Shifted the rest of the gui_bieri.execution.calc_modelfree() code into self.execute().
All the automatic model-free protocol, including execution, is now within the auto_model_free
module. This eliminates the last relics of the duplication of the full_analysis.py sample script.
