2d slices in 3d plot

[twmr]

Is it possible to create plots like the one below with matplotlib? If not, would it be difficult to add support for such plots in the mplot3d backend?

I know that such plots can be created with VTK but VTK is still limited to python 2 👎

[WeatherGod]

Slices, yes (sort of, with contourf(), not imshow()), but they can not
arbitrarially intersect. Because mplot3d is not a true 3d plotting library,
it can not handle intersecting artists, or even proper layering of separate
artists that have intersecting bounding boxes. If you want intersecting
artists, you would have to contourf() each quadrant separately, and even
then you may get rendering artifacts at different viewing angles.

They haven't ported mayavi yet?! Another project to take a look at for true
3d rendering is glumpy. Inspired from the matplotlib design, it has an
opengl rendering core which allows it to perform 3d plotting.

I hope that helps!

On Sat, Dec 13, 2014 at 2:18 PM, Thomas Hisch notifications@github.com
wrote:

Is it possible to create plots like the one below with matplotlib? If not,
would it be difficult to add support for such plots in the mplot3d backend?

[image: slice]
https://camo.githubusercontent.com/312d2210de78d9b491ddb0b26fc5185976ea0e9b/687474703a2f2f7777772e646172746d6f7574682e6564752f25374572632f636c61737365732f76697369742f74687265655f736c6963652e676966

I know that such plots can be created with VTK but VTK is still limited to
python 2 [image: 👎]

—
Reply to this email directly or view it on GitHub
#3919.

[twmr]

Do you have an example, which shows how to plot data on a plane (x,y,z + color) in a Axes3d ?

mayavi depends on vtk, therefore it is also py2 only.

[WeatherGod]

You would use the "offset" and the "zdir" arguments to the contourf() call
on a 3d axes:
http://matplotlib.org/mpl_toolkits/mplot3d/tutorial.html#filled-contour-plots

Keep in mind, those examples are surface data rather than a 3D image data,
which is what you would be handling. So, you would pass the coordinate
information for the first two arguments and then the 2D slice of the 3D
image data for the third argument. The zdir argument would indicate the
normal vector for the image slice, and the offset argument would specify
where along that normal vector the image slice should be placed.

I repeat the importance of needing to split up the image slices. You would
need to break up what would be 3 contourf() calls into 12 contourf() (4
quadrants for each slice). Luckily, numpy makes slicing and dicing your
data relatively easy...

On Sat, Dec 13, 2014 at 4:46 PM, Thomas Hisch notifications@github.com
wrote:

Do you have an example, which shows how to plot data on a plane (x,y,z +
color) in a Axes3d ?

mayavi depends on vtk, therefore it is also py2 only.

—
Reply to this email directly or view it on GitHub
#3919 (comment)
.

[twmr]

I see. Does this also work if I rotate the Axes interactively?

[twmr]

How do I set the zorder of the planes in the code below?

yplane=0.8
ind = Y[:,0] < yplane
ax.contourf(X[ind,:], Y[ind,:], Z[ind,:], zdir='z', offset=0.3)
ax.contourf(X, Z, Y, zdir='y', offset=yplane)

zorder=xx seems to have no effect.

edit:
https://stackoverflow.com/questions/14824893/how-to-draw-diagrams-like-this/14825951#14825951

[WeatherGod]

It should work with rotation just fine.

You don't mess around with zorder. mplot3d computes it dynamically to help
compose the 3d scene.

On Sat, Dec 13, 2014 at 5:43 PM, Thomas Hisch notifications@github.com
wrote:

How do I set the zorder of the planes in the code below?

yplane=0.8
ind = Y[:,0] < yplane
ax.contourf(X[ind,:], Y[ind,:], Z[ind,:], zdir='z', offset=0.3)
ax.contourf(X, Z, Y, zdir='y', offset=yplane)

zorder=xx seems to have no affect.

—
Reply to this email directly or view it on GitHub
#3919 (comment)
.

[petehuang]

Seems like we've resolved this @WeatherGod?

[twmr]

The last time I tested it, it did not work satisfactorily AFAIR.

[WeatherGod]

Have you tried glumpy, which I originally suggested? It is possible to create a function that could possibly do all of this for you, but it will still suffer from possible rendering artifacts at certain viewing angles, and it would still be a hack because it would use contourf() instead of imshow(). I would imagine glumpy would be much better suited for this need.

…

On Thu, Jan 5, 2017 at 11:13 PM, Thomas Hisch ***@***.***> wrote: The last time I tested it, it did not work satisfactorily AFAIR. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3919 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AARy-GVP1tnNsoMJXLDnY_qIJSjZM3Z8ks5rPb9dgaJpZM4DIDoZ> .

[louity]

As explained in this stackoverflow answer, matplotlib won't be able to make the 3 slices plots on the same figure "until OpenGL support is added to all of the backends".

