Virtual environments and dependencies

A virtual environment gives a project its own Python executable and installed packages. Create one with python -m venv .venv, using your working Python 3 command in place of python if needed. You can run its interpreter directly without activating: on Windows PowerShell use ./.venv/Scripts/python.exe -m pip --version; on macOS or Linux use ./.venv/bin/python -m pip --version. Standard-library modules need no pip install.

Worked example

import sys
print(f"Interpreter: {sys.executable}")
print(f"Inside a virtual environment: {sys.prefix != sys.base_prefix}")

With your base Python:
Interpreter: <path to your Python>
Inside a virtual environment: False

With the .venv Python:
Interpreter: <path to .venv Python>
Inside a virtual environment: True

Why this works

The interpreter path varies by machine. The second line is False for a base interpreter and True for a virtual environment. Run this script with the interpreter inside .venv to see the difference.

Another case

from importlib.util import find_spec
print(find_spec("json") is not None)
print(find_spec("not_installed_example") is not None)

True
False

find_spec() checks whether the interpreter you are running can import a module. After installing a package into .venv, run this check with the .venv interpreter to confirm the package is available in that environment.

Practice

Create a venv in a disposable folder, then run this example once with the base interpreter and once with the interpreter inside .venv.

Try another case

You install a package with pip while .venv is active, then run your script with the base Python in another terminal. Will the import work?