surface addAnnotationColumns in new_dataset

[allaway]

Sometimes we need to disable this (addAnnotationColumns = F) because synapse has conflicting annotation types, leading to duplicate cols in the schema (e.g. "Duplicate column name: 'age'")

Probably worth surfacing annotations at the same time. :)