refactor complex actions

[franzlst]

The issue is about to complex actions that take place and don't have a clear interface to be performed.
For now complex actions are, state_type_change, group_states, ungroup_state, substitute_state and
"paste". All this actions have separate code fragments or functions in multiple locations in core, models, gui.helpers and graphical editors. For every action the interface is different. In the last RAFCON meeting we discussed a solution which is now the starting point to re-factor this interface

Steps:

• encapsulate complex action by "start" and "end" signal generated on the state model which holed origin, target, affected models, action and before/after boolean -> this need to be integrated/substitute into modification history and graphical editors (and in future maybe in other components to improve performance)
• all complex actions will be separated and moved to helper modules
  • from the core elements
  • and model elements
  • the meta data re-size/adaptations in the graphical editor operations
• the complex actions will also use un-register and register observer methods to reduce notification load in the recursive notification pattern of the model (the start and end signals have to be notified recursively) and avoid unnecessary reactions (not sure if this is still the best way 2017-08)

Rough example from the meeting -> Action: Group

1. User: Triggers group for selected states
2. Menu entry calls state_helper.group
3. state_helper.group: emits signal in parent state: ActionSignal(action="group", origin="state_helper", parent, affected_models=selected states, after=False)
   (3.1 disable parent_m observer on core)
4. signal is propagated upwards to state_machine_m
5. graphical editor receives signal, creates temp list of afftected states, deactivates observers
6. state_helper.group: removes selected state_m, creates new container_m, adds affected as child_state_m, adapts meta data
7. state_helper.group: new container_m is added to parent_m's expected_future_models
8. state_helper.group: container_m.state is added to parent_m.state state's
9. container_m gets notification about new core state, finds and uses expected_future_model
   (9.1 enable parent_m observer on core)
10. state_helper.group: emits signal in parent state: ActionSignal(action="group", origin="state_helper", parent, affected_models=container_m, after=True)
11. graphical editor receives signal, create state_v for container_m and remaps state_v of temp list of afftected states, activates observers

Originally created by @Rbelder at 2017-03-28 09:40:34+00:00 (moved from RMC internal repository)