Indeed, occlusions of the different slices will not be rendered correctly, and you will get this kind of result (I plot the 3d function (x,y,z) -> x^2 + y^2 + z^2 over the domain [-1,1]^3) :

However, you can plot each slice on a separate figure :

or plot slices that do not occlude each other from the point of view you are looking from.

code :

import matplotlib.pyplot as plt
import numpy as np
from mpl_toolkits.mplot3d import Axes3D

def plot_3D_array_slices(array):
    min_val = array.min()
    max_val = array.max()
    n_x, n_y, n_z = array.shape
    colormap = plt.cm.YlOrRd

    x_cut = array[n_x//2,:,:]
    Y, Z = np.mgrid[0:n_y, 0:n_z]
    X = n_x//2 * np.ones((n_y, n_z))
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')
    ax.plot_surface(X, Y, Z, rstride=1, cstride=1, facecolors=colormap((x_cut-min_val)/(max_val-min_val)), shade=False)
    ax.set_title("x slice")

    y_cut = array[:,n_y//2,:]
    X, Z = np.mgrid[0:n_x, 0:n_z]
    Y = n_y//2 * np.ones((n_x, n_z))
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')
    ax.plot_surface(X, Y, Z, rstride=1, cstride=1, facecolors=colormap((y_cut-min_val)/(max_val-min_val)), shade=False)
    ax.set_title("y slice")

    z_cut = array[:,:,n_z//2]
    X, Y = np.mgrid[0:n_x, 0:n_y]
    Z = n_z//2 * np.ones((n_x, n_y))
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')
    ax.plot_surface(X, Y, Z, rstride=1, cstride=1, facecolors=colormap((z_cut-min_val)/(max_val-min_val)), shade=False)
    ax.set_title("z slice")

    plt.show()


n_pts = 100
r_square = (np.mgrid[-1:1:1j*n_pts, -1:1:1j*n_pts, -1:1:1j*n_pts]**2).sum(0)
plot_3D_array_slices(r_square)

[louity]

You might actually overcome this issue using matplotlib Poly3DCollection, see example, but as you can read in the doc :"this class (Poly3DCollection) does a bit of magic with the _facecolors and _edgecolors properties."... Let me know if you manage to do it :-).

[louity]

There is an example of solution using plotly.

[tacaswell]

Also see https://stackoverflow.com/questions/14824893/how-to-draw-intersecting-planes/14825951#14825951 . Manually chopping your slices up can also be made to work.

[jiadongdan]

Is it type of plot still possible? As of matplotlib v3, it seems there no proper way to handle this

[WeatherGod]

It should work, but without a code example to demonstrate the problem, I can't help you further.

…

On Tue, Jul 30, 2019 at 9:50 AM Chao Gao ***@***.***> wrote: 3D don't support zdir = 'x' ???? https://gc13141112.github.io/test/test.png <http://url> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3919?email_source=notifications&email_token=AACHF6AAJF4IXGLEZ7WUSEDQCBBMHA5CNFSM4AZAHIM2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3EA6VQ#issuecomment-516427606>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AACHF6HPQTGQUKBNNLMFLVTQCBBMHANCNFSM4AZAHIMQ> .

[github-actions]

This issue has been marked "inactive" because it has been 365 days since the last comment. If this issue is still present in recent Matplotlib releases, or the feature request is still wanted, please leave a comment and this label will be removed. If there are no updates in another 30 days, this issue will be automatically closed, but you are free to re-open or create a new issue if needed. We value issue reports, and this procedure is meant to help us resurface and prioritize issues that have not been addressed yet, not make them disappear. Thanks for your help!

[KrisSatie]

Is it type of plot still possible? As of matplotlib v3, it seems there no proper way to handle this

[dongjiaping]

yes, we need a better way to solve make 3 slices plots on the same figure.

[scottshambaugh]

Relabeling the reason for closure as "can't fix", as this is a renderer issue. However, plotting slices can be done by plotting each quadrant of the slice around the intersection separately. Here's an example (which could probably be made more compact but demonstrates the concept):

import matplotlib.pyplot as plt
import numpy as np

def plot_quadrants(ax, slice_data, fixed_coord, coord_val, colormap, shape, min_val, max_val):
    half_shape = (shape[1] // 2, shape[2] // 2)
    quadrants = [
        slice_data[:half_shape[0], :half_shape[1]],
        slice_data[:half_shape[0], half_shape[1]:],
        slice_data[half_shape[0]:, :half_shape[1]],
        slice_data[half_shape[0]:, half_shape[1]:]
    ]
    
