I am trying to import to Fipy a 3D mesh previously generated with Gmsh using the Gmsh3D function, like this:
mesh = Gmsh3D(join(output,'case_1.msh'),communicator=serialComm)
I get the same error as other similar posts, but the proposed solutions do not seem to work for me:
GmshException: Gmsh hasn't produced any cells! Check your Gmsh code.
Gmsh output:
Info : Running 'gmsh C:\Users\pcnou\AppData\Local\Temp\tmppouz68km.geo -3 -nopopup -format msh2 -o C:\Users\pcnou\AppData\Local\Temp\tmp5es89ie5.msh' [Gmsh 4.6.0, 1 node, max. 1 thread]
Info : Started on Wed May 05 10:22:49 2021
Info : Reading 'C:\Users\pcnou\AppData\Local\Temp\tmppouz68km.geo'...
Info : Done reading 'C:\Users\pcnou\AppData\Local\Temp\tmppouz68km.geo'
Info : Meshing 1D...
Info : Done meshing 1D (Wall 0s, CPU 0s)
Info : Meshing 2D...
Info : Done meshing 2D (Wall 0s, CPU 0s)
Info : Meshing 3D...
Info : Done meshing 3D (Wall 0s, CPU 0s)
Info : 0 nodes 0 elements
Info : Writing 'C:\Users\pcnou\AppData\Local\Temp\tmp5es89ie5.msh'...
Info : Done writing 'C:\Users\pcnou\AppData\Local\Temp\tmp5es89ie5.msh'
Info : Stopped on Wed May 05 10:22:49 2021 (From start: Wall 0.00999999s, CPU 0.03125s)
I'm working on Win10 with Fipy 3.4.2.1, Python 3.7.9 and Gmsh 4.8.4.
I tried to convert the .msh file into a .msh2 file but then the following error appears:
IndexError Traceback (most recent call last)
<ipython-input-47-38097a945970> in <module>
----> 1 mesh = Gmsh3D(join(output,'case_1.msh2'),communicator=serialComm) # Doesn't work
~\Anaconda3\envs\Ansys\Lib\site-packages\fipy\meshes\gmshMesh.py in __init__(self, arg, communicator,
overlap, background)
1991 self.cellGlobalIDs,
1992 self.gCellGlobalIDs,
-> 1993 self._orderedCellVertexIDs_data) = self.mshFile.read()
1994
1995 self.mshFile.close()
~\Anaconda3\envs\Ansys\Lib\site-packages\fipy\meshes\gmshMesh.py in read(self)
815 parprint("Recovering coords.")
816 parprint("numcells %d" % numCellsTotal)
--> 817 vertexCoords, vertIDtoIdx = self._vertexCoordsAndMap(cellsToGmshVerts)
818
819 # translate Gmsh IDs to `vertexCoord` indices
~\Anaconda3\envs\Ansys\Lib\site-packages\fipy\meshes\gmshMesh.py in _vertexCoordsAndMap(self,
cellsToGmshVerts)
1008 allVerts = nx.unique(nx.array(allVerts, dtype=nx.INT_DTYPE)) # remove dups
1009 allVerts = nx.sort(allVerts)
-> 1010 maxVertIdx = allVerts[-1] + 1 # add one to offset zero
1011 vertGIDtoIdx = nx.ones(maxVertIdx, 'l') * -1 # gmsh ID -> vertexCoords idx
1012 vertexCoords = nx.empty((len(allVerts), self.coordDimensions))
IndexError: index -1 is out of bounds for axis 0 with size 0
And the same happens when I use the .geo
file. Is it a problem of the mesh itself or am I doing something wrong with Gmsh3D? Any help will be highly appreciated.
I first tried with this mesh link that is the one that I expect to work with, but I'm not sure the mesh itself does not have any issues. So I then tried with the following .msh file: link and the corresponding .geo file: link, but did not work neither.
Thank you.
This line of output
Info : Running 'gmsh C:\Users\pcnou\AppData\Local\Temp\tmppouz68km.geo -3 -nopopup -format msh2 -o C:\Users\pcnou\AppData\Local\Temp\tmp5es89ie5.msh' [Gmsh 4.6.0, 1 node, max. 1 thread]
means that FiPy is interpreting the result of join(output,'case_1.msh')
as a GEO script, or, more precisely, that it can't find a file in that location, so it assumes that string must be a GEO script. Since it's a file path and not a GEO script, things go south from there.
I don't know why join(output,'case_1.msh')
isn't seen as a path to an existing file for you. Is output
an absolute path or relative? I'd suggest checking against the working directory you launched your script from.
case_1_amplazer_pos10001.geo
you provided is not a Gmsh geometry script, so neither FiPy nor Gmsh can read it.
test.msh
and case_1.msh
are both version 4.1 MSH files. FiPy cannot read these.
How did you convert to MSH2
format?
I used
gmsh -3 -format msh2 case_1.msh -o case_1.msh2
and FiPy successfully imported it.