Interactive Data Visualization

In version 3.3 of yt, an experimental, hardware-accelerated interactive volume renderer was introduced. This interactive renderer is based on OpenGL and natively understands adaptive mesh refinement data; this enables (GPU) memory-efficient loading of data. The data is copied from CPU memory onto the GPU as a series of 3D textures, which are then rendered to an interactive window. The window itself is the view from a conceptual “camera”, which can be rotated, zoomed, and so on. The color of each displayed pixel is computed by a “fragment shader” which is executed on each grid that is displayed. The fragment shaders currently implemented in yt enable computing (and then mapping to a colormap) the maximum value along each pixel’s line of sight and an unweighted integration of values along each pixel’s line of sight (and subsequent mapping to a colormap.) An experimental transfer function shader has been implemented, but is not yet functioning correctly. For more information, see Types of Projections.

A comprehensive description of the OpenGL volume rendering is beyond the scope of this document. However, a more detailed explanation can be found in this guide.

Much of the Interactive Data Visualization (IDV) interface is designed to mimic the interface available for software volume rendering (see 3D Visualization and Volume Rendering) so that in future versions API compatibility may lead to greater code reuse both for scripts that create visualizations and for internal visualization objects.

Installation

In order to use Interactive Data Visualization (IDV) you need to install PyOpenGL and cyglfw3 along with their respective dependencies, e.g. glfw3 is required to be installed before you can pip install cyglfw3. Please carefully read installation instructions provided on pypi pages of both packages.

If you are using conda, cyglfw3 is provided in our conda channel (pyopengl is shipped by Continuum already) and can be installed via:

conda install -c http://use.yt/with_conda/ cyglfw3 pyopengl

Using the interactive renderer

You can simply pass dataset to interactive_render(). By default it will load all data and render gas density:

import yt

ds = yt.load("IsolatedGalaxy/galaxy0030/galaxy0030")
yt.interactive_render(ds)

Alternatively you can provide a data object as a first argument to interactive_render() if your dataset is too big to fit GPU memory:

import yt

ds = yt.load("IsolatedGalaxy/galaxy0030/galaxy0030")
sp = ds.sphere("max", (0.1, "Mpc"))

cam_pos = ds.arr([0.1, 0.1, 0.1], "Mpc").in_units("code_length")
yt.interactive_render(sp, field="pressure", cam_position=cam_pos,
                      window_size=(512, 512))

A successful call to interactive_render() should create a new window called vol_render.

../_images/idv.jpg

By default it renders a Maximum Intensity Projection of the density field (see Types of Projections for more information). The rendering can be dynamically modified using the following keybindings:

1
Switch to MIP fragment shader
2
Switch to integration fragment shader
L
Switch between linear and logarithmic scales
W
Zoom in the camera
S
Zoom out the camera
C
Change the colormap

Pressing the h key will print all the available key bindings in a terminal window. The camera can be moved around by holding a left mouse button while moving the mouse.

More advanced initialization of interactive volume renderer can be found in Advanced Interactive Data Visualization.