neo4j_client.py computes semantic temporal similarity by scanning user event relationships with embeddings. This will not scale.
Acceptance criteria:
- Use Neo4j vector indexes or mirror temporal vectors into Pinecone
- Add date-range indexes and query paths.
- Benchmark temporal search at 10k, 100k, and 1M events per tenant.
neo4j_client.pycomputes semantic temporal similarity by scanning user event relationships with embeddings. This will not scale.Acceptance criteria: