Operating Modes
Cassiopeia provides five interfaces for different stages of the data integration workflow. This page describes each mode and when to use it.
Quick Reference
Mode |
Interface |
Typical use |
CLI |
Command line |
Batch jobs, CI/CD, one-off conversions |
Server |
REST + WebSocket |
Centralized hub, live data streams |
Explorer |
Terminal (TUI) |
Schema discovery, model inspection |
Wizard |
Terminal (TUI) |
Interactive mapping creation |
Configurator |
Web browser |
Same as Wizard, graphical UI |
CLI Mode
CLI mode runs a mapping pipeline from the command line. You provide an input source (file or URL) and a mapping configuration; output is written to a local directory or pushed to an NGSI-LD Context Broker.
Use CLI mode for scheduled batch jobs, one-off conversions, and local development. No persistent service is required.
Server Mode
Server mode runs Cassiopeia as a persistent backend with REST and WebSocket APIs.
-
REST API — Manage mapping configurations, list supported models, and trigger transformations on demand
-
WebSocket — Real-time communication via MessagePack, suitable for monitoring progress on large datasets or streaming updates
Use Server mode when Cassiopeia acts as a central transformation hub—for example, continuous IoT feeds or environments where external services trigger transformations programmatically.
Explorer Mode
Explorer is a terminal UI for browsing and inspecting FIWARE Smart Data Models. You can search the catalog, view schema hierarchies, and inspect attribute metadata (descriptions, types, constraints, required flags).
Use Explorer before building mappings to understand the target schema. It works offline against a local catalog.
Wizard Mode
Wizard is a terminal UI that guides you through mapping configuration creation:
-
Select a target Smart Data Model
-
Configure entity ID templates (site, service, group prefixes)
-
Map source fields to model attributes with schema-aware validation
-
Export a
.json5mapping file for CLI or Server modes
Use Wizard when creating mappings for large models—the guided flow and validation reduce manual errors and iteration cycles.