How epuerta9/kitchenai works
Main components
• Application / Control Plane: Represented by the “kitchenai” directory and its various submodules (apps, core, dashboard, etc.). This component provides the unified API for application developers.
• Messaging Layer: Emphasize the use of NATS (referenced by configuration files, e.g. NATS_URL and related configuration scripts) that enables messaging, service discovery, and asynchronous communication among components.
• External Dependencies: Include external systems such as the relational database (Postgres, as indicated by DATABASE_URL), S3-compatible storage (using Minio for local development), Redis for caching if enabled, and monitoring/observability tools (Grafana is configured).
• Plugin Ecosystem: Illustrate additional extensibility points such as those in “plugins-community” and “whisk” which provide additional logic, integrations, or experimental features.
Relationships and Data Flow
• External components (database, cache, S3 cloud storage, and monitoring dashboards) are connected to the Control Plane. For instance, database interactions for persistence and minio/S3 for media storage must be represented; similarly, monitoring tools (Grafana) ingest logs/metrics from various components.
Architectural Patterns and Design Principles
• Loose coupling and modularity: Components such as individual bento modules are independently deployable and replaceable without affecting the overall system.
• Support for plug & play components: Indicate the plugin ecosystem where additional modules (like the Deepeval plugin) can be integrated seamlessly.
• Notation of service dependency and communication protocols (HTTP for backend API and NATS messaging for inter-service communication) is important.
Relevant Technologies and Frameworks
• Backend: Likely Django (hinted by common Django file names and structure), possibly with additional async handlers.
• Messaging: NATS as the high-performance messaging layer.
• Data Storage: PostgreSQL (database), Redis (if used), and S3/Minio for media storage.
• Frontend/Presentation: While the emphasis is on the API and admin/dashboard layers indicated by templates and static assets, note the existence of interactive playgrounds.
• Containerization and Deployment: Docker compose files, deployment scripts, and configuration files (e.g., for Grafana and NATS).