🗂 Table of Contents

• Introduction
• Installation
• Usage

📖 Introduction

Starting with a pixi lockfile pixi.lock, you can create a packed environment that can be shared with others. This environment can be unpacked on any system using pixi-unpack to recreate the original environment.

In contrast to conda-pack, pixi-pack does not require the original conda environment to be present on the system for packing. Instead, it uses the lockfile to download the required packages and puts them into a .tar archive. This archive can then be shared with others and installed using pixi-unpack to recreate the original environment.

The original motivation behind pixi-pack was to create a conda-pack alternative that does not have the same reproducibility issues as conda-pack. It also aims to allow cross-platform building packs, so you can create a pack for win-64 on a linux-64 system.

💿 Installation

You can install pixi-pack and pixi-unpack using pixi:

pixi global install pixi-pack
pixi global install pixi-unpack

Or using cargo:

cargo install --locked --git https://github.com/quantco/pixi-pack.git

Or by downloading our pre-built binaries from the releases page.

Instead of installing pixi-pack and pixi-unpack globally, you can also use pixi exec to run pixi-pack in a temporary environment:

pixi exec pixi-pack
pixi exec pixi-unpack environment.tar

[!NOTE] You can also write pixi pack (and pixi unpack) if you have pixi-pack and pixi-unpack installed globally.

🎯 Usage

pixi-pack: Packing an environment

With pixi-pack, you can pack a conda environment into a environment.tar file:

pixi-pack --environment prod --platform linux-64 pixi.toml

This will create an environment.tar file that contains all conda packages required to create the environment.

# environment.tar
| pixi-pack.json
| environment.yml
| channel
|    ├── noarch
|    |    ├── tzdata-2024a-h0c530f3_0.conda
|    |    ├── ...
|    |    └── repodata.json
|    └── linux-64
|         ├── ca-certificates-2024.2.2-hbcca054_0.conda
|         ├── ...
|         └── repodata.json

pixi-unpack: Unpacking an environment

With pixi-unpack environment.tar, you can unpack the environment on your target system. This will create a new conda environment in ./env that contains all packages specified in your pixi.toml. It also creates an activate.sh (or activate.bat on Windows) file that lets you activate the environment without needing to have conda or micromamba installed.

$ pixi-unpack environment.tar
$ ls
env/
activate.sh
environment.tar
$ cat activate.sh
export PATH="/home/user/project/env/bin:..."
export CONDA_PREFIX="/home/user/project/env"
. "/home/user/project/env/etc/conda/activate.d/activate_custom_package.sh"

Cross-platform packs

Since pixi-pack just downloads the .conda and .tar.bz2 files from the conda repositories, you can trivially create packs for different platforms.

pixi-pack --platform win-64

[!NOTE] You can only pixi-unpack a pack on a system that has the same platform as the pack was created for.

Named platforms

If your pixi.toml defines named platforms (requires a lockfile in version 7 or later), you can pack them by name.

platforms = [{ name = "jetson", platform = "linux-aarch64", cuda = "13" }]

pixi-pack --platform jetson

You can still pass the underlying conda platform (e.g. linux-aarch64) as long as only one named platform uses it; otherwise pixi-pack asks you to use one of the platform names instead.

Self-extracting binaries

You can create a self-extracting binary that contains the packed environment and a script that unpacks the environment. This can be useful if you want to distribute the environment to users that don't have pixi-unpack installed.

# unix
$ pixi-pack --create-executable
$ ls
environment.sh
$ ./environment.sh
$ ls
env/
activate.sh
environment.sh

# windows
PS > pixi-pack --create-executable
PS > ls
environment.ps1
PS > .\environment.ps1
PS > ls
env/
activate.sh
environment.ps1

Custom pixi-unpack executable path

When creating a self-extracting binary, you can specify a custom path or URL to a pixi-unpack executable to avoid downloading it from the default location.

