Two types should be implemented: - [ ] Compare FDATA created by NMTRAN to Pharmpy dataset - [ ] Read models with IGNORE, write (or update_source is enough?), then reread and parse ignores
Activity