ADR-0022: Import Export Framework¶
Status¶
Accepted
Context¶
InfraLynx needs a consistent transfer layer for CSV, JSON, and API-based bulk operations across multiple domains. The platform also needs dry-run validation, clean error reporting, and a path to asynchronous import execution through the job engine.
Decision¶
InfraLynx will implement import and export as a standalone transfer service with:
- dataset-owned transfer schemas
- transport-neutral validation rules
- explicit CSV, JSON, and API format support
- synchronous commit for small valid payloads
- job-engine-backed asynchronous import for larger batches
- file-backed transfer state as the bootstrap runtime
Consequences¶
Transfer behavior is now centralized and consistent across formats. The current implementation is operational for bootstrap workflows, while later chunks can replace file-backed transfer state with domain-aware persistence and richer rollback behavior without breaking the import/export contract shape.