How jingw2/ontexus works
Ontexus turns connected or uploaded data into curated datasets and mapped domain ontologies, then exposes the resulting knowledge and actions to users and Agents. The main v2 path is connection and ingestion → pipeline transforms → curated review → ontology mapping and publication → graph, search, and executable actions. Separate branches provide document extraction, quality audits, and governed conversational Agent execution. Pipeline execution and step dispatch are supported by sampled source; several other capability links are documented in the README, while unsampled implementation details are not inferred.
Main entry points: dataset.py Datasets · publication Ontology Lifecycle · pipeline_run.py Pipeline Runner · PipelineBuilderPage.tsx Pipeline Builder
Components 20
Data Pipelines
- Pipeline Builder
frontend/src/pages/pipelines/builder/PipelineBuilderPage.tsx - Pipeline API
backend/app/routers/v2/pipelines.py - Pipeline Runner
backend/app/tasks/v2/pipeline_run.py - Transform Engine
backend/app/services/v2/pipeline/engine.py - Transform Steps
backend/app/services/v2/pipeline/steps - Datasets
backend/app/models/v2/dataset.py - Curated Review
backend/app/services/v2/curated/review_service.py - Data Connections
backend/app/services/connection
Ontology Platform
- Ontology Mapping
backend/app/services/v2/mapping/mapping_service.py - Ontology Lifecycle
backend/app/services/publication - Knowledge Graph
backend/app/services/v2/graph - Ontology Search
backend/app/services/v2/graph - Quality Audit
backend/app/routers/audit.py
Agent Runtime
- Agent Execution
backend/app/services/runtime - Ontology Tools
backend/app/services/ontology_tools.py
Platform Services
- LLM Extraction
backend/app/routers/extraction.py - Prompt Management
backend/app/routers/prompts.py - Data Export
backend/app/services/export_service.py - User Access
backend/app/services/auth_service.py
Other components
- User
Connections 20
- User → Pipeline Builder: builds pipelines
README.md - Pipeline Builder → Pipeline API: runs pipeline
- Pipeline API → Pipeline Runner: dispatches run
backend/app/routers/v2/pipelines.py - Pipeline Runner → Transform Engine: executes transforms
backend/app/tasks/v2/pipeline_run.py - Transform Engine → Transform Steps: runs steps
backend/app/services/v2/pipeline/engine.py - Data Connections → Datasets: ingests data
README.md - Pipeline Runner → Datasets: writes curated data
backend/app/tasks/v2/pipeline_run.py - Datasets → Curated Review: undergoes review
README.md - Datasets → Ontology Mapping: maps datasets
README.md - Ontology Mapping → Ontology Lifecycle: discovers rules
README.md - Ontology Lifecycle → Knowledge Graph: publishes knowledge
README.md - Ontology Lifecycle → Ontology Search: enables search
README.md - User → Quality Audit: requests audit
README.md - Quality Audit → Ontology Lifecycle: reviews quality
README.md - User → Agent Execution: starts conversation
README.md - Agent Execution → Ontology Tools: uses ontology tools
README.md - User → LLM Extraction: uploads documents
README.md - LLM Extraction → Prompt Management: uses prompts
README.md - User → Data Export: exports knowledge
README.md - User → User Access: authenticates
README.md