Data-Store
Graph Store Creation & Management
Build and manage knowledge graphs for advanced relationship visualization and network analysis.
Overview
Create graph stores to transform document collections into interactive knowledge networks showing relationships between concepts and entities through AI-powered extraction and visualization.
Store Creation
- "Graph Store" selection when creating new data stores
- Name and description configuration for graph identification
- Processing parameters for automatic graph generation from documents
- Document upload via drag and drop on the left side of the edit card
- Descriptive descriptions are crucial - the AI uses store descriptions to determine when to search specific graph stores for relevant relationship information
Access Control
- Permissions: Assign different access or edit permissions per user-role
- Private stores for personal access only
Document Management
- Document ingestion processing text to extract graph elements
- AI extraction of entities and relationships from uploaded content
- File type support: PDF, DOCX, CSV, XLSX, TXT, RTF, Markdown, Excel, JSON, XML
- Automatic processing to create nodes and edges from document content
Graph Creation (automatic)
- Automatic Entity recognition using selected LLM model
- Relationship detection for semantic connections
Graph Visualization
- Interactive exploration with zoom, pan, and navigation controls and click on a node to load all connected nodes and edges into the current visualization.
- Force-directed layouts for natural clustering and organization
- Visualization options including 2D and 3D modes with interactive exploration
- Highlight options to highlight specific entity types or relationship categories
- Source attribution click on a node to inspect all document-parts where the node entity is mentioned