blacksmith / py.lockfile

Tool for downloading of python packages for various OS, CPU and python versions.

py.lockfile

Install

pip install py.lockfile2

py.lockfile is a tool for downloading of python packages for various OS, CPU and python versions.

Motivation

Python projects managed by pipenv, poetry or pdm use lock files for freeze packages in specific version. It is very useful for stability of whole project. However, for installing these freeze packages, we have to use mentioned package managers, which bring many unwanted dependencies. It is uncomfortable especially for building of docker containers.

This tool takes Pipfile.lock/poetry.lock/pdm.lock file and downloads all required packages (for specific OS, CPU and python version) to target directory. After that we can install them by simple pip.

The simple example of usage

> py.lockfile -s tests/poetry.lock -t wheels/
📦 cffi 1.15.1  cffi-1.15.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

> pip install wheels/*
cffi-1.15.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl

Advance options

The tool accepts many arguments, the most useable are --python-version and --platform. They can help us to download packages for different python version or OS platform. Get the correct value of --platform attribute for your system, can be complicated. This command can help py_lockfile --help | grep 'Your current platform'

> py.lockfile --help
usage: py_lockfile [-h] [-s SOURCEFILE] [-t TARGET] [-g GROUP] [-p PYTHON_VERSION] [--platform PLATFORM] [--python-implementation {cp,ip,pp,jy}] [--ignore-missing] [--ignore-hash] [--no-binary] [--dryrun] [--no-color]

Python package downloader.
---------------------------------

This is a simple tool for download python packages managed by lock file.
Repository credentials can be set by environment variables
(PYLF_<NAME>_USERNAME, PYLF_<NAME>_PASSWORD and PYLF_<NAME>_URL).

Supported:
    * Pipfile.lock - Pipenv.
    * poetry.lock - Poetry, the repository credentials are automatically loaded
                    from ~/.config/pypoetry/auth.toml.
    * pdm.lock - PDM, the repository credentials are automatically loaded from
                 pyproject.toml

optional arguments:
  -h, --help            show this help message and exit
  -s SOURCEFILE, --sourcefile SOURCEFILE
                        source file (e.g. Pipfile.lock, poetry.lock or pdm.lock)
  -t TARGET, --target TARGET
                        Download folder (default: ./wheels)
  -g GROUP, --group GROUP
                        Append optional group of packages (e.g. dev)
  -p PYTHON_VERSION, --python-version PYTHON_VERSION
                        Download packages for python version (default: 3.9.5).
  --platform PLATFORM   Download packages for platform (default: manylinux_2_28_x86_64).
  --python-implementation {cp,ip,pp,jy}
                        Download packages for python implementation (default: cp).
  --ignore-missing      Skip missing packages.
  --ignore-hash         Ignore package hash checking.
  --no-binary           Download only source packages.
  --dryrun              Dry run. No download will be performed.
  --no-color            Disable color output.

SOURCEFILE:
    Automatically try to find supported files in current folder.

PLATFORM:
    Your current platform is 'manylinux_2_28_x86_64'
    See more https://peps.python.org/pep-0491/#file-name-convention
    e.g.:
        * macosx_10_9_x86_64
        * win_amd64
        * manylinux_2_17_x86_64    # glib linux systems (RHEL based system) https://peps.python.org/pep-0600/
        * musllinux_1_1_x86_64     # musl linux systems (AlpineLinux) https://peps.python.org/pep-0656/

PYTHON_IMPLEMENTATION:
   * 'cp' - CPython
   * 'pp' - Pypy
   * 'ip' - IronPython
   * 'jy' - Jython

As default values of --python-version, --platform and --python-implementation are set your current python configuration. However, you can download packages for different configuration as well.

> py.lockfile -s tests/poetry.lock -t wheels/ --python-version 2.7 --platform win_amd64
📦 cffi 1.15.1  cffi-1.15.1-cp27-cp27m-win_amd64.whl

Download packages from a private repository

Credentials are automatically loaded from pypoetry/auth.toml, pyproject.toml or we can overwrite them by environment variables.

PYLF_<REPO_NAME>_USERNAME=user PYLF_<REPO_NAME>_PASSWORD=password  py.lockfile ...

Optionally, we can set PYLF_<REPO_NAME>_URL as well, for appending custom pypi entry point.