Simple Python Dependency Management

While powerful tools like virtualenv, pipenv, conda, and Poetry exist to provide extra advantages for dependency management, I’ve found a simple system that works for the small projects I’ve been working on.

This setup involves built-in features and tools that come bundled with Python by default.

When I start a new project, I create a new directory for all of its files.

mkdir my_python_project

Then, I change into the directory.

cd my_python_project

Next, I create a virtual environment.

python -m venv venv

This creates a new directory called venv. You can call it whatever you want, but I stick with this for all my projects for consistency. The new directory contains all the files and scripts necessary to manage a virtual environment.

Each session, before starting any work, I activate the virtual environment

# macOS and Linux
source venv/bin/activate

# Windows
.\venv\Scripts\activate

You’ll know you did it correctly when your prompt changes to have (venv) at the beginning. Remember, you must execute the above command each time you start a new coding session.

Installing packages works as usual with pip install.

pip install scikit-learn

But each time I use pip install, I freeze the dependencies into requirements.txt immediately after.

pip freeze > requirements.txt

This generates a requirements.txt file which keeps track all the required dependencies.

Any time I’m pulling in code on a new project, whether it’s my own code or someone else’s, I know when I see requirements.txt, that I can install the required packages with pip.

pip install -r requirements.txt

Managing dependencies this way is simple but powerful and it works great for the kinds of projects I work on.