I think the support for Gremlin database is not available. You can extract the data and put it into an SQL DB or mongo and then use MongoVwectorSearch for mongo DB or SQLTableRetrieverQueryEngine.
Or you can extract the data and create the documents run time and do query