Automatically assign a label to unlabelled data
To ease the manual tagging process in the case of behaviors of interest as opposed to all other behaviors, and still include the data for these other behaviors e.g. in a training dataset, we would like to automatically tag all the data as some fallback state/behavior, and manually tag the behaviors of interest only.
To do so, we need a command like:
larvatagger.jl import <input-file-path> <output-file-name> --default-label=<fallback-state>
This should support both raw track data and manually (and partially) labelled data, and generate a label file with all the tracks and time steps defined, possibly with the specified default label as behavior tag.