post installation tests

Question

post installation tests

Opened this issue a year ago · 8 comments

hello im trying to get pyleecan working on windows. im using python 3.8.10 also i cloned pyleecan repo. and i downloaded all requirements through pip.

im assuming if i run the command : python -m pytest ./Tests -m "not long_5s".
then i should be getting 0 failed tests. however im getting 24 failed tests.

this is the summary

=========================== short test summary info ===========================
FAILED Tests/Classes/test_classes.py::test_class_methods[class_dict103] - Imp...
FAILED Tests/Classes/test_classes.py::test_class_methods[class_dict104] - Imp...
FAILED Tests/Classes/test_classes.py::test_class_methods[class_dict105] - Imp...
FAILED Tests/Classes/test_classes.py::test_class_methods[class_dict135] - Imp...
FAILED Tests/Classes/test_classes.py::test_class_methods[class_dict136] - Imp...
FAILED Tests/Classes/test_classes.py::test_class_methods[class_dict137] - Imp...
FAILED Tests/Classes/test_classes.py::test_class_methods[class_dict149] - Imp...
FAILED Tests/Classes/test_classes.py::test_class_methods[class_dict151] - Imp...
FAILED Tests/Functions/test_save_load.py::test_save_load_simu[h5] - Attribute...
FAILED Tests/Functions/test_save_load.py::test_save_load_slash_h5 - Attribute...
FAILED Tests/GUI/Dialog/DMatLib/test_DMatSetup.py::TestDMatSetup::test_BH_setup
FAILED Tests/Methods/Import/test_import_mesh_unv.py::test_import_mesh_unv[unv_file0]
FAILED Tests/Methods/Import/test_import_mesh_unv.py::test_import_mesh_unv[unv_file1]
FAILED Tests/Methods/Import/test_import_mesh_unv.py::test_import_mesh_unv[unv_file2]
FAILED Tests/Methods/Import/test_import_mesh_unv.py::test_import_mesh_unv[unv_file3]
FAILED Tests/Methods/Import/test_import_mesh_unv.py::test_import_mesh_unv[unv_file4]
FAILED Tests/Methods/Mesh/test_perm_coord.py::test_perm_coord - ImportError: ...
FAILED Tests/Methods/Mesh/test_plot_contour.py::test_plot_contour_1group - Im...
FAILED Tests/Methods/Mesh/test_plot_contour.py::test_plot_contour_2group - Im...
FAILED Tests/Methods/Mesh/test_plot_mesh.py::test_plot_mesh - ImportError: ca...
FAILED Tests/Methods/Mesh/test_plot_meshsol.py::test_plot_glyph - ImportError...
FAILED Tests/Methods/Mesh/test_plot_meshsol.py::test_plot_deflection - Import...
FAILED Tests/Methods/Mesh/Mesh/test_convert_MeshVTK.py::test_convert_MeshVTK
FAILED Tests/Methods/Simulation/test_VarParamSweep_fct_save_load.py::test_VarParamSweep_fct_save_load
= 24 failed, 5400 passed, 16 skipped, 93 deselected, 120 warnings in 358.84s (0:05:58) =

im also having an issue through the tutorial "how to define a simulation to call femm".

when i call this command :
out_femm.mag.meshsolution.plot_contour(label="B", group_names="stator core") i get an error relating to vtk. i guess the error is related to the following issue.
pyvista/pyvista#2580

im looking forward for your feedback!, thank you in advance :)

Answer 1 · 2023-04-25T10:07:21.000Z

so i could reduce the number of errors to 6 errors through reinstalling pyvista. however its version 0.38.5. so i get this

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
pyleecan 1.4.2 requires pyvista<=0.31.3,>=0.25.3, but you have pyvista 0.38.5 which is incompatible.

after running python -m pytest ./Tests -m "not long_5s". i get now these failed tests.

FAILED Tests/Functions/test_save_load.py::test_save_load_simu[h5] - Attribute...
FAILED Tests/Functions/test_save_load.py::test_save_load_slash_h5 - Attribute...
FAILED Tests/GUI/Dialog/DMatLib/test_DMatSetup.py::TestDMatSetup::test_BH_setup
FAILED Tests/Methods/Mesh/test_plot_meshsol.py::test_plot_glyph - TypeError: ...
FAILED Tests/Methods/Mesh/test_plot_meshsol.py::test_plot_deflection - pyvist...
FAILED Tests/Methods/Simulation/test_VarParamSweep_fct_save_load.py::test_VarParamSweep_fct_save_load
= 6 failed, 5418 passed, 16 skipped, 93 deselected, 119 warnings in 356.92s (0:05:56) =

now doing the call femm tutorial and calling the function
out_femm.mag.meshsolution.plot_contour(label="B", group_names="stator core")
gives this warning

WARNING:root:VTK requires 3D points, but 2D points given. Appending 0 third component.
however the plot opens but the results are wrong.

Answer 2 · 2023-04-25T15:18:04.000Z

Hello,

Thank you for your interest in pyleecan. I just checked and tests are running fine on my side on the master branch. I'm using python 3.8.10 on a windows 11 with pyvista 0.31.3 can you check with this version ?

Can you detail the error messages you are receiving ? We should soon release the next version of pyleecan, if you find something to update in the requirements we can add the correction to it.

Best regards,
Pierre

Answer 3 · 2023-04-25T18:36:41.000Z

thank you for your reply. im also on windows 11, python 3.8.10 and now i installed pyvista 0.38.1 and openpyxl which resolved alot of issues. however i see there is a common error which is the following:

  raise AttributeError("module {!r} has no attribute "

                         "{!r}".format(__name__, attr))

E AttributeError: module 'numpy' has no attribute 'warnings'

im a beginner in python, so im not exactly sure what this error mean.

Nasser

Answer 4 · 2023-04-26T07:21:01.000Z

I was able to reduce the errors to zero, I just downgraded Vtk and numpy to previous versions. However I’m still getting this warning while running „calling Femm simulation“.

out_femm.mag.meshsolution.plot_contour(label="B", group_names="stator core")
gives this warning

WARNING:root:VTK requires 3D points, but 2D points given. Appending 0 third component.
however the plot opens but the results are wrong.

I’m not sure if I should open a new issue for this ?

Thanks
Nasser

Answer 5 · 2023-04-26T09:27:53.000Z

Hello,

I think that the error message is "normal". We always have some struggle with the vtk plots and the jupyter notebook. Can you share the wrong plot result ? Did you change something in the notebook ?

Best regards,
Pierre

Answer 6 · 2023-04-26T10:39:54.000Z

hello,

im actually following the tutorial and i didnt change anything to jupyter. i can see that the scaling is also different from the one expected

Here is a link to the tutorial: https://www.pyleecan.org/tuto_Simulation_FEMM.html

Answer 7 · 2023-04-27T12:03:51.000Z

Hello,

I'm updating the tutorials and here are some observations/explanations:

The image is on pyleecan website so it is rather old
The image has more element than the default mesh, so it is not the result of that jupyter notebook plot
The colormap on your screenshot is from -2.48 to 2.8. I have the same on my side, when setting colormap from 0 to 3, I get something closer to the website image.
If I remember well at some point we changed the Prius winding.

Best regards,
Pierre

Answer 8 · 2023-04-28T14:54:28.000Z

Hello,

I just release pyleecan1.5.0 and updated all the tutorials on the website (there is a new loss tutorial).
Do you need further help or can we close this issue ?

Best regards,
Pierre