Endpoint to retrieve samples for a given slice
- client requests samples for a set of features and a range specified for each feature
- return all samples within those bounds or downsample on the fly if some constant MAX_SAMPLE_COUNT is exceeded
Challenge: We probably need some index structure to filter was enough (instead of running over the whole data set).
Alternatively: Render the scatterplot server side and only send the result to the client.
Thoughts?