story/internal/privacy/access_processor.go

package privacy

import (
	"context"
	"encoding/json"
	"errors"
	"fmt"
	"time"

	"data-terminal-cloud/internal/audit"
	"data-terminal-cloud/internal/database"
	"data-terminal-cloud/internal/outbox"

	"github.com/jackc/pgx/v5"
	"github.com/jackc/pgx/v5/pgxpool"
)

type AccessProcessor struct {
	pool  *pgxpool.Pool
	audit *audit.Repository
}

func NewAccessProcessor(pool *pgxpool.Pool) (*AccessProcessor, error) {
	if pool == nil {
		return nil, errors.New("database pool is required")
	}
	return &AccessProcessor{pool: pool, audit: audit.NewRepository()}, nil
}

func (p *AccessProcessor) Handle(ctx context.Context, event outbox.Event) error {
	return database.Transact(ctx, p.pool, func(tx pgx.Tx) error {
		var tenantID, subjectID, publicID, status, requestType string
		var scopeJSON []byte
		if err := tx.QueryRow(ctx, `SELECT tenant_id::text,subject_id::text,public_id::text,status,request_type,scope FROM privacy_request WHERE id=$1::uuid FOR UPDATE`, event.AggregateID).Scan(&tenantID, &subjectID, &publicID, &status, &requestType, &scopeJSON); errors.Is(err, pgx.ErrNoRows) {
			return nil
		} else if err != nil {
			return fmt.Errorf("load privacy access request: %w", err)
		}
		if status == "COMPLETED" {
			return nil
		}
		if requestType != "ACCESS" {
			return fmt.Errorf("privacy request type %s is not handled by access processor", requestType)
		}
		if _, err := tx.Exec(ctx, `UPDATE privacy_request SET status='VERIFYING',version=version+1,updated_at=now() WHERE id=$1::uuid AND status='SUBMITTED'`, event.AggregateID); err != nil {
			return err
		}
		if _, err := tx.Exec(ctx, `UPDATE privacy_request SET status='EXECUTING',version=version+1,updated_at=now() WHERE id=$1::uuid AND status IN ('VERIFYING','PARTIAL_FAILED')`, event.AggregateID); err != nil {
			return err
		}
		if _, err := tx.Exec(ctx, `UPDATE privacy_subtask SET status='RUNNING',attempts=attempts+1,last_error=NULL,updated_at=now() WHERE privacy_request_id=$1::uuid AND subsystem='postgresql' AND target_type='subject_data_inventory'`, event.AggregateID); err != nil {
			return err
		}

		counts, err := collectInventory(ctx, tx, tenantID, subjectID)
		if err != nil {
			return err
		}
		var scope map[string]any
		if err := json.Unmarshal(scopeJSON, &scope); err != nil {
			return fmt.Errorf("decode privacy scope: %w", err)
		}
		result := map[string]any{
			"inventory_counts": filterInventory(counts, scope),
			"generated_at":     time.Now().UTC().Format(time.RFC3339Nano),
			"content_included": false,
		}
		resultJSON, err := json.Marshal(result)
		if err != nil {
			return err
		}
		if _, err := tx.Exec(ctx, `UPDATE privacy_subtask SET status='COMPLETED',completed_at=now(),updated_at=now() WHERE privacy_request_id=$1::uuid AND subsystem='postgresql' AND target_type='subject_data_inventory'`, event.AggregateID); err != nil {
			return err
		}
		if _, err := tx.Exec(ctx, `UPDATE privacy_request SET status='COMPLETED',result_summary=$2::jsonb,completed_at=now(),version=version+1,updated_at=now() WHERE id=$1::uuid`, event.AggregateID, resultJSON); err != nil {
			return err
		}
		return p.audit.Record(ctx, tx, audit.Event{TenantID: &tenantID, ActorType: "system", Action: "privacy.access.complete", ResourceType: "privacy_request", ResourceID: &publicID, Result: "success", Detail: map[string]any{"subject_digest_scope": "tenant_subject", "content_included": false}})
	})
}

func collectInventory(ctx context.Context, tx pgx.Tx, tenantID, subjectID string) (map[string]int64, error) {
	queries := map[string]string{
		"profile":         `SELECT count(*) FROM subject_profile WHERE tenant_id=$1::uuid AND id=$2::uuid`,
		"consents":        `SELECT count(*) FROM consent_record WHERE tenant_id=$1::uuid AND subject_id=$2::uuid`,
		"device_bindings": `SELECT count(*) FROM device_binding WHERE tenant_id=$1::uuid AND subject_id=$2::uuid`,
		"conversations":   `SELECT count(*) FROM conversation WHERE tenant_id=$1::uuid AND subject_id=$2::uuid`,
		"transcripts":     `SELECT count(*) FROM conversation_turn WHERE tenant_id=$1::uuid AND subject_id=$2::uuid`,
		"summaries":       `SELECT count(*) FROM conversation_summary WHERE tenant_id=$1::uuid AND subject_id=$2::uuid`,
		"stories":         `SELECT count(*) FROM story_draft WHERE tenant_id=$1::uuid AND subject_id=$2::uuid`,
		"memories":        `SELECT count(*) FROM curated_memory WHERE tenant_id=$1::uuid AND subject_id=$2::uuid`,
	}
	result := make(map[string]int64, len(queries))
	for category, query := range queries {
		var count int64
		if err := tx.QueryRow(ctx, query, tenantID, subjectID).Scan(&count); err != nil {
			return nil, fmt.Errorf("count privacy inventory category %s: %w", category, err)
		}
		result[category] = count
	}
	return result, nil
}

func filterInventory(counts map[string]int64, scope map[string]any) map[string]int64 {
	requested, ok := scope["categories"].([]any)
	if !ok || len(requested) == 0 {
		return counts
	}
	result := make(map[string]int64, len(requested))
	for _, value := range requested {
		category, ok := value.(string)
		if !ok {
			continue
		}
		if count, exists := counts[category]; exists {
			result[category] = count
		}
	}
	return result
}