`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: