Search Docs…
Jan 6, 2025
Getting Started
Data Flow
How does data flow through Leora, and how do the APIs work
Flow 1: Data Ingestion
Client (Auth: Bearer <access_token>) └─1. POST /v1/datastores └─▶ API Gateway (validate token, rate-limit) └─▶ Ingestion Service └─▶ Schema Registry └─▶ S3 Storage (profiles, interactions, content)
Flow 2: Model Training
Client (Auth: Bearer <access_token>) └─2. POST /v1/datastores/{ds}/train └─▶ API Gateway (validate token, rate-limit) └─▶ Training Orchestrator └─▶ Model Registry └─▶ S3 Storage (model checkpoints)
Flow 3a: User-Based Recommendations
Client (Auth: Bearer <access_token>) └─3a. GET /v1/datastores/{ds}/recommendations/user └─▶ API Gateway (validate token, rate-limit, check scopes) └─▶ Inference Service └─▶ Model Registry └─▶ Response: Top‑K Items for User
Flow 3b: Session-Based Recommendations
Client (Auth: Bearer <access_token>) └─3b. GET /v1/datastores/{ds}/recommendations/session └─▶ API Gateway (validate token, rate-limit, check scopes) └─▶ Inference Service └─▶ Model Registry └─▶ Response: Top‑K Items for Session