The existing MLDP DataSet Annotation mechanism could be used to annotate individual data points (e.g., as suspect or invalid) but is probably not very efficient for navigating the annotations. We could develop a new API optimized for annotating an individual data point and navigating those annotations.
- maybe annotateDataSet() is too broad for this, and we should add something like annotateData()
show annotations within the bucket that they target, don't add them to the main dataset annotations collection