Installation

Installation#

PyVista is supported on Python versions 3.9+.

For the best experience, please considering using Anaconda as a virtual environment and package manager for Python and following the instructions to install PyVista with Anaconda.

Dependencies#

PyVista is built on top of the Visualization Toolkit (VTK) and NumPy - as such, the following projects are required dependencies of PyVista:

vtk - PyVista directly inherits types from the VTK library.
NumPy - NumPy arrays provide a core foundation for PyVista’s data array access.
matplotlib - Used for colormaps and 2D plotting with pyvista.ChartMPL.
pillow - PIL fork used for saving screenshots.
imageio - This library is used for reading images and writing animations.
pooch - Data management for our example datasets so users can download tutorials on the fly.
scooby - Reporting and debugging tools.

Optional Dependencies#

PyVista includes several optional dependencies for visualization and reading a variety of additional file formats, including:

cmocean - Colormaps for Oceanography.
colorcet - Perceptually accurate 256-color colormaps for use with Python.
trame - Used for client and server-side rendering in Jupyter.
meshio - Input/Output for many mesh formats.

PyPI#

PyVista can be installed from PyPI using pip:

pip install pyvista

To install all the additional packages that extend PyVista, install using pip with:

pip install pyvista[all]

Anaconda#

To install this package with conda run:

conda install -c conda-forge pyvista

Installing the Current Development Branch from GitHub#

There may be features or bug-fixes that have been implemented in PyVista but have not made their way into a release. To install pyvista from the latest up-to-date development branch from github, use one of the following

pip install -U git+https://github.com/pyvista/pyvista.git@main

Alternatively, you can clone the repository with git and install it with pip.

git clone https://github.com/pyvista/pyvista.git
cd pyvista
pip install -e .

Note the development flag -e. This allows you to change pyvista in-place without having to reinstall it for each change.

Optional Dependencies#

The following are a list of optional dependencies and their purpose:

Package	Purpose
`trame`	Interactive notebook rendering
`sphinx_gallery`	Capturing PyVista output for docs
`colorcet`	Perceptually uniform colormaps
`cmocean`	Oceanographic colormaps
`imageio-ffmpeg`	Saving movie files
`tqdm`	Status bars for monitoring filters
`trimesh` `rtree` `pyembree`	Vectorised ray tracing

Source / Developers#

Alternatively, you can install the latest version from GitHub by visiting PyVista, and downloading the source (cloning) by running:

git clone https://github.com/pyvista/pyvista.git
cd pyvista
python -m pip install -e .

The latest documentation for the main branch of PyVista can be found at dev.pyvista.org.

Test Installation#

You can test your installation by running an example:

>>> from pyvista import demos
>>> demos.plot_wave()

See other examples and demos:

>>> from pyvista import examples
>>> from pyvista import demos

List all available examples.

>>> print(dir(examples))

List all available demos.

Note

A more comprehensive testing suite is available after cloning the source repository. For details on how to clone and test the PyVista source, please see our Contributing Guide and specifically, the Testing section.

Running on CI Services#

Please head over to pyvista/setup-headless-display-action for details on setting up a headless display in GitHub Actions Workflows using PyVista.

In the past, we have also supported pyvista/gl-ci-helpers for setting up CI services like Travis and Azure Pipelines to run PyVista.

Running on MyBinder#

This section is for advanced users that would like to install and use PyVista with headless displays on notebook hosting services like MyBinder.

Please see this project for a convenient Cookiecutter to get started using PyVista on the notebook hosting service MyBinder.

To get started, the Docker container will need to have libgl1-mesa-dev and xvfb installed through apt-get. For MyBinder, include the following in a file called apt.txt:

libgl1-mesa-dev
xvfb

Then, you need to configure the headless display, for MyBinder, create a file called start and include the following set up script that will run every time your Docker container is launched:

#!/bin/bash
set -x
export DISPLAY=:99.0
export PYVISTA_OFF_SCREEN=true
which Xvfb
Xvfb :99 -screen 0 1024x768x24 > /dev/null 2>&1 &
sleep 3
set +x
exec "$@"

All you have to do next is include PyVista in your Python requirements and you can get to visualizing your data. If you need more help than this on setting up PyVista for these types of services, hop on Discussions page and chat with the developers or take a look at this repository that is currently using PyVista on MyBinder.

Running on Remote Servers#

Using PyVista on remote servers requires similar setup steps as in the above Docker case. As an example, here are the complete steps to use PyVista on AWS EC2 Ubuntu 18.04 LTS (ami-0a313d6098716f372 in us-east-1). Other servers would work similarly.

