How orange-opensource/tripleflow works
Tripleflow is a web application for semi-supervised knowledge-graph population from unstructured documents. Users submit text or files to configured extractors, then review and validate extracted triples, inspect graph data, and publish approved triples to configured targets. The API also manages extractor and target configurations, file provenance, and audit history. The frontend and API source indexes support the main routing and service boundaries; some API-to-frontend and external-service links are shown as unverified because their runtime wiring was not excerpted.
Main entry points: main.py FastAPI application · App.vue Vue application · extraction.py Extraction routes · sinks_service.py Target configuration
Components 27
Frontend
- Vue application
tripleflow-front/src/App.vue - Extraction workspace
tripleflow-front/src/pages/HomePage.vue - Workflow view
tripleflow-front/src/components/organisms/Workflow.vue - Extraction state
tripleflow-front/src/composables/UseExtractionState.js - Extraction API client
tripleflow-front/src/composables/extraction/api.js - Validation workspace
tripleflow-front/src/pages/ValidationPage.vue - Graph browser
tripleflow-front/src/pages/GraphPage.vue - Extractor management
tripleflow-front/src/pages/ExtractorsPage.vue - Target management
tripleflow-front/src/pages/TargetsPage.vue
Extraction
- Extraction routes
tripleflow-api/app/routes/extraction.py - Document parsing
tripleflow-api/app/services/docling_service.py - Text chunking
tripleflow-api/app/services/chunking_service.py - Extractor execution
tripleflow-api/app/services/extraction_service.py
Configuration
- Extractor configuration
tripleflow-api/app/services/extractors_service.py - Extractor schemas
tripleflow-api/app/schemas/extractors.py - Target configuration
tripleflow-api/app/services/sinks_service.py - Target schemas
tripleflow-api/app/schemas/sinks.py
Review and publishing
- Triple and file routes
tripleflow-api/app/routes/triples.py - File routes
tripleflow-api/app/routes/files.py - Entity search
tripleflow-api/app/routes/search.py - Publication routes
tripleflow-api/app/routes/sinks.py - Triple and file data
tripleflow-api/app/services/triple_db_service.py
Persistence
- MongoDB collections
tripleflow-api/app/db.py - FastAPI application
tripleflow-api/app/main.py
Other components
- Reviewer
- Extractor endpoint
- Knowledge graph target
Connections 40
- Reviewer → Vue application: uses
tripleflow-front/src/App.vue - Vue application → Extraction workspace: routes to
tripleflow-front/src/App.vue - Vue application → Validation workspace: routes to
tripleflow-front/src/App.vue - Vue application → Graph browser: routes to
tripleflow-front/src/App.vue - Vue application → Extractor management: routes to
tripleflow-front/src/App.vue - Vue application → Target management: routes to
tripleflow-front/src/App.vue - Extraction workspace → Workflow view: renders
tripleflow-front/src/pages/HomePage.vue - Extraction workspace → Extraction state: uses
tripleflow-front/src/pages/HomePage.vue - Workflow view → Extraction state: uses
tripleflow-front/src/components/organisms/Workflow.vue - Extraction state → Extraction API client: requests extraction
tripleflow-front/src/composables/UseExtractionState.js - Extraction API client → Extraction routes: calls API
- FastAPI application → Extraction routes: registers route
tripleflow-api/app/main.py - FastAPI application → Triple and file routes: registers route
tripleflow-api/app/main.py - FastAPI application → File routes: registers route
tripleflow-api/app/main.py - FastAPI application → Entity search: registers route
tripleflow-api/app/main.py - FastAPI application → Publication routes: registers route
tripleflow-api/app/main.py - FastAPI application → MongoDB collections: ensures indexes
tripleflow-api/app/main.py - Extraction routes → Document parsing: uses parser
tripleflow-api/app/routes/extraction.py - Extraction routes → Text chunking: uses chunking
tripleflow-api/app/routes/extraction.py - Extraction routes → Extractor execution: runs extraction
tripleflow-api/app/routes/extraction.py - Extraction routes → Triple and file data: stores results
tripleflow-api/app/routes/extraction.py - Extractor execution → Extractor configuration: resolves extractor
tripleflow-api/app/services/extraction_service.py - Extractor execution → Extractor endpoint: sends text
- Extractor configuration → Extractor schemas: validates config
tripleflow-api/app/services/extractors_service.py - Extractor configuration → MongoDB collections: reads and writes
tripleflow-api/app/services/extractors_service.py - Extractor management → Extractor configuration: manages via API
- Target configuration → Target schemas: validates config
tripleflow-api/app/services/sinks_service.py - Target configuration → MongoDB collections: reads and writes
tripleflow-api/app/services/sinks_service.py - Target configuration → Knowledge graph target: publishes triples
- Target management → Target configuration: manages via API
- Validation workspace → Publication routes: publishes via API
- Validation workspace → Triple and file routes: submits review
- Validation workspace → Entity search: searches entities
- Graph browser → Triple and file routes: loads graph data
- Triple and file routes → Triple and file data: reads and updates
tripleflow-api/app/routes/triples.py - File routes → Triple and file data: manages files
tripleflow-api/app/routes/files.py - Entity search → Extractor configuration: uses extractor config
tripleflow-api/app/routes/search.py - Publication routes → Target configuration: publishes and manages
tripleflow-api/app/routes/sinks.py - Publication routes → Triple and file data: loads triples
tripleflow-api/app/routes/sinks.py - Triple and file data → MongoDB collections: reads and writes
tripleflow-api/app/services/triple_db_service.py