Overview

This section provides a high-level overview of ScriptBee's architecture and the roles of its core components.

UI

The User Interface provides a web-based experience for managing projects, loading data from various sources, and running analysis scripts.

The Gateway is the central orchestration component. Its responsibilities include:

NOTE

Authentication & Authorization are handled by a separate, external service. The Gateway integrates with this service to manage project access.

Analysis is responsible for the core logic of processing models and executing scripts. It performs the following functions:

Data Loading: Ingesting data from various sources using loader plugins.
Model Linking: Connecting loaded models using linker plugins to create a unified data structure.
Script Execution: Running C#, Javascript, or Python scripts against the linked models.

Analysis can be performed in two modes:

Permanent Instances: Manually managed by the user for long-running exploration and interactive analysis.
Temporary Instances: Automatically created for one-time processing, ideal for automated pipelines or quick checks.

Both the Gateway and Analysis services rely on a MongoDB database for data persistence:

Gateway: Stores project configurations, workspace metadata, and instance statuses.
Analysis: Uses the database for temporary data storage during processing and for persisting analysis results.

The orchestration of analysis components varies based on the environment:

In standard containerized environments, a Docker Executor manages analysis instances dynamically. It handles:

Networking: Ensuring the Gateway can communicate with the analysis instance.
Data Access: Mapping local data folders to the analysis container so loaders and scripts can access source files.

See Docker Compose for more information about deployment to Docker

In Kubernetes deployments, analysis is managed through specialized controllers that handle the lifecycle of each instance as a native resource.

See Kubernetes for more information about deployment to Kubernetes