Skip to content

Latest commit

 

History

History
309 lines (202 loc) · 6.88 KB

README.md

File metadata and controls

309 lines (202 loc) · 6.88 KB

pixi

pixi: package management made easy.

pixi is a cross-platform, multi-language package manager and workflow tool built on the foundation of the conda ecosystem. It's entirely written in Rust, and supports, and can spins up development environments quickly on Linux, Windows, and MacOS for multi-languages, such as Python, R, C/C++, Rust, Ruby, and so on.

pixi provides a more modern experience for developers managing Python environments. It provides a clean and simple command-line interface, centralizes the dependency management by using TOML file, and supports locking dependencies for reproducible builds.

Moreover, pixi is ready for production!

How to install it?

Linux & MacOS

For Linux and MacOS users, you can install pixi by running the following command in your terminal:

$ curl -fsSL https://pixi.sh/install.sh | bash

This command automatically download, extract, and move the pixi package to ~/.pixi/bin. Depending on your OS system and shell environment, it would update your ~/.bash, ~/.zsh, or ~/.bash_profile to include ~/.pixi/bin in your PATH, making pixi command invokable from anywhere.

You may need to restart or source your shell, for example,

$ source ~/.zsh`

To enable autocompletion for your shell,

For Bash users, you can add the following to your ~/.bashrc:

eval "$(pixi completion --shell bash)"

For Zsh users, you can add the following to your ~/.zshrc:

eval "$(pixi completion --shell zsh)"

Windows

For Windows PowerShell users,

$ iwr -useb https://pixi.sh/install.ps1 | iex

For Windowns winget users,

$ winget install prefix-dev.pixi

The installation location of pixi on Windows is LocalAppData/pixi/bin.

To enable autocompletion for your shell environment,

for PowerShell,

$ (& pixi completion --shell powershell) | Out-String | Invoke-Expression

for Fish, you can add the following into ~/.config/fish/config.fish:

pixi completion --shell fish | source

Beside for the instructions above, pixi could be installed and configured in various other methods as well. For example, you can use Homebrew for MacOS user, use msi Windows installer for Windows user, and can install from the source. For more detailed information about installation and autocompletion, see the official documentation - https://pixi.sh/latest/#installation.

After the installation, you can check by executing the following command:

$ pixi --version

How to use it?

1. Create a virtual environment

pixi supports two manifest formats: pyproject.toml and pixi.toml. We use pyproject.toml here, as it is the most common format for Python projects.

Creating a new project,

$ pixi init <project-name> --format pyproject

What's inside?

$ tree pixi-demo-project
pixi-demo-project
├── pyproject.toml
└── src
    └── pixi_demo_project
        └── __init__.py

2 directories, 2 files

Enter the project directory, then you can create the default environment by running the following command:

$ pixi install

A .pixi directory would be created under your project, which contains a envs diretory with the default environment.

$ tree .pixi -L 3
.pixi
└── envs
    └── default
        ├── bin
        ├── conda-meta
        ├── include
        ├── lib
        ├── man
        ├── share
        ├── shared
        ├── ssl
        ├── x86_64-conda_cos6-linux-gnu
        └── x86_64-conda-linux-gnu

12 directories, 0 files

If you like to create another environment, e.g. test environment, for your project. You can run the following command:

$ pixi project environment add --feature test test
✔ Added environment test

Your pyproject.toml would be updated automatically with the following content:

[tool.pixi.environments]
test = ["test"]

2. Activate the virtual environment

There are multiple options to activate the virtual environment,

Use pixi shell

Open a shell with the environment activated,

$ pixi shell

To activate the test environment,

$ pixi shell -e test

Use `pixi shell-hook

This command would print the command to activate the environment in your current shell.

$ eval "$(pixi shell-hook)"

Use pixi run

This helps run a command in the environment. For example,

$ pixi run python

3. Manage Python packages

Add a single package to the project,

$ pixi add <package-name>

For example,

$ pixi add requests

If you need pytest for test environments, then you can run

$ pixi add --feature test pytest

To remove a dependency from the project,

$ pixi remove <package-name>

If you need to install all Python packages in pyproject.toml into the default environment, then

$ pixi install

If to install into the test environment, then

$ pixi install -e test

4. Deactivate the virtual environment

To exit the activated shell environment,

$ exit

5. Remove the virtual environment

There is not a direct delete or remove command to manage the environment, you can do this:

$ rm -rf .pixi/envs/<environment-name>

Different Python versions?

By default, when you run:

$ pixi install

The latest Python version fron conda-forge would be applied. As the time of writting, the latest Python version is python 3.13,

$ pixi run python

Python 3.13.1 | packaged by conda-forge | (main, Dec  5 2024, 21:23:54) [GCC 13.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>>

To install different Python versions, you can do this, for example:

For Python 3.12,

$ python add python==3.12

Then you can validate by running:

$ pixi run python
Python 3.12.0 | packaged by conda-forge | (main, Oct  3 2023, 08:43:22) [GCC 12.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>>

If you need multiple versions of Python in your project for testing, then you can manage multiple version through pixi features.

To avoid conflicts, you need to allow multiple versions by adding:

$ pixi add python

Then, can run the following:

$ pixi add --feature py311 python==3.11
$ pixi project add environment --feature py311 py311
$ pixi shell -e py311

After the shell is activated,

$ python
Python 3.11.0 | packaged by conda-forge | (main, Jan 14 2023, 12:27:40) [GCC 11.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>>

For more information about pixi, please refer to the official documentation https://pixi.sh/latest/.

Happy coding!