add PoissonEnergy

nifty4/library/__init__.py

@@ -3,4 +3,5 @@ from .wiener_filter_curvature import WienerFilterCurvature
 from .noise_energy import NoiseEnergy
 from .nonlinear_power_energy import NonlinearPowerEnergy
 from .nonlinear_wiener_filter_energy import NonlinearWienerFilterEnergy
+from .poisson_energy import PoissonEnergy
 from .nonlinearities import Exponential, Linear, Tanh, PositiveTanh

nifty4/library/nonlinear_power_energy.py

@@ -24,11 +24,11 @@ from ..utilities import memo
 
 
 def _LinearizedPowerResponse(Instrument, nonlinearity, ht, Distributor, tau,
-                            xi):
-     power = exp(0.5*tau)
-     position = ht(Distributor(power)*xi)
-     linearization = nonlinearity.derivative(position)
-     return 0.5*Instrument*linearization*ht*xi*Distributor*power
+                             xi):
+    power = exp(0.5*tau)
+    position = ht(Distributor(power)*xi)
+    linearization = nonlinearity.derivative(position)
+    return 0.5*Instrument*linearization*ht*xi*Distributor*power
 
 
 class NonlinearPowerEnergy(Energy):

nifty4/library/wiener_filter_energy.py

@@ -20,7 +20,7 @@ from ..minimization.quadratic_energy import QuadraticEnergy
 from .wiener_filter_curvature import WienerFilterCurvature
 
 
-def WienerFilterEnergy(position, d, R, N, S, inverter):
+def WienerFilterEnergy(position, d, R, N, S, inverter=None):
     """The Energy for the Wiener filter.
 
     It covers the case of linear measurement with
@@ -39,8 +39,9 @@ def WienerFilterEnergy(position, d, R, N, S, inverter):
         The noise covariance in data space.
     S : EndomorphicOperator
         The prior signal covariance in harmonic space.
-    inverter : Minimizer
+    inverter : Minimizer, optional
         the minimization strategy to use for operator inversion
+        If None, the energy object will not support curvature computation.
     """
     op = WienerFilterCurvature(R, N, S, inverter)
     vec = R.adjoint_times(N.inverse_times(d))