Modeling & application code generation
ecosystem-server hosts modeling projects: UML diagrams, ontology context, and async jobs that generate application code from your model. Workbench Project Generate on a deployment creates plugin/git artefacts for the Runtime; modeling generation is the server-side path for full application stacks.
Ontology concepts: Modeling Server · Server catalog.
Modeling project shape
| Piece | Role |
|---|---|
ModelingProject | Container for diagrams and generation settings |
Diagram | Class, sequence, or other UML views |
OntologyContext | Domain ids, structured model, prompt context, roles |
UmlGenerationJob | Analyze or generate UML from prompts |
AppGenerationJob | Async application code generation from the model |
OntologySynthesisRun | Merge diagrams / data / roles → Turtle export |
Domain seed catalog (public): ecosystem-domain-catalog — business-domain Turtle and SHACL used when usesDomainCatalog is set on the ontology context.
Generated client stacks
Application generation can emit integration code for the same closed loop documented on the developer site:
| Output | Use |
|---|---|
| TypeScript / JavaScript | Browser or Node clients |
| Python | Scripts and notebooks |
| React | Component examples |
| Vue | Composition-API friendly snippets |
Workbench Deployment Test → Integrate tab produces the same stack list for an already pushed deployment (Deployment Test). Modeling generation targets the application implied by your UML/ontology project rather than a single deployment’s plugin JAR.
Deployment → Generate (see Deployment parameters — Build and Generate) creates a git branch with Runtime plugin code for that deployment. Modeling → App generation on ecosystem-server creates application projects from diagrams and ontology context. Use both when you need plugins on the Runtime and a separate client app.
Ontology-aware generation
Recent ontology work connects:
- Workbench Ontology Management — mappings,
resolve-term, MCPgenerate_ontology_query/query_data_via_ontology(Ontology for agents) - Domain catalog — shared industry vocabulary and optional source mappings (Data and Ontology)
- Modeling synthesis —
OntologySynthesisRunexports Turtle aligned with diagrams
When Ontology Management is enabled, prefer resolving business terms to mapped fields before generating queries or client payloads.
Docker images (reference only)
Ecogenetic publishes refreshed ecosystem-server, ecosystem-modeling, and related images on Docker Hub (ecosystemai/*). Pin by digest or an explicit tag your install documents — do not assume latest matches a specific semver unless your release notes say so.
Related
- Project — Workbench project document
- Deployment parameters — plugin Generate/Build flags
- Public API — Ontology