    for i, quad in enumerate(quadrants):
        if fixed_coord == 'x':
            Y, Z = np.mgrid[0:half_shape[0], 0:half_shape[1]]
            X = coord_val * np.ones_like(Y)
            Y_offset = (i // 2) * half_shape[0]
            Z_offset = (i % 2) * half_shape[1]
            ax.plot_surface(X, Y + Y_offset, Z + Z_offset, rstride=1, cstride=1, facecolors=colormap((quad-min_val)/(max_val-min_val)), shade=False)
        elif fixed_coord == 'y':
            X, Z = np.mgrid[0:half_shape[0], 0:half_shape[1]]
            Y = coord_val * np.ones_like(X)
            X_offset = (i // 2) * half_shape[0]
            Z_offset = (i % 2) * half_shape[1]
            ax.plot_surface(X + X_offset, Y, Z + Z_offset, rstride=1, cstride=1, facecolors=colormap((quad-min_val)/(max_val-min_val)), shade=False)
        elif fixed_coord == 'z':
            X, Y = np.mgrid[0:half_shape[0], 0:half_shape[1]]
            Z = coord_val * np.ones_like(X)
            X_offset = (i // 2) * half_shape[0]
            Y_offset = (i % 2) * half_shape[1]
            ax.plot_surface(X + X_offset, Y + Y_offset, Z, rstride=1, cstride=1, facecolors=colormap((quad-min_val)/(max_val-min_val)), shade=False)

def plot_3D_array_slices(array):
    colormap = plt.cm.YlOrRd
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')

    min_val = array.min()
    max_val = array.max()
    
    x_slice = array[array.shape[0]//2, :, :]
    y_slice = array[:, array.shape[1]//2, :]
    z_slice = array[:, :, array.shape[2]//2]
    
    plot_quadrants(ax, x_slice, 'x', array.shape[0]//2, colormap, array.shape, min_val, max_val)
    plot_quadrants(ax, y_slice, 'y', array.shape[1]//2, colormap, array.shape, min_val, max_val)
    plot_quadrants(ax, z_slice, 'z', array.shape[2]//2, colormap, array.shape, min_val, max_val)
    
    plt.show()

n_pts = 100
r_square = (np.mgrid[-1:1:1j*n_pts, -1:1:1j*n_pts, -1:1:1j*n_pts]**2).sum(0)
plot_3D_array_slices(r_square)

[timhoffm]

Thanks @scottshambaugh! I think this would be good as a gallery example. There were some indices mixed up (noticed when using different shapes for the dimensions). Below is a corrected version, which also simplifies the plot_quadrants interface:

import matplotlib.pyplot as plt
import numpy as np


def plot_quadrants(ax, array, fixed_coord, cmap):
    """For a given 3d *array* plot a plane with *fixed_coord*, using four individual quadrants."""
    nx, ny, nz = array.shape
    index = {
        'x': (nx//2, slice(None), slice(None)),
        'y': (slice(None), ny//2, slice(None)),
        'z': (slice(None), slice(None), nz//2),
    }[fixed_coord]
    plane_data = array[index]

    n0, n1 = plane_data.shape
    quadrants = [
        plane_data[:n0//2, :n1//2],
        plane_data[:n0//2, n1//2:],
        plane_data[n0//2:, :n1//2],
        plane_data[n0//2:, n1//2:]
    ]

    min_val = array.min()
    max_val = array.max()
    
    cmap = plt.get_cmap(cmap)
    
    for i, quadrant in enumerate(quadrants):
        facecolors = cmap((quadrant - min_val) / (max_val-min_val))
        if fixed_coord == 'x':
            Y, Z = np.mgrid[0:ny//2, 0:nz//2]
            X = nx//2 * np.ones_like(Y)
            Y_offset = (i // 2) * ny//2
            Z_offset = (i % 2) * nz//2
            ax.plot_surface(X, Y + Y_offset, Z + Z_offset, rstride=1, cstride=1, facecolors=facecolors, shade=False)
        elif fixed_coord == 'y':
            X, Z = np.mgrid[0:nx//2, 0:nz//2]
            Y = ny//2 * np.ones_like(X)
            X_offset = (i // 2) * nx//2
            Z_offset = (i % 2) * nz//2
            ax.plot_surface(X + X_offset, Y, Z + Z_offset, rstride=1, cstride=1, facecolors=facecolors, shade=False)
        elif fixed_coord == 'z':
            X, Y = np.mgrid[0:nx//2, 0:ny//2]
            Z = nz//2 * np.ones_like(X)
            X_offset = (i // 2) * nx//2
            Y_offset = (i % 2) * ny//2
            ax.plot_surface(X + X_offset, Y + Y_offset, Z, rstride=1, cstride=1, facecolors=facecolors, shade=False)


def figure_3D_array_slices(array, cmap=None):
    """Plot a 3d array using three intersecting centered planes."""
    fig = plt.figure()
    ax = fig.add_subplot(projection='3d')
    ax.set_box_aspect(array.shape)
    plot_quadrants(ax, array, 'x', cmap=cmap)
    plot_quadrants(ax, array, 'y', cmap=cmap)
    plot_quadrants(ax, array, 'z', cmap=cmap)
    return fig, ax


nx, ny, nz = 70, 100, 50
r_square = (np.mgrid[-1:1:1j*nx, -1:1:1j*ny, -1:1:1j*nz]**2).sum(0)

figure_3D_array_slices(r_square, cmap='viridis_r')
plt.show()

[scottshambaugh]

That is cleaner, full disclosure I had chatgpt do my example so I'm not surprised it messed up indexing! Agreed it'd make a nice gallery example, I can whip that up.