Packages and pip

15 min · Beginner

A package is a folder of modules. Python's standard library uses packages, and so do the tools you download. pip is the installer that fetches new packages from the internet.

A package is a folder of modules

Package is a folder of modules, usually with an __init__.py file. You reach a module inside it with dots, such as shop.pricing.

A project with many modules needs order. Put related modules in one folder with a file named __init__.py. That folder is a package, and you reach inside with dots.

Text

1shop/2    __init__.py3    pricing.py4    orders.py56# main.py, next to the shop folder7from shop.pricing import total_with_tax8import shop.orders

Example 1 · Counting with a standard library class

Python

1from collections import Counter2import json3
4status = Counter(["open", "closed", "open"])5print(status.most_common(1))6print(json.dumps({"id": 7, "paid": True}))

Output

[('open', 2)]
{"id": 7, "paid": true}

What this code does

from collections import Counter picks one class out of a larger module.
most_common(1) gives the top item and its count.
json.dumps writes a dictionary as JSON text. Python's True becomes true.

shop.pricing means the module pricing inside the package shop. The standard library has packages too: import json is a module, but from os import path and from collections import Counter reach into bigger ones.

Installing with pip

pip is Python's installer for packages from the Python Package Index. Run it as python -m pip install name so it uses the same Python as your code.

The standard library cannot cover everything. Other people publish packages on the Python Package Index. Run pip in the terminal, not inside a Python file.

Terminal

python -m pip install requestspython -m pip listpython -m pip uninstall requests

How a package reaches your code. PyPI is the online package index. pip install requests downloads the package and installs it into the site-packages folder of your environment. Your code then runs import requests and reads the installed copy.

Example 2 · Is a package available?

Python

import importlib.util
for name in ["json", "surely_not_installed_pkg"]:    found = importlib.util.find_spec(name) is not None    print(name, found)

Output

json True
surely_not_installed_pkg False

What this code does

find_spec looks for a module without loading it. It returns None when nothing is found.
json ships with Python, so it is found. The made-up name is not.

install downloads a package, list shows what is installed, and uninstall removes it. Writing python -m pip makes sure pip belongs to the same Python that runs your code. Install only packages you trust.

Keep each project separate

Virtual environment is a private folder with its own Python packages for one project. It keeps one project's packages from clashing with another's.

Two projects may need different versions of one package. A virtual environment gives each project its own set of packages, and a requirements.txt file records them.

Terminal

python -m venv .venvpython -m pip freeze > requirements.txtpython -m pip install -r requirements.txt

Example 3 · Which Python runs the code?

Python

import sys
print(sys.version_info.major)print(sys.version_info >= (3, 8))

Output

3
True

What this code does

sys.version_info describes the running Python.
.major is the first number of the version. Comparing with a tuple such as (3, 8) checks a minimum version.

The first line creates an environment in the folder .venv; your editor or terminal then activates it. freeze writes installed names and versions to a file. install -r reads that file, so a teammate can rebuild the same set.

Next step

Check your understanding

9 quick questions

Installing with pip

pip is Python's installer for packages from the Python Package Index. Run it as python -m pip install name so it uses the same Python as your code.

The standard library cannot cover everything. Other people publish packages on the Python Package Index. Run pip in the terminal, not inside a Python file.

Keep each project separate

Virtual environment is a private folder with its own Python packages for one project. It keeps one project's packages from clashing with another's.

Two projects may need different versions of one package. A virtual environment gives each project its own set of packages, and a requirements.txt file records them.