Replies: 1 comment 1 reply
-
|
The knowledge graph in RAGFlow is stored in the document store backend, which can be Elasticsearch, OpenSearch, or InfinityDB, depending on your configuration—not in MySQL, Redis, or MinIO. The graph data is saved as serialized JSON (nodes and edges) and indexed by knowledge base and tenant ID. You can access and manage it via the API or the web UI, but it’s not stored as files inside Docker containers or in MinIO buckets. RAGFlow does not support connecting to external graph databases like Neo4j; all knowledge graph data stays in the configured document store backend. If you need to back up or migrate the knowledge graph, export/import the relevant indices from your document store (e.g., using elasticdump for Elasticsearch) rather than from MinIO or MySQL [source] [source] [source]. To reply, just mention @dosu. How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
After generating a knowledge graph, where is the graph stored at? i can see there are many component running in docker such as mysql, redis and minio. and can use external graph database to connect RAGFlow? @dosu
Beta Was this translation helpful? Give feedback.
All reactions