[FEATURE] Neo4J RETRIEVER

[retail-amelis]

Integration of Neo4J retriever, as done for REDIS, PINECONE, etc. This provides the opportunity to use graph databases as data retrievers.

https://python.langchain.com/docs/integrations/providers/neo4j https://python.langchain.com/docs/integrations/vectorstores/neo4jvector

[gkorland]

That can be a great addition but please make sure it's a generic Graph API, so it can support other Graph databases but Neo4J (like FalkorDB)

[themeaningofmeaning]

Yes, integration with graph API would open doors.

[qdrddr]

GRAG = Graph RAG great article explaining how Grapg Knowledge DB can improve quality of RAG

https://medium.com/towards-data-science/text-to-knowledge-graph-made-easy-with-graph-maker-f3f890c0dbe8

[abdarwish23]

Dears any progress in this feature, it is an added value ✌️

[qdrddr]

Hope Knowledge Graph Data Bases such as Neo4j, FalkorDB, Nebula Graph can be integrated with Flowise LangChain has it for a while now: https://python.langchain.com/v0.2/docs/integrations/graphs/neo4j_cypher/

[qdrddr]

Note, Neo4J can work as both: Vector Database, and Knowledge Graph Data Bases that is using Cypher queries (similar to SQL but for KG).

[qdrddr]

https://docs.llamaindex.ai/en/stable/api_reference/storage/graph_stores/neo4j/

[qdrddr]

Feature Description:

The integration of Neo4j into Flowise will bring support for powerful knowledge graph database retrieval capabilities, extending Flowise's versatility in handling diverse data storage solutions. Neo4j, Bebula Graph, and FalcorDB are leading Knowledge Graph Databases, that can serve as both a vector database and a knowledge graph using Cypher queries. This feature will mirror existing integrations like Redis and Pinecone, empowering users to efficiently leverage graph structures for retrieval tasks.

https://python.langchain.com/v0.2/docs/integrations/graphs/neo4j_cypher/ https://python.langchain.com/docs/integrations/providers/neo4j https://python.langchain.com/docs/integrations/vectorstores/neo4jvector

Motivation

Why This is Essential:

Graph Databases as Data Retrievers: The ability to retrieve data from knowledge graph databases opens new avenues for complex, interconnected data queries. KGDB structure allows for highly sophisticated data relationships, offering more flexible and insightful retrieval than traditional databases.

Unlocking Knowledge Graph Potential: Knowledge Graph Databases such as Neo4j, FalkorDB, and Nebula Graph are becoming crucial in industries where understanding relationships between data points is key—such as recommendation systems, fraud detection, and semantic search. Integrating these with Flowise will allow users to query graph data with ease and precision, using graph-based algorithms.

Dual Functionality with Vector Databases: Neo4j, Bebula Graph, and FalnorDB can also act as a vector database, adding to Flowise's capacity to handle vectorized data alongside traditional knowledge graph structures. This flexibility allows users to seamlessly move between vector searches (e.g., for similarity) and more intricate relationship-based queries via Cypher, a powerful query language.

Aligning with LangChain: LangChain, a recognized leader in the space, has already integrated Neo4j for both vector store and graph database functionalities. Aligning Flowise with LangChain ensures that Flowise remains competitive and compatible with the broader AI and machine learning ecosystem. Users will appreciate the continuity of experience and the shared ecosystem between these tools.

Extending Flowise’s Reach: The integration will expand Flowise's use cases, making it more appealing to sectors that rely on graph-based data representation and advanced knowledge modeling. Neo4j’s integration would create a path for future inclusion of other graph databases like FalkorDB and Nebula Graph, thereby creating a wider, more flexible infrastructure for Flowise users.

Business Impact:

By enabling Neo4j, Nebula Graph, FalcorDB and other Knowledge Graph Databases, Flowise will become more attractive to industries focused on complex data interrelationships, ensuring broader adoption and deeper penetration into sectors like fintech, healthcare, cybersecurity, and recommendation engines.

[qdrddr]

Your competitor RAGflow has already added GraphRAG support

https://github.com/infiniflow/ragflow/tree/v0.9.0

[jbeltran73-2]

Please add it.

[qdrddr]

Here is why knowledge graphs are strategically important for AI & RAG. Please vote for this task.