How can I show edges in a VTU mesh?

[XushanLu]

I am trying to use vedo to plot an unstructured 3D surface mesh and here's what I did: from vedo import load, show surf_mesh = load('surface_mesh_with_topo.vtu) show(surf_mesh)

And here's what was plotted:

As you can see, there is not a single edge shown here. The following screenshot shows how the mesh looks like in Paraview:

How can I show the edges of the unstructured mesh and set its colors?

Here is the mesh file:

surface_mesh_with_topo.vtu.zip

[marcomusy]

You can use:

pip install -U git+https://github.com/marcomusy/vedo.git

then:

from vedo import *

ug = UGrid('data/surface_mesh_with_topo.vtu')
print(ug.getArrayNames())

# scale makes the z more visible
msh = ug.tomesh().scale([1,1,10]).lineWidth(0.1)

msh.cmap("jet", 'names001', on="cells").addScalarBar(c='w')

show(msh, axes=dict(digits=3), zoom=50)

(you can pan the image by dragging with middle mouse button)

[XushanLu]

Thanks very much for your help, @marcomusy. That is awesome. Now I can show the edges without any problems.

[XushanLu]

Hi @marcomusy ,

I've managed to show the edges but I am not sure how to show all nodes in a vtu file. Here is the code that I use to try to plot the edges and nodes at the same time but only edges are shown:

from vedo import UGrid, show, screenshot

surf_file = 'surface_mesh_with_topo.vtu'
ug = UGrid(surf_file)

msh = ug.tomesh().lineWidth(2).lineColor('w')
bounds = [504827, 505227, 6416234, 6416634, 400, 600]
msh.crop(bounds=bounds)
sargs = dict(
    color='k',
    title='Elevation (m)',
    titleFontSize=12,
    horizontal=True,
    pos=(0.2, 0.01),
)
msh.cmap("coolwarm", 'Z', on='cells')
msh.addScalarBar(
    c='k',
    title='Elevation (m)',
    titleFontSize=40,
    pos=(0.75, 0.25),
    size=(50, 800)
)

# font = 'Calco'
# font = 'Comae'
font = 'Kanopus'
axes_opt = dict(
    xtitle='Easting (m)',
    ytitle='Northing (m)',
    ztitle='',
    xTitleColor='k',
    yTitleColor='k',
    xLabelColor='k',
    yLabelColor='k',
    xLabelSize=0.015,
    xTitlePosition=0.65,
    yTitlePosition=0.65,
    xLineColor='k',
    axesLineWidth=2,
    labelFont=font,
    titleFont=font,
)

# Add the tetrahedra for the observation points
tr_list = ['013', '014', '015', '016', '017']
ele_list = ['0116311', '0116349', '0116134', '0116171', '0267813', '0267886',
            '0282449', '0282517', '0253337', '0253404']
ele_vtu = []
iele = 0
for itr in tr_list:
    filename = 'element_iTr=' + itr + '_iObs=0001_idxEle=' + ele_list[iele] + \
        '.vtu'
    ele_vtu.append(UGrid(filename))
    iele += 1
    filename = 'element_iTr=' + itr + '_iObs=0001_idxEle=' + ele_list[iele] + \
        '.vtu'
    ele_vtu.append(UGrid(filename))
    iele += 1

# Add the nodes of the source refinements
node_vtu = UGrid('src_node_with_topo.vtu').tomesh()
# points = node_vtu.insidePoints()
node_crop = node_vtu.crop(bounds=bounds).c('r')

size = [2560, 1600]
show(msh, ele_vtu, node_crop, axes=axes_opt, size=size)
screenshot('src_obs_refinement_vedo.png')

And here is what I got (as you can see, no points are plotted and I would like to plot them as spheres and be able to control the point sizes)

Here is the zip file contains all the necessary data files: Archive.zip

I guess I can convert the mesh to points and try to plot points instead of meshes but I do not know how to crop the points there, and it would be more straightforward if I can just plot things with the UGrid object.

Thanks in advance for your help!

[marcomusy]

If I understand you want to visualize the vertices of the colored mesh? If so try:

from vedo import UGrid, Spheres, show, screenshot, settings

settings.defaultFont = 'Kanopus'
settings.useParallelProjection = True # avoid perspective parallax
path = 'data/chddem'

ug = UGrid(f'{path}/surface_mesh_with_topo.vtu')

msh = ug.tomesh().lineWidth(2).lineColor('w')
bounds = [504827, 505227, 6416234, 6416634, 400, 600]
msh.crop(bounds=bounds)

msh.cmap("coolwarm", 'Z', on='cells').lighting('off')
msh.addScalarBar(
    title='Elevation (m)',
    titleFontSize=40,
    pos=(0.75, 0.25),
    size=(50, 800),
)

axes_opt = dict(
    xtitle='Easting (m)',
    ytitle='Northing (m)',
    ztitle='',
    xLabelSize=0.015,
    xTitlePosition=0.65,
    yTitlePosition=0.65,
    axesLineWidth=4,
    xrange=msh.xbounds(),
    yrange=msh.ybounds(),
)

# Add the tetrahedra for the observation points
tr_list = ['013', '014', '015', '016', '017']
ele_list = ['0116311', '0116349', '0116134', '0116171', '0267813', '0267886',
            '0282449', '0282517', '0253337', '0253404']
ele_vtu = []
iele = 0
for itr in tr_list:
    filename = f'{path}/element_iTr={itr}_iObs=0001_idxEle={ele_list[iele]}.vtu'
    ele_vtu.append(UGrid(filename))
    iele += 1
    filename = f'{path}/element_iTr={itr}_iObs=0001_idxEle={ele_list[iele]}.vtu'
    ele_vtu.append(UGrid(filename))
    iele += 1

# Add the nodes of the source refinements
node_vtu = UGrid(f'{path}/src_node_with_topo.vtu').tomesh()
node_crop = node_vtu.crop(bounds=bounds).c('r')

nodes = Spheres(msh.points(), r=1).c('k')

size = [2560, 1600]
show(msh, nodes, ele_vtu, node_crop, axes=axes_opt, size=size, zoom=1.2)
# screenshot('src_obs_refinement_vedo.png')

PS: your code looks REALLY good :) you're already mastering vedo :) I made a few changes that you might like..