You can provide one of the following as the --pixi-unpack-source:

• a URL to a pixi-unpack executable like https://my.mirror/pixi-pack/pixi-unpack-x86_64-unknown-linux-musl
• a path to a pixi-unpack binary like ./pixi-unpack-x86_64-unknown-linux-musl

Example Usage

Using a URL:

pixi-pack --create-executable --pixi-unpack-source https://my.mirror/pixi-pack/pixi-unpack-x86_64-unknown-linux-musl

Using a path:

pixi-pack --create-executable --pixi-unpack-source ./pixi-unpack-x86_64-unknown-linux-musl

[!TIP] The produced executable is a simple shell script that contains both the pixi-unpack binary as well as the packed environment.

Inject additional packages

You can inject additional packages into the environment that are not specified in pixi.lock by using the --inject flag:

pixi-pack --inject local-package-1.0.0-hbefa133_0.conda pixi.toml

This can be particularly useful if you build the project itself and want to include the built package in the environment but still want to use pixi.lock from the project. Before creating the pack, pixi-pack will ensure that the injected packages' dependencies and constraints are compatible with the packages in the environment.

PyPi support

You can also pack PyPi wheel packages into your environment. pixi-pack only supports wheel packages and not source distributions. If you happen to use source distributions, you can ignore them by using the --ignore-pypi-non-wheel flag. This will skip the bundling of all PyPi source distributions.

The --inject option also supports wheels.

pixi-pack --ignore-pypi-non-wheel --inject my_webserver-0.1.0-py3-none-any.whl

[!WARNING] In contrast to injecting from conda packages, we cannot verify that injected wheels are compatible with the target environment. Please make sure the packages are compatible.

Mirror and S3 middleware

You can use mirror middleware by creating a configuration file as described in the pixi documentation and referencing it using --config.

[mirrors]
"https://conda.anaconda.org/conda-forge" = ["https://my.artifactory/conda-forge"]

If you are using S3 in pixi, you can also add the appropriate S3 config in your config file and reference it.

[s3-options.my-s3-bucket]
endpoint-url = "https://s3.eu-central-1.amazonaws.com"
region = "eu-central-1"
force-path-style = false

Setting maximum number of parallel downloads

[concurrency]
downloads = 5

Use pixi-pack --config config.toml to use the custom configuration file. See pixi docs for more information.

Cache downloaded packages

You can cache downloaded packages to speed up subsequent pack operations by using the --use-cache flag:

pixi-pack --use-cache ~/.pixi-pack/cache

This will store all downloaded packages in the specified directory and reuse them in future pack operations. The cache follows the same structure as conda channels, organizing packages by platform subdirectories (e.g., linux-64, win-64, etc.).

Using a cache is particularly useful when:

• Creating multiple packs with overlapping dependencies
• Working with large packages that take time to download
• Operating in environments with limited bandwidth
• Running CI/CD pipelines where package caching can significantly improve build times

Unpacking without pixi-unpack

If you don't have pixi-unpack available on your target system, you can still install the environment if you have conda or micromamba available. Just unarchive the environment.tar, then you have a local channel named pixi-unpack on your system where all necessary packages are available. Next to this local channel, you will find an environment.yml file that contains the environment specification. You can then install the environment using conda or micromamba:

tar -xvf environment.tar
micromamba create -p ./env --file environment.yml
# or
conda env create -p ./env --file environment.yml

[!NOTE] The environment.yml and repodata.json files are only for this use case, pixi-unpack does not use them.

[!NOTE] Both conda and mamba are always installing pip as a side effect when they install python, see conda's documentation. This is different from how pixi works and can lead to solver errors when using pixi-pack's compatibility mode since pixi doesn't include pip by default. You can fix this issue in two ways:

• Add pip to your pixi.lock file using pixi add pip.
• Configuring conda (or mamba) to not install pip by default by running conda config --set add_pip_as_python_dependency false (or by adding add_pip_as_python_dependency: False to your ~/.condarc)

Build attestations

Build attestations are uploaded to GitHub Artifact Attestations. You can verify a binary using the gh CLI:

gh attestation verify --repo Quantco/pixi-pack pixi-pack-<architecture>