After logging into the remote server, install Miniconda and related packages:

wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh -O miniconda.sh
bash miniconda.sh -b -p miniconda
echo '. $HOME/miniconda/etc/profile.d/conda.sh' >> ~/.bashrc && source ~/.bashrc
conda create --name vtk_env python=3.9
conda activate vtk_env
conda install nodejs  # required when importing pyvista in Jupyter
pip install jupyter pyvista trame

# To avoid "ModuleNotFoundError: No module named 'vtkOpenGLKitPython' " when importing vtk
# https://stackoverflow.com/q/32389599
# https://askubuntu.com/q/629692
sudo apt update && sudo apt install python-qt4 libgl1-mesa-glx

Then, configure the headless display:

sudo apt-get install xvfb
export DISPLAY=:99.0
export PYVISTA_OFF_SCREEN=true
Xvfb :99 -screen 0 1024x768x24 > /dev/null 2>&1 &
sleep 3

Reconnect to the server with port-forwarding, and start Jupyter:

ssh -i "your-ssh-key" your-user-name@your-server-ip -L 8888:localhost:8888
conda activate vtk_env
jupyter lab --NotebookApp.token='' --no-browser --port=8888

Visit localhost:8888 in the web browser.

Running on WSL#

Similar to the example of the remote server above, the windows subsystem for Linux does not provide an x-server for visualization. Instead, the fastest way to get up and running on WSL is through JupyterLab.

First, make sure you have installed the correct environment through Miniconda and related packages:

wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh -O miniconda.sh
bash miniconda.sh -b -p miniconda
echo '. $HOME/miniconda/etc/profile.d/conda.sh' >> ~/.bashrc && source ~/.bashrc
conda create --name vtk_env python=3.9
conda activate vtk_env
conda install nodejs  # required when importing pyvista in Jupyter
pip install jupyter pyvista[jupyter] trame

# To avoid "ModuleNotFoundError: No module named 'vtkOpenGLKitPython' " when importing vtk
# https://stackoverflow.com/q/32389599
# https://askubuntu.com/q/629692
sudo apt update && sudo apt install python-qt4 libgl1-mesa-glx

VTK Link to Jupyter#

There are two ways to get vtk rendering 3D objects in JupyterLab. First you can follow the example above for remote servers, skipping over the ssh instructions.

Configure the headless display:

sudo apt-get install xvfb
export DISPLAY=:99.0
export PYVISTA_OFF_SCREEN=true
Xvfb :99 -screen 0 1024x768x24 > /dev/null 2>&1 &
sleep 3

Start Jupyter:

jupyter lab --NotebookApp.token='' --no-browser --port=8888

Visit localhost:8888 in the web browser.

Finally add this example code and your interactive visualizations should be displayed in JupyterLab.

import pyvista
pl = pyvista.Plotter(shape=(1, 2))
actor = pl.add_mesh(pyvista.Cube())
pl.subplot(0, 1)
actor = pl.add_mesh(pyvista.Sphere())
pl.set_background('orange', all_renderers=False)
pl.show()

Your visualizations should now be showing directly in the Jupyter frontend.

Running with Sphinx-Gallery#

In your conf.py, add the following:

import pyvista
# necessary when building the sphinx gallery
pyvista.BUILDING_GALLERY = True
pyvista.OFF_SCREEN = True

# Optional - set parameters like theme or window size
pyvista.set_plot_theme('document')
pyvista.global_theme.window_size = np.array([1024, 768]) * 2

extensions = [
    ...
    "sphinx_gallery.gen_gallery",
]

# Add the PyVista image scraper to SG
sphinx_gallery_conf = {
    ...
    "image_scrapers": ('pyvista', ..., ),
    ...
}

We also have a Sphinx-Gallery scraper for embedding dynamic 3D scenes instead of static screenshots. This scraper can be enabled by passing an instance directly to the image_scrapers parameter instead of the string 'pyvista' above and by registering the pyvista.ext.viewer_directive extension

import pyvista
from pyvista.plotting.utilities.sphinx_gallery import DynamicScraper

# necessary when building the sphinx gallery
pyvista.BUILDING_GALLERY = True
pyvista.OFF_SCREEN = True

# Optional - set parameters like theme or window size
pyvista.set_plot_theme('document')
pyvista.global_theme.window_size = np.array([1024, 768]) * 2

extensions = [
    ...
    "sphinx_gallery.gen_gallery",
    "pyvista.ext.viewer_directive",
]

# Add the PyVista image scraper to SG
sphinx_gallery_conf = {
    ...
    "image_scrapers": (DynamicScraper(), ..., ),
    ...
}

Installation

Contents

Installation#

Dependencies#

Optional Dependencies#

PyPI#

Anaconda#

Installing the Current Development Branch from GitHub#

Optional Dependencies#

Source / Developers#

Test Installation#

Running on CI Services#

Running on MyBinder#

Running on Remote Servers#

Running on WSL#

VTK Link to Jupyter#

Running with Sphinx-Gallery#