Overview & Result
neo4jev, created by Neo4j expert Michael Hunger, integrates TypeSafe's Jev model into graph database exploration. As an agent walks through connected entity nodes, Jev evaluates outgoing relationships and property previews, classifying which edge is most promising to traverse next toward resolving the user's inquiry.
How Jev fits in the loop
- Queries adjacent edges and neighboring node properties from Neo4j via Cypher.
- Presents candidate relationship edges to Jev with the global search goal.
- Follows the selected path iteratively until destination nodes or insights are reached.
How to reproduce
- Clone https://github.com/jexp/neo4jev and connect to a running Neo4j instance.
- Set your research query and starting node in the graph.
- Run the explorer script and trace Jev choosing candidate relationships to traverse to target answers.
Why this build matters
Solves combinatorial path explosion in large knowledge graphs by replacing blind brute-force graph search with intelligent edge classification.
Limitations
- Requires an active Neo4j database instance.
- Optimal performance requires indexed property graphs.