packing-packages 0.0.2

This module provides functionality to pack conda environments and their dependencies into a specified directory.

PACKING-PACKAGES

This module provides functionality to pack conda environments and their dependencies into a specified directory.

You can use this package to migrate a conda environment to another offline machine with the same operating system.

Install

# Need python, setuptools, pip package
pip install packing-packages

How to use

see --help.

Pack

packing-package pack -d .

Install

packing-package install .

Example

Source device

conda activate <envname>
python -m pip install packing-packages
python -m packing_packages pack -d .

Destination device (offline)

conda create -yn <envname> --offline
conda activate <envname>
conda install --use-local --offline ./conda/*
python -m pip install --no-deps --no-build-isolation ./pypi/*

Notes

If you want to use pip to download from a source other than pypi.org (e.g. pytorch), the above will not completely collect. You will see packages that failed to download, so you will have to download them manually. This is very complicated and we do not plan to support it.

# install command
pip install torch==2.5.1 --index-url https://download.pytorch.org/whl/cu124

# download command
# --no-deps: only target package
# -d: destination directory to download
pip download torch==2.5.1 --index-url https://download.pytorch.org/whl/cu124 --no-deps -d .

Some older packages may not install with pip install. In this case, you may succeed by installing only the failed packages with the --use-pep517 option. Please perform this manually.

There are two methods of installation: one is to use the standard functions of conda and pip, and the second is to use the paciking-package install command, which wraps them. The advantages of each are summarized in the table below.

Commands	Advantages	Disadvantages
conda install or pip install	Fast	Stop immediately when cause error
packing-package install	Skip when error occurs and tell you the package	Slow