PDM is meant to be a next generation Python package management tool. It was originally built for personal use. If you feel you are going well with Pipenv or Poetry and don't want to introduce another package manager, just stick to it. But if you are missing something that is not present in those tools, you can probably find some goodness in pdm.

• Opt-in PEP 582 support, no virtualenv involved at all.
• Simple and fast dependency resolver, mainly for large binary distributions.
• A PEP 517 build backend.
• PEP 621 project metadata.
• Flexible and powerful plug-in system.
• Versatile user scripts.
• Opt-in centralized installation cache like pnpm.

The majority of Python packaging tools also act as virtualenv managers to gain the ability to isolate project environments. But things get tricky when it comes to nested venvs: One installs the virtualenv manager using a venv encapsulated Python, and create more venvs using the tool which is based on an encapsulated Python. One day a minor release of Python is released and one has to check all those venvs and upgrade them if required.

PEP 582, on the other hand, introduces a way to decouple the Python interpreter from project environments. It is a relatively new proposal and there are not many tools supporting it (one that does is pyflow, but it is written with Rust and thus can't get much help from the big Python community and for the same reason it can't act as a PEP 517 backend).

PEP 582 proposes a project structure as below:

foo
    __pypackages__
        3.8
            lib
                bottle
    myscript.py

There is a __pypackages__ directory in the project root to hold all dependent libraries, just like what npm does. Read more about the specification here.

PDM requires python version 3.7 or higher.

Like Pip, PDM provides an installation script that will install PDM into an isolated environment.

For Linux/Mac

curl -sSL https://raw.githubusercontent.com/pdm-project/pdm/main/install-pdm.py | python3 -

For Windows

(Invoke-WebRequest -Uri https://raw.githubusercontent.com/pdm-project/pdm/main/install-pdm.py -UseBasicParsing).Content | python -

For security reasons, you should verify the checksum of install-pdm.py. The sha256 checksum is: f09accb8a530315be312cf9ce7af987ccb608aa90d3972968d73e7ef7d8c547b

The installer will install PDM into the user site and the location depends on the system:

• $HOME/.local/bin for Unix
• %APPDATA%\Python\Scripts on Windows

You can pass additional options to the script to control how PDM is installed:

usage: install-pdm.py [-h] [-v VERSION] [--prerelease] [--remove] [-p PATH] [-d DEP]

optional arguments:
  -h, --help            show this help message and exit
  -v VERSION, --version VERSION | envvar: PDM_VERSION
                        Specify the version to be installed, or HEAD to install from the main branch
  --prerelease | envvar: PDM_PRERELEASE    Allow prereleases to be installed
  --remove | envvar: PDM_REMOVE            Remove the PDM installation
  -p PATH, --path PATH | envvar: PDM_HOME  Specify the location to install PDM
  -d DEP, --dep DEP | envvar: PDM_DEPS     Specify additional dependencies, can be given multiple times

You can either pass the options after the script or set the env var value.

If you are on MacOS and using homebrew, install it by:

brew install pdm

If you are on Windows and using Scoop, install it by:

scoop bucket add frostming https://github.com/frostming/scoop-frostming.git
scoop install pdm

Otherwise, it is recommended to install pdm in an isolated environment with pipx:

pipx install pdm

Or you can install it under a user site:

pip install --user pdm

With asdf-vm

asdf plugin add pdm
asdf install pdm latest

Initialize a new PDM project

pdm init

Answer the questions following the guide, and a PDM project with a pyproject.toml file will be ready to use.

Install dependencies

pdm add requests flask

You can add multiple dependencies in the same command. After a while, check the pdm.lock file to see what is locked for each package.

Run your script with PEP 582 support

By default, PDM will create .venv in the project root, when doing pdm install on an existing project, as other package managers do. But you can make PEP 582 the default by pdm config python.use_venv false. To enable the full power of PEP 582, do the following steps to make the Python interpreter use it.

Suppose you have a script app.py placed next to the __pypackages__ directory with the following content(taken from Flask's website):

from flask import Flask
app = Flask(__name__)

@app.route('/')
def hello_world():
    return 'Hello World!'

if __name__ == '__main__':
    app.run()

If you are a Bash user, set the environment variable by eval "$(pdm --pep582)". Now you can run the app directly with your familiar Python interpreter:

$ python /home/frostming/workspace/flask_app/app.py
 * Serving Flask app "app" (lazy loading)
 ...
 * Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)

Ta-da! You are running an app with its dependencies installed in an isolated place, while no virtualenv is involved.

For Windows users, please refer to the doc about how to make it work, it also includes a simple explanation of how it works.

Tell people you are using PDM in your project by including the markdown code in README.md:

[![pdm-managed](https://img.shields.io/badge/pdm-managed-blueviolet)](https://pdm.fming.dev)

Awesome PDM is a curated list of awesome PDM plugins and resources.

PEP 582 is a draft proposal which still needs a lot of polishing. For instance, it doesn't mention how to manage CLI executables. PDM makes the decision to put bin and include together with lib under __pypackages__/X.Y.

The recommended way is to prefix your command with pdm run. It is also possible to run CLI scripts directly from the outside. PDM's installer has already injected the package path to the sys.path in the entry script file.

Packages in the local __pypackages__ directory will be loaded before the system-level site-packages for isolation.

You'd better not. The packages installed inside __pypackages__ are OS dependent. Instead, you should keep pdm.lock in VCS and do pdm sync on the target environment to deploy.

This project is strongly inspired by pyflow and poetry.

This project is open sourced under MIT license, see the LICENSE file for more details.