-
|
I am experimenting to build text-to-graphQL query generator based in the GraphQL schema. My schema.graphql is too big to fix in to context. So exploring right GraphRAG product, the reason I am thing helix-db might be a choice for this knowledge base is, it support graph+vector DB can you guys give any feedback? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
Hi Sumanth, we've obviously discussed this privately but thought I'd reply here now for the rest of the community. This looks like a good use case. You can embed the types/inputs in your schema and store the properties as edges that point to the other types/inputs that they represent. You can then perform similarity searches against the vectors and then traverse the graph to retrieve all the type dependencies. You then feed this context to the LLM and it has an exact overview of all the types and inputs necessary to answer your particular question. |
Beta Was this translation helpful? Give feedback.
Hi Sumanth, we've obviously discussed this privately but thought I'd reply here now for the rest of the community.
This looks like a good use case. You can embed the types/inputs in your schema and store the properties as edges that point to the other types/inputs that they represent.
You can then perform similarity searches against the vectors and then traverse the graph to retrieve all the type dependencies.
You then feed this context to the LLM and it has an exact overview of all the types and inputs necessary to answer your particular question.