Accessing Fields for DC 2.5D

In DC 2.5D, trying to access any other field than “Phi” raises the current error (btw, those other fields appear in the documentation and the 2.5Dcode, they are just not fully implemented yet):

Side note: Accessing the ‘Phi’ field is working. However, I have a question about its shape. See example below. The second dimension is the number of spatial frequencies but I am unsure what is the first dimension: it is neither the size of the whole mesh nor the size of the active cells…

Extra information: here is the scripts I ended up using to compute and visualize the field. The em-apps are using a similar approach. So improving on the Fields object will likely improve the em-apps.

# Plot vector field of one quadrupole for mtrue
Mx = meshList20[0].gridCC
rx = DC.Rx.Pole_ky(Mx)
src = DC.Src.Dipole([rx],dcsurvey20[0].srcList[0].loc[0],dcsurvey20[0].srcList[0].loc[1])
dcsurvey = DC.Survey_ky([src])
dcprob = DC.Problem2D_N(meshList20[0],sigmaMap=mapping20[0],Solver=PardisoSolver)
dcprob.pair(dcsurvey)

mtrue = 1./res20[0]*1000
phipotential = dcsurvey.dpred(mtrue[actind_air20[0]])
efield = -meshList20[0].cellGrad * phipotential
efieldx = -meshList20[0].cellGradx * phipotential
efieldy = -meshList20[0].cellGrady * phipotential

jfield = dcprob.MfRhoI * efield

from scipy.interpolate import LinearNDInterpolator, NearestNDInterpolator
import matplotlib

tensormesh = Mesh.TensorMesh(meshList20[0].h)
tensormesh.x0 = meshList20[0].x0
indcore, tensormesh = Utils.ExtractCoreMesh(np.c_[[-3,-1],[3,1]],tensormesh)

InterpolateEField_x = LinearNDInterpolator(meshList20[0].gridEy,efieldx)
ExField_tensor = InterpolateEField_x(tensormesh.gridEy)
InterpolateEField_y = LinearNDInterpolator(meshList20[0].gridEx,efieldy)
EyField_tensor = InterpolateEField_y(tensormesh.gridEx)
EField_tensor = np.r_[ExField_tensor,EyField_tensor]

InterpolateJField_x = NearestNDInterpolator(meshList20[0].gridEy,jfield[meshList20[0].gridEx.shape[0]:])
JxField_tensor = InterpolateField_x(tensormesh.gridEy)
InterpolateJField_y = NearestNDInterpolator(meshList20[0].gridEx,jfield[:meshList20[0].gridEx.shape[0]])
JyField_tensor = InterpolateField_y(tensormesh.gridEx)
JField_tensor = np.r_[JxField_tensor,JyField_tensor]

tensormesh.plotImage(EField_tensor,view='vec',vType='F',pcolorOpts={'norm':matplotlib.colors.LogNorm()})

[dccowan]

@thast @sgkang I believe I found the problem. I match a patch. See PR #868

[jcapriot]

This has now been taken care of in 0.14.1. You can directly access the phi and other solutions from the 2.5D fields object (it will take care of the transformation object behind the scenes).