[marcomusy]

..you may also like this instead of addScalarBar:

msh.addScalarBar3D(
    title='Elevation (m)',
    titleSize=1,
    titleXOffset=2,
)

[XushanLu]

Wow, that is great! I like the 3D scalar bar. The ability to have the labels of the axes rotate with the mesh (being 3D) is awesome. However, I'd still like to ask two more questions:

1. How can I get the camera parameters once I rotate the view under the interactive mode? For example, the following picture is actually what I wanted for the final plot I need to rotate the view in order to get what I want exactly. But if I need to change any parameters and rerun the script, then I need to rotate it again. I know I can set the camera position when calling show. But how can I know it in the first place?

2. I like all the axes labels that vedo can do! I can have control over the fontsize, etc, and it rotates with the view, which is awesome! I could not achieve this with PyVista (the reason why I am looking into vedo) because whatever I set as the fontsize, it will not be reflected in the final plot. I previously was using Paraview, and it was a mess when it comes to this issue. It is nothing even remotely close to being smart. However, as can be seen in the above picture, the text is still not super sharp and it is a bit blurry on the edges of the text. While being absolutely better than the other plotting tools I have tried, could it be made of publication quality?

Thanks a lot for your help!

[marcomusy]

How can I get the camera parameters once I rotate the view under the interactive mode?

Just press shift-C on the screen when you are happy with the view and it will print out a camera script that you can paste directly in your code.

The text looks blurry because you lack antialiasing (the actual text is polygonal, so it won't be blurry at any scale!). What is your OS? Are you sure you did: pip install -U git+https://github.com/marcomusy/vedo.git

What happens if you press shift-A in the scene (that should toggle antialiasing). Otherwise you can set directly settings.multiSamples = 8

You can also use pyvista with vedo axes, something like:

# ...
ax = vedo.Axes(mesh)
pyvista.add_actor(ax)

Thanks for the feedback

PS, ah, for the camera, you can also use show(..., elevation=-45)

[XushanLu]

Sorry, I am pretty sure I did not use the code in Github. Now I do, and the issue is fixed. That elevation parameter is really convenient!

I can confirm that Shift-C and Shift-A do what you described. Now I am happy about this picture. I should be trying to reproduce something other pictures that I used to plot with Paraview just for fun. Thanks!

[XushanLu]

Oh, maybe one more question: How do I set the background color of the plot? Or, is there a way to trim the boundary white spaces after saving the screenshot? In PyVista, I would set the background using 'w' color and then I can use the following code to get rid of the white space of the screenshot

from wand import image

with image.Image(filename='src_obs_refinement_vedo.png') as imag:
    imag.trim(color=None, fuzz=0)
    imag.save(filename='src_obs_refinement_vedo_trim.png')

However, both GIMP and wand seem to be unable to trim the extra whitespace near the boundary for the screenshot taken using vedo. Any ideas?

[marcomusy]

vedo can crop and visualize images, add at the end of the script:

from vedo import Picture
# ...
show(msh, nodes, ele_vtu, node_crop, axes=axes_opt,
     size=size, zoom=1.2, elevation=-45)

npimg = screenshot(returnNumpy=True)
pic = Picture(npimg)
pic.crop(left=0.15, right=0.1, top=0.1, bottom=0.1).write("a.png")
show(pic, axes=1, new=True, bg='black')

To change background color use: show()..., bg="pink", bg2="yellow") # bg2 defines a gradient color Pressing 5and 6also does. Press hto see tthe full list of options.

[XushanLu]

Thanks for pointing out the Picture class, and also the magic 'h' button. That was totally hidden to me.

Using Picture works and I am able to crop the picture to what I want. However, I still think an automatic process would be great if possible. I think the algorithms used in wand just picks up the color at one of the corners and then crops anything that has the same color in the picture from the boundary to inside.

I have set both bg and bg2 to 'w', and I still cannot use wand to crop the picture, which is different from the pictures saved using Paraview and PyVista. With these two, as long as I set the background to 'w', both GIMP and wand works.

I am happy even just using Picture, though.

[marcomusy]

uhm.. maybe if you add at the beginning: settings.rendererFrameWidth = 0 then it might work? What is your operative system?

[XushanLu]

Sorry, I forgot to mention it. My OS is Mac OS Big Sur. When you say it might work, what are you referring to? Is it the wand thing?

Adding settings.rederFrameWidth = 0 did not work for the wand solution. What does that do?

[marcomusy]

Mac OS Big Sur version 11.3 ? (i'm asking for other reasons)

What it does is to remove a thin border which is drawn to separate multiple subwindows in the same rendering window (when they exist). If that doesn't work I'll try to investigate why wand is unable to do the job.

[XushanLu]

11.2.3. I am about to update the OS but have not done that.

I think from what I saw, that did not solve the problem for wand.

[marcomusy]

well.. this is strange... your wand code above works perfectly fine on my mac! it trims the image correctly..

[XushanLu]

Does that mean I need to upgrade my OS??!!

[marcomusy]

I have no idea, ..but I doubt it!