Writing additional columns to output tables

[dhensle]

The user cannot add columns that are calculated in the pre-processor to the final output tables. (Some models do have optional "annotate" steps to populate columns, but not all. They also do no have access to the variables calculated in the pre-processor).

An example case for this would be calculating value of time bins in trip mode choice depending on attributes of the trip, but not being able to segment output trip matrices by the those value of time bins because they are not included in the final trip table.

Solution would look something like allowing the user to specify additional columns from the chooser table in the yaml file for that submodel to keep in the final output table.