DGLink introduces semantic interoperability within data portals through automated metadata extraction and knowledge graph construction. DGLink connects data semantically within a data portal and maps data to external knowledge to enable knowledge-driven data interpretation and discovery.
scripts- Scripts for recreating KG and artifactsexamples- Example code, so far just the original proof of concept not using the package.dglink/core- Core DGLink code, applicable for all portals.dglink/portals- Portal specific code.dglink/applications- Applications of KGs constructed with DGLink.dglink/applications/semantic_search- bring up web UI for semantic searching on the KG (see instructions bellow)dglink/applications/project_similarity- Leverage KG embedding methods for accessing project similarity (need to install additional dependencies with)uv pip install ".[graph_embedding]"
dglink/resources- data and container for brining up the Neo4j instance.
- Go into the semantic search directory with
cd dglink/applications/semantic_search - Bring up the service (and build images if required with)
docker-compose up --build - Connect to the services.
- semantic search UI http://localhost:5000/
- neo4j browser http://localhost:7474/browser/