pybind11使用

有时需要python调用cpp编写的函数(比如做一些高性能计算),这时可以考虑使用pybind11来对你的cpp代码封装成库方便python调用.

1. 安装pybind11

python -m pip install pybind11

2. 链接cpp
   最好的教程自然是官方的,入门请看链接
   如果你需要使用cmake来集成,可以用下面的方式集成

find_package(pybind11 CONFIG REQUIRED)

# Create an extension module
add_library(mylib MODULE main.cpp)
target_link_libraries(mylib PUBLIC pybind11::module)

# Or embed the Python interpreter into an executable
add_executable(myexe main.cpp)
target_link_libraries(myexe PUBLIC pybind11::embed)

当然,有时我们的安装位置cmake找不到,可以用下面命令来指定pybind11包的查找路径.

python3.8 -m pybind11 --cmakedir
# /home/lyy/.local/lib/python3.8/site-packages/pybind11/share/cmake/pybind11
ls /home/lyy/.local/lib/python3.8/site-packages/pybind11/share/cmake/pybind11
# FindPythonLibsNew.cmake  pybind11Common.cmake  pybind11Config.cmake  pybind11ConfigVersion.cmake  pybind11NewTools.cmake  pybind11Targets.cmake  pybind11Tools.cmake

可以看到pybind11的cmake安装路径并不是cmake默认的查找路径,pybind11Config.cmake中有这么几行:

``pybind11_DIR``
  CMake variable, set to directory containing this Config file.

所以可以在cmakelists中添加下面一行来指定查找路径

set(pybind11_DIR  /home/lyy/.local/lib/python3.8/site-packages/pybind11/share/cmake/pybind11)

接下来再去find_package就不会有问题了.

github有个仓库pybind11_opencv_numpy, 作者定义了一个ndarray_converter来让cpp代码与python共享一个numpy数据,速度还是很快的.