Virtual Environments

Virtual environments are a handy way to store the dependencies your Python application requires. Creating a virtual environment makes a self-contained folder that includes a full Python installation, allowing you to install dependencies specific to your application (or even a specific version of Python if you want) without relying on the global system version of Python.

As you saw earlier, you can create a virtual environment (or virtualenv for short) by running:

$ python3.9 -m venv env
$ source env/bin/activate
> py -3.9 -m venv env
> .\env\Scripts\Activate

In either case, you’ve now created a folder called venv that includes a full Python install, and your shell is now set up to call this local Python version when you run the python command.

requirements.txt

requirements.txt is a special file that we use to tell pip, the Python package manager, which dependencies to install. The format is simple: you can create one manually by putting each dependency you import into a text file named requirements.txt, one dependency on each line.

Alternatively, you can install each dependency in your virtualenv using pip. Once you’re done (your application stops complaining about missing imports), you can run:

$ pip freeze > requirements.txt

This will automatically generate a requirements.txt file for you, using the specific version of each dependency you installed in your virtualenv.

Later on, when you move your Python application to a new virtualenv, a new computer, or deploy it to the world, you can bring all your dependencies with you with requirements.txt. To install all your requirements again in a new virtual environment you can simply run:

$ pip install -r requirements.txt