Getting Started¶
Clone the GitHub repository:
$ git clone https://github.com/yoctoyotta1024/microphysics_testcases.git
and install the pre-commit hooks:
$ pre-commit install
That’s it, you’re done!
Using PyBind11 on ARM64 (Macbooks with M1 Processors) instead of x86_64 Architectures¶
If you are using Python on an ARM64 architecture (e.g. you have a Mac with an M1 silicon chip) then
you need to ensure the python libraries have been installed for ARM64 and not x86_64
architectures. You can check this by entering either file [your_python_library].dylib
or
file [your_python_library].so
, which will tell you which architecture your python is installed
for.
If you don’t know where your python libraries are installed, have a look at which python
to find
the path to your python interpreter, e.g. something like /path/to/python/bin/python
. Then
look in /path/to/python/lib/
for something like libpython[X.Y].dylib
or libpython[X.Y].so
where X.Y is your python version (which you can get from python --version
).
If your python libraries are not ARM64, then you will have problems… One solution is to create a conda enviroment which specifically uses conda-forge’s osx-arm64 directory to find packages, e.g. via
$ CONDA_SUBDIR=osx-arm64 conda create -n [my_arm64_env] numpy -c conda-forge
$ conda activate [name_my_arm64_env]
$ conda config --env --set subdir osx-arm64
where [name_my_arm64_env]
is the name of your new enviroment.