To install pyplatypus use command:
pip install pyplatypus
See documentation at: GitHub Pages.
To build and deploy your updated docs, use:
mkodcs build
Which will create site/ directory in the project folder. Then type in:
mkdocs gh-deploy
to automatically push the docs to GitHub Pages.
To build the package locally just move to its main directory run:
pip install --use-feature=in-tree-build .
Then it may be imported as any other package installed with the use of pip.
If you wish to create the source distribution (in the .egg format):
python setup.py sdist
To build a wheel run:
python setup.py bdist_wheel --universal
Important!
Before uploading the source distribution to PyPI test it using the TestPyPI using:
twine upload --repository-url https://test.pypi.org/legacy/ source/package.tar.gz
To upload all the distributions from a certain folder, for instance if you created
wheels aside from the archive source distribution, use this line:
twine upload --repository-url https://test.pypi.org/legacy/ folder/*
Then try installing from TestPyPI:
pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple pyplatypus
Only after testing it there should you proceed with the release:
twine upload dist/*
And then install your package!
pip install pyplatypus
Note!
Pip dependencies will be installed during the package build but there is one conda dependency that
needs to be installed by hand if you are planning to play around with the GPU.
conda install -c conda-forge tensorflow==2.9.1
We are working on creating the deployment script for this :)