A template for building Blender addons with C++ and Eigen based on numpyeigen.
This template implements a simple Blender addon including a C++ shared library, a Blender operator and a UI panel. Additionally, a small utility addon is included to reload addons during development.
This template is based on the numpyeigen example project, with example code
from the libigl tutorial.
C++ sources are stored in the src dir while python sources go in
the cppTemplate dir. For a full introduction to Blender addons refer to
the official Addon Tutorial and the Blender Python API documentation.
You can build the project using CMake. From the project root run
mkdir build
cd build
cmake ..
make
This will generate the C++ shared library build/libtemplate.*.so
which can be imported from Python as import libtemplate.
Additional C++ dependencies can simply be added to the CMakeLists.txt file.
Python dependencies can be installed using
/path/to/blender/python -m pip install myDependency
To find the location of Blender's python binary run bpy.app.binary_path_python
in the Blender python console.
It is recommended to use the provided Reloader addon to install and
reload development addons. The Reloader addon can be installed by executing reload_addons.py file under Blender's python interpreter.
blender -b -P reload_addons.py
This addon provides two commands, "Reload All Addons" and
"Reload Enabled Addons" which can be accessed from the Blender
menu in the topbar under System, or through the tool search bar.
This will search the user addons directory (default ~/src/Blender,
configurable in Preferences -> Addons -> Reload Addons) for addon sources,
install them and reload them in Blender's interpreter.
The Reloader will call the shell script package_addon.sh before reloading.
This can be used to perform whatever build actions you may need.
By default it will recompile the C++ source and copy the library into the
cppTemplate directory.
You can also choose to install the addon manually. To do this copy the compiled
library from build to cppTemplate, then zip the cppTemplate directory and
install it in Blender under Preferences -> Addons -> Install.
In addition to the addon skeleton, the template includes a couple
of useful development utilities in the cppTemplate/utils directory.
NumpyMesh is a context manager which handles transfering mesh
data between Numpy and Blender. Wrap your geometry code in
in a NumpyMesh context and it'll handle extracting the vertices
and faces into numpy ndarrays and putting them back into the
Blender mesh object when you're done
myMesh = bpy.data.meshes['Suzanne']
with NumpyMesh(myMesh) as mesh:
V = mesh.V
F = mesh.F
(newV, newF) = libtemplate.my_cool_geometry_function(V, F)
mesh.V = V
mesh.F = F
During development it's helpful to be able to recompile and reload
your addon without restarting Blender. Unfortunately, Python does not
support "hot-reloading" external C++ modules. However, this can be circumvented
using the SharedLib wrapper. Use this wrapper instead of the normal module import
to executes each function call in a separate process which imports a fresh version
of the library.
class MyOperator(bpy.types.Operator):
def execute(self, context):
libtemplate = SharedLib('libtemplate')
I = libtemplate.eye(3)
Note that if the library is imported normally this will override the wrapper's version and the module will not be updated. You will need to restart Blender to allow the new version to be imported.