v1.4.0

Resilience & Scale: Type safety hardening and storage pagination. Backward compatible with v1.3.0.

Added

Type Safety (S1)

• TaskRequest: TaskRequestConfig model for config (timeout_seconds, priority, idempotency_key, streaming, persist_state, model, temperature)
• TaskResponse: TaskMetrics model for metrics (duration_ms, tokens_in, tokens_out, tokens_used, api_calls)
• Entities: CommonMetadata model for conversation metadata (purpose, ttl_hours, source, timestamp, tags; extra allowed)
• Envelope: Payload typed as discriminated union; validator parses by payload_type; payload_dict property for backward compatibility

Storage Pagination (S2)

• SLAStorage: query_metrics(..., limit, offset) and count_metrics(...); GET /sla/history?limit=&offset= with total in response
• MeteringStorage: query(..., limit, offset); GET /usage?limit=&offset= with paginated results
• APIs: SLA history and Usage endpoints accept limit (default 100, max 1000) and offset; responses include count and (SLA) total

Examples

• v1.4.0 Showcase: uv run python -m asap.examples.v1_4_0_showcase — demonstrates pagination on Usage and SLA history APIs

Changed

• README: v1.4.0 Quick Info and showcase command; v1.4 (Resilience & Scale) marked complete in roadmap
• AGENTS.md: Status updated to v1.4.0

Technical Details

• Python: 3.13+
• Tests: 2335+ passing; type checker (mypy) and full test suite verified
• Coverage: Maintained

---

---

Full Changelog: v0.1.0...v1.4.0