ON ua.id=hm.user_id AND ua.tenant_id=et.tenant_id AND ua.status='active' WHERE et.public_id=$1::uuid AND ua.oidc_issuer=$2 AND ua.oidc_subject=$3`, publicID, principal.Issuer, principal.Subject).Scan(&role); err != nil || !identity.Can(identity.Role(role), identity.CapabilityPrivacyRead) { return ExportTask{}, ErrAccessDenied } if item.Status == "READY" && item.ExpiresAt != nil { if s.exportSigner == nil || s.exportDownloadTTL <= 0 { return ExportTask{}, ErrExportUnavailable } ttl := min(s.exportDownloadTTL, time.Until(*item.ExpiresAt)) if ttl <= 0 { item.Status = "EXPIRED" return item, nil } url, err := s.exportSigner.PresignDownload(ctx, item.objectRef, ttl, "subject-data-"+item.ID+".json") if err != nil { return ExportTask{}, fmt.Errorf("presign export download: %w", err) } until := time.Now().Add(ttl) item.DownloadURL = url item.DownloadUntil = &until } return item, nil } const exportTaskSelectSQL = `SELECT et.public_id::text,sp.public_id::text,et.scope,et.status,et.object_ref,et.object_size,COALESCE(et.content_digest,''),et.expires_at,et.attempts,COALESCE(et.last_error,''),et.created_at,et.updated_at,et.completed_at FROM export_task et JOIN subject_profile sp ON sp.id=et.subject_id AND sp.tenant_id=et.tenant_id` func scanExportTask(row rowScanner) (ExportTask, error) { var item ExportTask var scopeJSON []byte var objectRef *string if err := row.Scan(&item.ID, &item.SubjectID, &scopeJSON, &item.Status, &objectRef, &item.ObjectSize, &item.ContentDigest, &item.ExpiresAt, &item.Attempts, &item.LastError, &item.CreatedAt, &item.UpdatedAt, &item.CompletedAt); err != nil { return ExportTask{}, err } if objectRef != nil { item.objectRef = *objectRef } if err := decodeObject(scopeJSON, &item.Scope); err != nil { return ExportTask{}, err } return item, nil } func NormalizeExportCategories(values []string) ([]string, error) { allowed := map[string]bool{"profile": true, "consents": true, "device_bindings": true, "conversations": true, "transcripts": true, "summaries": true, "stories": true, "memories": true} seen := map[string]bool{} result := make([]string, 0, len(values)) for _, raw := range values { value := strings.TrimSpace(raw) if !allowed[value] { return nil, fmt.Errorf("unsupported export category %q", value) } if !seen[value] { seen[value] = true result = append(result, value) } } return result, nil }