SaveOutputEveryIteration Directive errors

Can no longer use the SaveOutputEveryIteration directive on the the dcip/dev branch. I get the following error when calling this function using Tikhonov regularization on a treemesh.

SimPEG.InvProblem is setting bfgsH0 to the inverse of the eval2Deriv.
    ***Done using same Solver and solverOpts as the problem***
Use Tikhonov
SimPEG.SaveOutputEveryIteration will save your inversion progress as: '###-InversionModel-2017-12-20-12-06.txt'
SimPEG.SaveModelEveryIteration will save your models as: '###-InversionModel-2017-12-20-12-06.npy'
model has any nan: 0
============================ Inexact Gauss Newton ============================
  #     beta     phi_d     phi_m       f      |proj(x-g)-x|  LS    Comment   
-----------------------------------------------------------------------------
x0 has any nan: 0
   0  1.22e-01  1.64e+05  7.60e+01  1.64e+05    2.35e+05      0              
Traceback (most recent call last):
  File "InvScript.py", line 225, in <module>
    mopt = inv.run(m0)
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Utils/CounterUtils.py", line 99, in wrapper
    out = f(self, *args, **kwargs)
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Inversion.py", line 67, in run
    self.m = self.opt.minimize(self.invProb.evalFunction, self.invProb.model)
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Utils/CounterUtils.py", line 99, in wrapper
    out = f(self, *args, **kwargs)
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Optimization.py", line 291, in minimize
    self.doEndIteration(xt)
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Utils/CounterUtils.py", line 88, in wrapper
    out = f(self, *args, **kwargs)
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Utils/codeutils.py", line 144, in wrapper
    return f(self,*args,**kwargs)
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Optimization.py", line 564, in doEndIteration
    self.callback(xt)
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Inversion.py", line 73, in _optCallback
    self.directiveList.call('endIter')
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Directives.py", line 170, in call
    getattr(r, ruleType)()
  File "/tera_raid/mmitchel/Programs/simpeg/SimPEG/Directives.py", line 351, in endIter
    self.reg.objfcts[0](self.invProb.model) * self.reg.alpha_s
AttributeError: 'ComboObjectiveFunction' object has no attribute 'alpha_s'

[thast]

@fourndo: I understand we want to be able to handle several regularizers (that can be so powerful!). What I am less sure about, and not a big fan so far, is the discrepancy between the type of object when calling *.reg.

Moreover, This also create inconsistency of interaction (as reported by @micmitch ) and naming convention. For example there is also now an inconsistency for the use of the term objfcts: in one case sel.reg.objfcts is a full regularization (Simple or Tikhnov object) while in the case self.invProb.reg.objfcts it calls the individual norms or portions ( Small or SmoothDeriv …) . It seems we now define objfcts as a handle to always access the next subdivision, but I am not sure of the name anymore in this case.

So with my current understanding, Would not it be better then to also turn invProb.reg into a ComboObjectiveFunction object then to ensure consistency when calling a reg object?

[jcapriot]

I believe the underlying bug here was fixed a while ago, as this is working in the more recent versions. However the discussion on how directives being tied to specific regularizations is still relevant.