unevenly spaced coordinates for volume rendering

[rabernat]

I just discovered ipyvolume and think it is fantastic! Thanks for this contribution to the python visualization world.

I work in ocean modeling, and I often want to visualize the output of my simulations using volume rendering. I have been using paraview (example), but it is pretty heavy duty and doesn't integrate well with the rest of my analysis pipeline. I would be thrilled to be able to stay in a notebook environment.

Ipyvolume could really fill a missing need for me if it supported unevenly spaced coordinates (as in a VTK rectilinear grid). I would like to be able to do something like this:

ipyvolume.volshow(data, x=x, y=y, z=z)

where x, y, and z, are one-dimensional coordinate variables. (Even better would be to be able to pass and xarray DataArray, which has the coordinates built in.)

Would this be a difficult feature to add?

[maartenbreddels]

Hi, thanks for your kind words, nice to hear it's being used (or planned to be)in different sciences.

Would binning the data to a regular grid work? How are the samples distributed? If you have many per gridcell you could take a look at vaex.

import vaex
ds = vaex.from_arrays(x=x, y=y, z=z)
vol = ds.count(binby=['x', 'y', 'z'], shape=128)

And volshow that. Vaex can process about 1 billion samples a second, it the mother project of ipyvolume.
Btw, your links are not working.

[rabernat]

Sorry, here is the fixed link to xarray: http://xarray.pydata.org/en/stable

[rabernat]

Would binning the data to a regular grid work? How are the samples distributed? If you have many per gridcell you could take a look at vaex.

The samples are distributed regularly, on a grid, but unevenly (i.e closer grid spacing near the surface.) The thing I don't understand from your example is, where do I input the actual data? In my example data is a scalar (e.g. temperature) located at the x, y, and z points.

[maartenbreddels]

The video link is now also working, so i understand it better now.
Using vaex you'd have to use ds.mean, but what i think you need is: https://docs.scipy.org/doc/scipy-0.18.1/reference/generated/scipy.interpolate.griddata.html#scipy.interpolate.griddata
Interpolate on a regular grid and volshow that. Let me know it that works out, would be nice if you could share the data/notebook for maybe one of the examples for how ipyvolume can be used.

[rabernat]

Thanks @maartenbreddels for these tips! I am an experienced numpy/scipy user and am familiar with various approaches for interpolating and regridding data.

What I am asking with my issue is whether you think ipyvolume should do these things for the user automatically, using the api I suggested above. Your response suggests that the answer is "no"; such features are outside the scope of what you aim to provide.

[maartenbreddels]

Ah ok, I misunderstood you then. Well, i'm not against supporting this, although I wouldn't directly add it in volshow now. I did think of having say 'incubator.py' module, where a function that does this could go without problem (with the regridding), and where when things are more mature/sorted out/merged etc could then move to the pylab.py file. Another thing might be to have a per science module, 'astro.py', 'earth.py', 'bio.py' etc where more science specific things could do.

But I think this example is quite general. I also think it must be possible to do this without regridding/interpolating, so in the long term I do see this as something that volshow could support. Until (or if) that happens, I'm happy to accept a PR with an incubator.py file that does sth like you propose, volshow_rectilinear(data, x=x, y=y, z=z) using scipy to regrid the data.

[victortxa]

@maartenbreddels thanks for your explanation above. However I would like to confirm some information. Now, it is only necessary to have a 3d numpy array to plot using ipyvolume. So, if I have any input I would like to plot using ipyvolume I need to "convert" it to a 3d numpy array, right?
I am asking it also because I have a python package which 3d output I would like to plot with ipyvolume. The output are the bounds of the geometric element (polyhedron) with a correspondent property value for it. For example, I have the coordinates of the bounds of a cube and inside this cube I consider I have a property with value 2.0. I was wondering if it would be possible to only give these bounds and a property value in order to have a plot of my figure.

[maartenbreddels]

Hi @victortxa, please open a new issue the next time.

I am not sure I understand your question, but ipyvolume does both volume rendering (volshow, displayed 3d cubes/arrays) and glyphs (a shape/symbol at a specific location). Currently it does not do arbitrary objects, like meshes, but pythreejs would solve think I'd think.
Let me know if that answers your question.

[mfouesneau]

With python 3.10, ipyvolume=='0.6.0-alpha.10'

I cannot show a volume from xarray dataarray

import ipyvolume as ipv
ipv.volshow(ds)

I obtain the following error message

Open Browser Console for more detailed log - Double click to close this message]
Failed to create view for 'VolumeView' from module 'ipyvolume' with model 'VolumeModel' from module 'ipyvolume'
TypeError: Cannot read properties of undefined (reading 'getRenderSize')
    at d.render (http://localhost:8888/lab/extensions/ipyvolume/static/568.9cf325ecd07f6f56b10f.js?v=9cf325ecd07f6f56b10f:1:255709)
    at http://localhost:8888/lab/extensions/@jupyter-widgets/jupyterlab-manager/static/150.b0e841b75317744a7595.js?v=b0e841b75317744a7595:1:2959
    at async w.renderModel (http://localhost:8888/lab/extensions/@jupyter-widgets/jupyterlab-manager/static/134.40eaa5b8e976096d50b2.js?v=40eaa5b8e976096d50b2:1:72264)

ds is a regular grid 64x64x64
description: xarray.DataArray gmag: 64 g_min_rp: 64 dmod: 64

quickvolshow seems to work but does not read the coordinates.