Let's try: uv for faster Python packages

These days, I have been using uv and here I would like to share about this.

---

What is uv

uv is an alternative to pip for Python with the tag line saying that uv is super fast than pip. As far as I have tried, it’s very fast to install packages, for example, pyspark that takes around 3 minutes with pip but several seconds with uv.

I like how quick it performs for installing packages, plus the template files I don’t have to write from scratch, especially “pyproject.toml” that is necessary for building packages.

For the official repo, it’s here.

GitHub - astral-sh/uv: An extremely fast Python package and project manager, written in Rust.

An extremely fast Python package and project manager, written in Rust. - astral-sh/uv

GitHub

---

Alternatives to uv

• pip: A package installer for Python.
• conda: Open-source package and environment management. Well-known for data science.
• poetry: A tool for dependency management and packaging in Python.

---

Install uv

According to the repo, there are many ways to install uv such as homebrew, curl, pip. I prefer to use devbox.

• devbox
• others

• Start from initializing devbox and generating direnv then add uv.

  # init devbox and direnv
  devbox init
  devbox generate direnv --force
  
  # add uv into devbox
  devbox add uv
  

  Then edit “devbox.json” to activate Python venv when entering the directory by adding row #7.

  {
    "$schema":  "https://raw.githubusercontent.com/jetify-com/devbox/0.14.0/.schema/devbox.schema.json",
    "packages": ["uv@latest"],
    "shell": {
      "init_hook": [
        "echo 'Welcome to devbox!' > /dev/null",
        "source .venv/bin/activate"
      ],
      "scripts": {
        "test": [
          "echo \"Error: no test specified\" && exit 1"
        ]
      }
    }
  }
  

  Read more about devbox and direnv by following the link below.

  Isolated development with direnv & devbox (& gum)

  In this blog we will talk about 3 tools to make a deal with multiple environments easier.

  bluebirz

  

• We can install via curl or pip.

  # On macOS and Linux.
  curl -LsSf https://astral.sh/uv/install.sh | sh
  
  # With pip.
  pip install uv
  

  Other solutions can be found below.

  Installation | uv

  charliermarsh

---

Initialize a project

After installation, we need to initialize a project with uv.

# format
uv init <flags> <path>

# example
uv init
uv init .
uv init test # create folder `test`

Then we can see new files generated.

.
├── .git
├── .gitignore
├── .python-version
├── README.md
├── main.py
└── pyproject.toml

and setup Python with this command.

uv python install <version>

# example
uv python install 3.12

Then create venv.

uv venv

And we can activate and deactivate the venv.

• devbox
• others

• If we install uv through devbox, we don’t have to do it manually.

  devbox should activate it automatically or we would run devbox shell if needed.

• Run the commands.

  # activate
  source .venv/bin/activate
  
  # deactivate
  deactivate
  

---

add/remove packages

Run this command to install dependencies, or the packages, into “pyproject.toml”. It’s equivalent to pip install.

# command
uv add <packages>

# example
uv add pandas

# add from requirements.txt
uv add -r requirements.txt

# remove packages
uv remove <packages>

---

run a script

Execute it by this command.

# command
uv run <script>

# example
uv run main.py # equivalent to `python3 -m main.py`

---

update uv

In order to update uv itself, we can follow this.

• devbox
• others

• For devbox users, we can update through devbox.

  devbox update uv
  

• If we install uv via the standalone installer, we can do this.

  uv self update
  

---

These are basic commands I currently use a lot with uv.

Hope this helps you guys not to wait for so long in development.