We could make a unified API call that hides most of the details of the writer. That way, scientists would only need to pass in a CSV and the column names for the track, centroids, bounds, etc. data and a list of FeatureInfo objects for their features, plus any metadata.
One potential wrinkle would be needing to handle different kinds of image transformations, since different projects want to flatten their 3D z-stack images differently. This could be handled through a callback, maybe, that takes a URL/filepath and returns the 2D segmentations.
This would make it much easier to start working with colorizer data!
We could make a unified API call that hides most of the details of the writer. That way, scientists would only need to pass in a CSV and the column names for the track, centroids, bounds, etc. data and a list of FeatureInfo objects for their features, plus any metadata.
One potential wrinkle would be needing to handle different kinds of image transformations, since different projects want to flatten their 3D z-stack images differently. This could be handled through a callback, maybe, that takes a URL/filepath and returns the 2D segmentations.
This would make it much easier to start working with colorizer data!