samyama-graph

关于

Project URL Category Rust Project Title A Rust graph database with built-in vector search Project Description A graph database written in Rust that runs OpenCypher queries and vector search in a single engine. Setup is one docker pull, and it ships 11 case studies built from public data - cricket, IMDB, court judgments - so you can run real queries on day one. Highlights Most projects in this space make you run two systems: a graph database for relationships and a vector store for embeddings, kept in sync by your own code. Samyama Graph runs OpenCypher traversal and HNSW vector search in one process, so a "who is connected to whom" query and a "what is similar to this" query hit the same engine and the same data. It is also easy to start, which matters more than the architecture for anyone trying it out. One docker pull from a public registry - no account, no signup, no sales call - and the casestudies directory has 11 datasets that rebuild themselves from public sources: cricket ball-by-ball records, IMDB, Indian Supreme Court judgments, drug interactions, protein pathways. Each one ships its loader and its queries, so you can go from empty database to a real answer in an evening. Apache-2.0, written in Rust, 14 graph algorithms. The benchmark page publishes its own unflattering results, including two runs of the same binary that differed by 24%. Example Code markdown // Which bowler has dismissed which batsman most often? // 36,619 nodes and 1,392,017 edges of ball-by-ball cricket data. MATCH (bowler:Player)-[d:DISMISSED]->(batsman:Player) RETURN bowler.name, batsman.name, count(d) AS times ORDER BY times DESC LIMIT 5 Screenshots or Demo Videos A one-minute demo of Claude querying six of these knowledge graphs over MCP, picking the server and writing the Cypher itself: The eleven case studies, each with its loader and queries:

收录记录

1. HelloGitHub·用户投稿

   2026年8月21日 · 1 次收录