Reputation: 87
I created a miniconda environment using the following command:
conda create -n build_a_python_cpp_module xtensor-python -c conda-forge
Activated the environment and created a cookiecutter project for x-tensor (a C++ library for fast array computations like NumPy) in the project folder.
pip3 install cookiecutter
cookiecutter https://github.com/QuantStack/xtensor-python-cookiecutter.git
After the project was created, I wrote a simple program to count the commas and new lines in a file in both C++ and python to compare the performance difference.
I bound the function in C++ using pybind11 and ran
python3 setup.py install
Importing the package and running the C++ method inside it using the python binding works but I want to know:
What should I do to be able to package it and transfer the code to another machine?
Upvotes: 1
Views: 1488
Reputation: 424
You should be able to run python setup.py bdist_wheel
. This will create a wheel that you can upload on pypi.
You can test if it works with pip install dist/name_of_your_wheel
before uploading it.
Let me know if you encounter any issue
Upvotes: 3