I've been looking at the "Example Docs" functionality in Graph and I had a few questions about it. Currently, when you click "Example Docs", it simply takes the current selection of nodes, and puts them in a bool query and sends that off to the ES _search API endpoint. I was curious as to how one should go about creating a more contextual search for relevant documents. As an example problem, the same node may show up in several different graphs, but hitting "Example Docs" would always give back the same docs.
This touches upon an issue that's come up as I've gotten to use graph more: graph is really good at telling you "what" connections exist in your documents, but it's not so great at telling you "why." We've been experimenting with using graph to visualize our documents (which are enriched and indexed news sources), but when we see interesting connections, graph doesn't natively provide a way to view the documents that created these connections. It very well could be that our use case doesn't it lend itself well to graph, but I'd imagine users of graph would benefit from being able to see what documents were used to create the nodes and edges.
Thanks again for all your time and effort!