Hi everyone! Is this channel active for Nebula Graph users to post questions? If not, would you help direct us where we can ask questions? Here is our question: we are building GraphRAG using NebulaGraph database on the customer account and activity data. To simplify the context here, we have 6 million customer account IDs associated with 6 industry verticals. We have tried two designs. 1) Create customer_account_ID as node, industry_vertical as node, and their mapping relationship as edge. In this design, one industry vertical node is connected to 1 million customer_account_ID nodes on average. 2) Create customer_account as node with customer_account_id and industry_vertical as its properties/attributes, and then create a tag index on customer_account node. For these two designs, we compared their query performance as a part of whole graph with activity nodes. When asking questions involving the industry vertical, the query on graphRAG with the 1st design is extremely slow, while the query on graphRAG with the 2nd design is much much faster. For the query difference, our understanding is that by having industry_vertical as nodes, it increases the edges in the graph by millions, which causes the searching extremely slow. Is this correct? And what are the best practices on what should be created as nodes vs. properties in general? Thank you very much.