The pipeline currently builds the graph for RAG when the application is started, and while this doesn't take too long currently because of the relatively small number of repositories we're pulling from, eventually we'll need to use a pre-built db that is updated independently. So the pipeline and graph RAG agent needs to be adjusted to allow plugging in the pre-built db.
The pipeline currently builds the graph for RAG when the application is started, and while this doesn't take too long currently because of the relatively small number of repositories we're pulling from, eventually we'll need to use a pre-built db that is updated independently. So the pipeline and graph RAG agent needs to be adjusted to allow plugging in the pre-built db.