PoC / Paraview integration for mesh_doctor
#2790
Conversation
jafranc
added
type: feature
jafranc
changed the title
There was a problem hiding this comment.
Thanks @jafranc, a really good start!
@jeannepellerin you may be interested in this PR!?
| from checks import element_volumes | ||
| import vtk | ||
| # | ||
| res = element_volumes.__check(inputs[0].VTKObject, element_volumes.Options(tol)) |
There was a problem hiding this comment.
We'll need to change the name of __check 🤣
Also, if using inputs[0] is an issue, maybe mesh_doctor can evolve?
There was a problem hiding this comment.
Agreed, going into private function is hacky 😛
There was a problem hiding this comment.
Yes, that will definitely be interesting for GEOS user working with external meshes.
@jafranc Have you considered Paraview Python plugin instead of Programmable Filter in Paraview?
There was a problem hiding this comment.
I did not yet but will have a look
There was a problem hiding this comment.
Perhaps @AntoineMazuyer might be able to help on this
| :linenos: | ||
|
|
||
| input0 = inputs[0] | ||
| tol = 1.2e-6 |
There was a problem hiding this comment.
Do you know how to define tol from the GUI?
There was a problem hiding this comment.
That I think would remain to be set from the script. This would maybe possible via Plugin to have a QLineEdit (just guessing though)
| print("There are {} cells under {} m3 vol".format(extracted.GetOutput().GetNumberOfCells(), tol)) | ||
| output.ShallowCopy(extracted.GetOutput()) | ||
|
|
||
| Here we rely on ``pyvtk`` interface more than on Paraview adaptation, for legacy and reusability reasons. This is the reason |
There was a problem hiding this comment.
Do you mean that we cannot access vtkSelectionNode from within paraview?
We may want to find a way to detect that we're in the "paraview plugin env" to rely on the paraview vtk not to risk any conflict.
But the requirements.txt part remains. However, we'll move forward!
There was a problem hiding this comment.
That is exactly that. As I tried to launched it having from paraview import vtk instead of import vtk, I bumped into an error saying that there was no vtkSelectionNode. As Paraview is able to do the operation, I guess this is just that this is not exposed but wrapped. But there must be a way through Paraview, it is just a question of 1. I started from a pyvtk script, which I found more versatile but not Paraview compativle and 2. we might want to rewrite the script in a more Paraview way then.
| .. note:: | ||
| ``pip`` is installing the ``mesh_doctor`` package from the test.pypi repo, which is intended to test package deployment. | ||
| Once stabilized and ``mesh_doctor`` uploaded onto the main package repo, this should be dropped out. |
There was a problem hiding this comment.
is installing it like this working for you? When I tried it seemed to me that there were some issue in the setup.py that did not allow for this to work. Basically I could not find the modules when trying to import them.
There was a problem hiding this comment.
I got this output
Looking in indexes: https://test.pypi.org/simple/
Collecting mesh_doctor
Downloading https://test-files.pythonhosted.org/packages/97/cc/1ece6916f6e87f95cff4b887f2e6ce50488b146a2e025a959c3665a12d6e/mesh_doctor-0.0.1-py3-none-any.whl (42 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 42.4/42.4 KB 2.1 MB/s eta 0:00:00
Installing collected packages: mesh_doctor
Successfully installed mesh_doctor-0.0.1
and I can checked afterwards that checks is in `/path/to/Paraview/lib/PythonX.X' but this need refactor/cleanup obv.
There was a problem hiding this comment.
As a workaround you can try while being in mesh_doctor directory conatining the .toml
python3 -m pip install --upgrade build
python3 -m build
which should produce de dist/ directory with .whl and .tar.gz.
The install command on the .tar.gz might also work, as
python3 -m pip install dist/mesh_doctor-0.0.1.tar.gz --target /path/to/paraview/lib/pythonX.X'
Co-authored-by: TotoGaz <[email protected]>
jafranc
changed the title
mesh_doctor
|
|
Of course, I am still figuring out few things but I definitively should move that there |
|
Great to know. I will then test and amend the documentation in that sense. Thank you 🙂
…________________________________
De : Lionel Untereiner ***@***.***>
Envoyé : mercredi 27 mars 2024 11:50
À : GEOS-DEV/GEOS ***@***.***>
Cc : Jacques Franc ***@***.***>; Mention ***@***.***>
Objet : Re: [GEOS-DEV/GEOS] PoC / Paraview integration for ``mesh_doctor`` (PR #2790)
What system are you targeting ? If a user does export PV_PLUGIN_PATH=${PV_PLUGIN_PATH}:/path/to/python/module paraview will find the python plugin anywhere it lives
—
Reply to this email directly, view it on GitHub<#2790 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AL5OYFWBMF5OYT7RNQT6273Y2KJABAVCNFSM6AAAAAA6ZEZ4GOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMRSGQ3DCOBZHE>.
You are receiving this because you were mentioned.Message ID: ***@***.***>
|
|
Closed and moved to GEOS-DEV/geosPythonPackages#16 |
This is a test to leverage
mesh_doctorinParavieweither as a Python Programmable Filter or a python pluginA example is given for
element_volumescheck.It is run over the unstructured version of spe11a join here.
spe11a_mesh.zip
As a programmable filter
mesh_doctorhas to be deploy in the Paraview python hierarchy so it is able to get the import pathAs a python plugin
As examplified here, going through the loading procedure allow load and access to Mesh Doctor(GEOS) as a new filter (for now only volume check is available).