Relationship Graph

Relationship Graph

Relationship Graph Module

The graph module models relationships between your business entities using a dynamic ontology and D3.js visualisations.

Main models

  • OntologieType — Entity types (person, company, account, project, etc.)
  • Entite — Instance of a type with dynamic properties (JSONField)
  • Relation — Typed link between two entities (client_of, employee_of, etc.)
  • Anomalie — Automatically detected graph anomaly
  • RequeteSauvegardee — Saved search queries

Features

  • Dynamic ontology — Create your own entity and relationship types
  • D3.js visualisation — Interactive graph with forces, zoom and filters
  • Semantic search — Similarity search via embeddings (pgvector)
  • Anomaly detection — Automatic identification of inconsistencies:
  • Missing relationships
  • Contradictory data
  • Isolated entities
  • Potential duplicates
  • Timeline — Chronological view of entity-related events
  • CSV import — Bulk import of entities and relationships
  • Suggestions — Relationship recommendations based on embeddings

API

Endpoint Description
/api/v1/graph-types/ Ontology types (CRUD)
/api/v1/graph-entites/ Entities (CRUD + search)
/api/v1/graph-relations/ Relations (CRUD)
/api/v1/graph-anomalies/ Detected anomalies
/api/v1/graph/recherche/ Semantic search
/api/v1/graph/stats/ Graph statistics
/api/v1/graph/import/csv/ CSV import

MCP Server

The graph is accessible via the MCP server (Model Context Protocol) at the /mcp/ endpoint for integration with AI clients (Claude Desktop, VS Code, etc.).

Interactions

  • Core — Entities can be linked to business objects
  • Accounting — Relationships between accounts, entries and third parties
  • Analytics — Graph visualisations in dashboards
  • AI Chat — Query the graph in natural language
On this page