For the cluster visualizations, when a dot is clicked, it should fetch the associated article/chunk from the server and show it. Probably slide in a modal from the side.
This requires that we have an ID or something so that when we request it from the server it knows how to get it. For articles that's probably not that hard, but for other chunks (e.g., paragraphs), it means that whenever we do the chunk splitting we need to assign IDs to each AND keep them around so we can search for them later.
For particularly large datasets you'd want to build an index on those IDs or else the server is going to spend all its time just looking the IDs up, but that's an optimization we can think about later.
For the cluster visualizations, when a dot is clicked, it should fetch the associated article/chunk from the server and show it. Probably slide in a modal from the side.
This requires that we have an ID or something so that when we request it from the server it knows how to get it. For articles that's probably not that hard, but for other chunks (e.g., paragraphs), it means that whenever we do the chunk splitting we need to assign IDs to each AND keep them around so we can search for them later.
For particularly large datasets you'd want to build an index on those IDs or else the server is going to spend all its time just looking the IDs up, but that's an optimization we can think about later.