little-cheesemonger

little-cheesemonger is a package that can set up a manylinux build environment to compile a correctly linked manylinux wheel. Configuration data is loaded from a pyproject.toml file, or a custom data loader can be implemented.

Installation

pip install little-cheesemonger==0.2.1

Quickstart

To trigger environment setup, either run little-cheesemonger from the root directory of a package, or pass the path to the package you want compiled. The package must contain configuration data in its pyproject.toml file for environment setup to work.

little-cheesemonger path/to/package

Default Configuration

The manylinux image version is controlled by last section of the configuration heading. The latest tag is always used.

[tool.little-cheesemonger.manylinux2014_x86_64]
environment_variables = [
  "FOO=BAR"
]
system_dependencies = [
  "atlas"
]
python_dependencies = [
  "nyancat==0.1.2"
]
python_versions = [
  "cp36-cp36m"
  "cp38-cp38"
]
steps = [
  "touch /foobar.txt"
]

• environment_variables is a list of environment variables to set in the container prior to building the wheel. They are expected to be in KEY=VALUE format.
• system_dependencies is a list of CentOS system dependencies to install via YUM. They are expected to be in package-version format.
• python_dependencies is a list of Python dependencies to install via pip. They are expected to be in package==version format. They are currently installed for all available versions of Python in a given manylinux image.
• python_versions is a list of Python versions to install Python dependencies for in a given manylinux image. They are expected to follow the Python installation directory naming convention from in the specified manyliniux image, ex. cp36-cp36m. Python version names can be found by running ls /opt/python in a manylinux image.
• steps is a list of steps to execute via bash. Package building and uploading could be implemented here, or little-cheesemonger could be integrated into a larger system!

Custom Data Loader

You can pass the path to an importable function, as well as positional and keyword arguments to little-cheesemonger to customize how configuration data is loaded.

little-cheesemonger --data-loader path.to.function --loader-arg foo --loader-kwarg foo=bar

Custom data loader functions must accept directory as the first argument with positional and keyword arguments passed to the loader function after. The function must return a dictionary of configuration data matching the fields above.

def my_custom_loader(directory: Path, ...) -> Dict:
  pass