Installing opencv for pypyp3-CodePudding

I've installed pypy3 on Mac, and would like to use it to speed up a python script to analyse a live video feed. I've tried to install opencv-python with pip and pip_pypy3, but get the following error:

Building wheels for collected packages: opencv-python
  Building wheel for opencv-python (pyproject.toml) ... error
  error: subprocess-exited-with-error
  
  × Building wheel for opencv-python (pyproject.toml) did not run successfully.
  │ exit code: 1
  ╰─> [9 lines of output]
      Traceback (most recent call last):
        File "/private/var/folders/7v/1cg_j6kd34l4k8zgvj782ndc0000gn/T/pip-build-env-nymhzjmp/overlay/site-packages/skbuild/setuptools_wrap.py", line 612, in setup
          cmkr = cmaker.CMaker(cmake_executable)
        File "/private/var/folders/7v/1cg_j6kd34l4k8zgvj782ndc0000gn/T/pip-build-env-nymhzjmp/overlay/site-packages/skbuild/cmaker.py", line 148, in __init__
          self.cmake_version = get_cmake_version(self.cmake_executable)
        File "/private/var/folders/7v/1cg_j6kd34l4k8zgvj782ndc0000gn/T/pip-build-env-nymhzjmp/overlay/site-packages/skbuild/cmaker.py", line 105, in get_cmake_version
          ) from err
      
      Problem with the CMake installation, aborting build. CMake executable is cmake
      [end of output]
  
  note: This error originates from a subprocess, and is likely not a problem with pip.
  ERROR: Failed building wheel for opencv-python
Failed to build opencv-python
ERROR: Could not build wheels for opencv-python, which is required to install pyproject.toml-based projects

Any idea how I can get around this error? Thanks!

(The advice from people who have previously had this problem is rather confusing...)

CodePudding user response：

How did you tried to install opencv?. Try this full path if you are in Mac

pip install opencv-python==4.5.3.56

CodePudding user response：

opencv-python does not release binary wheels for PyPy, so you are on your own to compile it, which is non-trivial. I would recommend using conda-forge instead since there are many more binary packages available there. If you want, you could ask politely for PyPy wheels on this issue on the opencv-python issue tracker.

With that, PyPy is unlikely to be faster if much of your code uses opencv. In order to support the CPython C-API, PyPy needs to emulate C so each function call is about 4x slower, and of course the JIT cannot look inside compiled C/C code. If your code is pure python, PyPy may be able to help, this sobel filter in python is what got me involved in PyPy over a decade ago.