This is a collection of simple to more involved examples to scripting in Blender with Python.
Table of Contents
- Simple Sphere
- Parametric Torus
- Voronoi Landscape
- Tetrahedron Fractal
- Phyllotaxis Flower
- Rugged Donut
- Fisher Iris Visualization
- Voronoi Sphere
To run the examples, open your favorite console in the example folder, make sure that the Blender executable is an environment variable or in the PATH environment variable in Windows and run the following command. Make sure to edit in run_script.py the
scriptFile variable to the Python script in the scripts folder you want to execute.
blender -b -P run_script.py
Another option is to open the script in Blender and run run_script.py inside Blender, which is a nice way to test and tweak the files and to see and play with the generated result before rendering.
Some frequently used functions in blender, which will be used in most of the scripts.
Simple rendering of a smooth sphere. First an icosphere is added with
import bpy bpy.ops.mesh.primitive_ico_sphere_add(location=(0, 0, 0)) obj = bpy.context.object
Then the subdivision surface modifier is added to the object to increase the resolution of the mesh and afterwards all the faces of the object are set to a smooth shading
modifier = obj.modifiers.new('Subsurf', 'SUBSURF') modifier.levels = 2 modifier.render_levels = 2 mesh = obj.data for p in mesh.polygons: p.use_smooth = True
Alternatively the icosphere can be subdivided with the
subdivisions argument in the function
bpy.ops.mesh.primitive_ico_sphere_add(subdivisions=4, location=(0, 0, 0))
Parametric generation of a torus. The torus is created with the following parameterization of a grid of the variables u, v
where the values u, v are between 0 and 1 and are then mapped to x, y, z coordinates. In parametric_torus.py, the function
torusSurface(r0, r1) returns the surface parameterization function for a torus which is then used in
createSurface(surface, n, m) as the first argument, which creates the object from a n by m grid. The function
createSurface(surface, n, m) can be also used for other parameterizations such as surfaces of revolution or other parametric surfaces.
Generate random metaballs in Blender inspired by this tutorial.
This is a more advanced example for using a Voronoi diagram. The Voronoi diagram is implemented with the module
scipy.spatial which can be added with Scipy, or can be found in the Python distribution Anaconda. The steps to use Anaconda as the Interpreter in Blender 2.77 are shown in this solution.
This is an example for a fractal tetrahedron, where each tetrahedron is subdivided into smaller pieces with a recursive function. In order to create a material for the tetrahedron the material is assigned as shown here:
color = (0.5, 0.5, 0.5) mat = bpy.data.materials.new('Material') # Diffuse mat.diffuse_shader = 'LAMBERT' mat.diffuse_intensity = 0.9 mat.diffuse_color = color # Specular mat.specular_intensity = 0 obj.data.materials.append(mat)
This script implements a Phyllotaxis Flower which aranges leaves or the petals according to the golden angle. Additionally The flower is animated by appending an application handler for frame change by
def handler(scene): frame = scene.frame_current # Create new geometry for new frame # ... # Append frame change handler on frame change for playback and rendering (before) bpy.app.handlers.frame_change_pre.append(handler)
In order to render all frames you can run
The animation is inspired by the mesmerizing sculptures by John Edmark.
This script implements a number of different things available in Blender. For one it applies a Displace modifier to a torus which displaces the object with a texture as follows.
# Create musgrave texture texture = bpy.data.textures.new('Texture', 'MUSGRAVE') # Create displace modifier and apply texture displace = obj.modifiers.new('Displace', 'DISPLACE') displace.texture = texture
Further we can control the texture by an object such as an Empty object
# Create Empty to control texture coordinates empty = bpy.data.objects.new('Empty', None) bpy.context.scene.objects.link(empty) # Take the texture coordinates from empty’s coordinate system displace.texture_coords = 'OBJECT' displace.texture_coords_object = empty
Additionally we want to add a material with additional bump map to our torus object which is done in the following way.
# Create bump map texture bumptex = bpy.data.textures.new('BumpMapTexture', 'CLOUDS') # Create material mat = bpy.data.materials.new('BumpMapMaterial') # Add texture slot for material and add texture to this slot slot = mat.texture_slots.add() slot.texture = bumptex slot.texture_coords = 'GLOBAL' slot.use_map_color_diffuse = False slot.use_map_normal = True # Append material to object obj.data.materials.append(mat)
Now we want to animate the empty in order to animate the texture. We can achieve this by inserting keyframes for the location of our empty as shown in this quick tutorial and in the next snippet.
for frame in range(1, num_frames): t = frame / num_frames x = 0.7*cos(2*pi*t) y = 0.7*sin(2*pi*t) z = 0.4*sin(2*pi*t) empty.location = (x, y, z) empty.keyframe_insert(data_path="location", index=-1, frame=frame)
Fisher Iris Visualization
This script implements a visualization of the famous Fisher's Iris data set. The data set consists of 50 samples for each of three flower species of Iris setosa, Iris virginica and Iris versicolor. Each sample consists of four features (sepal length, sepal width, petal length and petal width). In order to visualize the data set in three dimensions we apply dimensionality reduction by using Principal Component Analysis. The data set and PCA are both included in the scikit-learn library for Python. This script works both with or without sklearn which is not part of the Blender Python distribution. You can use sklearn by using Anaconda in Blender which I show in this quick tutorial.
from sklearn import datasets from sklearn import decomposition # Load Dataset iris = datasets.load_iris() X = iris.data y = iris.target labels = iris.target_names # Reduce components by Principal Component Analysis from sklearn X = decomposition.PCA(n_components=3).fit_transform(X)
The data set in /scripts/data/iris/ is downloaded from the UCI Machine Learning Repository and PCA is implemented manually with the help of the included Numpy library. If sklearn is not in the current Python distribution the Iris data set is loaded as in the next code snippet.
path = os.path.join('data', 'iris', 'iris.data') iris_data = np.genfromtxt(path, dtype='str', delimiter=',') X = iris_data[:, :4].astype(dtype=float) y = np.ndarray((X.shape,), dtype=int) # Create target vector y and corresponding labels labels, idx = , 0 for i, label in enumerate(iris_data[:, 4]): if label not in labels: labels.append(label); idx += 1 y[i] = idx - 1 # Reduce components by implemented Principal Component Analysis X = PCA(X, 3)
The data set is loaded into the scene as a 3D scatter plot with different shape primitives for each class of flower from the BMesh Operators. Additionally each collection of shapes in a class has different materials assigned to them. Each class has corresponding labels which are rotated toward the camera by a Locked Track Constraint.
This is another example using the Voronoi diagram, but this time in the 3rd dimension. It is implemented as well with the module
scipy.spatial which can be added with Scipy and it is even used in a similar way as the previous Voronoi example in 2D.