feature: minimal topology endpoint (stretch)

[xnorkl]

Hackathon positioning

Stretch goal for v0.3.0-Hackathon. Not required for PCAP replay → Zeek → asset ingest unless product/judges explicitly need /api/topology. Core milestone path is #49, #50, #116.

---

Summary

Ship a read-only minimal topology endpoint that returns one snapshot document matching minimal_network-topology-schema-v0.json and the example in-repo.

Context

This is the v0.3.0 topology contract. The endpoint should be small, deterministic, and schema-first. Asset.hostname is used as asset.id; until #137 is complete, assets without hostname are excluded and that behavior is documented.

Goal

Expose a single topology snapshot route that returns schema-valid JSON and is documented in OpenAPI as the same contract version.

Acceptance criteria

• GET /api/topology/ (or agreed equivalent) returns one snapshot object that validates against minimal_network-topology-schema-v0.json (automated test with jsonschema).
• Response includes required root fields (schema_version, snapshot_id, timestamp, assets) and uses the minimal schema's field types/enums/patterns.
• asset.id is sourced from Asset.hostname; assets with null/empty hostname are excluded until Enforce hostname as stable identity on Asset for topology #137, and this is stated in API docs/OpenAPI description.
• Empty-state/error behavior is explicitly defined and tested (pick one: 200 with empty assets or 404).
• OpenAPI for the topology endpoint points to the same minimal schema artifact/version used by tests.

Implementation guidance (non-binding)

• Reuse existing models (Asset, NetworkEndpoint) without introducing new topology models in this issue.
• Keep the endpoint read-only and non-paginated.
• Derive services in a way that stays schema-valid; detailed enrichment logic can evolve in follow-ups.

Non-goals

• Full-schema topology work (feat: AssetInterface model (full-schema topology, post-v0.3.0) #132, feat: full-schema topology endpoint (post-v0.3.0) #133).
• Changing the minimal schema itself.
• Expanding scope to broader REST cleanup work.