Python Environments - Dependecy & Workspace management

Installing Python

• Use pyenv for switch between Python versions without effort.

Commands

• After you have it installed, you can easily install a specific version of Python:

$ pyenv install 3.8.5
$ pyenv install 3.8.6
$ pyenv install 3.9.0
$ pyenv install 3.10.2

$ pyenv versions
* system
  3.8.5
  3.8.6
  3.9.0
  3.10.2

• You can set a "global" Python version to use at any place of your system:

$ pyenv global 3.8.6

$ pyenv versions
  system
  3.8.5
* 3.8.6 (set by /home/will/.pyenv/version)
  3.9.0
  3.10.2

$ python -V
Python 3.8.6

This will not alter anything of your system Python.

• You can also set a "local" Python version to use at your current work directory, like so:

$ pyenv local 3.10.2

$ pyenv versions
  system
  3.8.5
  3.8.6
  3.9.0
* 3.10.2 (set by /home/will/Documents/studies/about-python-environments/.python-version

$ python -V
Python 3.10.2

A file called .python-version will be created to pyenv identify which version of Python you are using at the moment. You can add this file on your .gitignore as well.

Managing Dependencies

Tools for managing Python dependencies and virtual environment.

venv + pip

• Using venv and pip is the simplest way of creating a virtual environment to get things done quickly. They already come installed with the most versions of Python, so you don't need to install anything.

Usage

1. To create a new virtual environment called venv you can just do:

$ python -m venv venv

2. With the created virtual environment you still need to activate it with the command:

$ source venv/bin/activate
(venv)$
(venv)$ deactivate # Deactivate the virtual environment

3. You can check which Python interpreter the system are appointing to and it will be the Python interpreter of the virtual environment:

(venv)$ which python
/home/will/Documents/studies/about-python-environments/venv/bin/python

4. With that, you can install packages local to your projects with pip inside your virtual environment:

(venv)$ pip install requests
...

5. To things working properly, you need a track of each dependency you need to install to run the project, usually you need a file called requirements.txt inside of the root directory of your project. You can manually create one, or just create with pip freeze

(venv)$ python -m pip freeze > requirements.txt
(venv)$ cat requirements.txt
certifi==2022.12.7
charset-normalizer==3.0.1
idna==3.4
requests==2.28.2
urllib3==1.26.14

To only get top-level dependecy (requests==2.28.2), you can check pip-chill for that.

• It's great and easy, but we have a problem: we need to manage each requirements.txt depending on the environment we are in, for example a requirements.txt with the prod packages and a requirements.dev.txt with packages that we only use on development.

poetry

• poetry is the most feature-rich dependency management tool for Python, it comes with a powerful CLI to create and manage Python projects.

• Since Poetry is a more complicated to use, I do recommend reading the oficial docs about it, it's easy when you get it.

Usage

1. You can create a new project with everything configured just with the command: poetry new <project-name>

Or you can just init a poetry project with poetry init

• With that command it will create the following files and folders:

sample-project
├── README.md
├── pyproject.toml
├── sample_project
│   └── __init__.py
└── tests
    ├── __init__.py
    └── test_sample_project.py

• Dependencies and configurations are managed inside the pyproject.toml:

[tool.poetry]
name = "sample-project"
version = "0.1.0"
description = ""
authors = ["William Rodrigues <williamrodrigues2442@gmail.com>"]
readme = "README.md"
packages = [{include = "sample_project"}]

[tool.poetry.dependencies]
python = "^3.10"
flask = "^2.2.2"

[tool.poetry.group.dev.dependencies]
pytest = "^7.2.1"

[build-system]
requires = ["poetry-core"]
build-backend = "poetry.core.masonry.api"

2. To install (add) a new dependency, you can just run: poetry add [--dev] <package-name>

The --dev flag indicates that the dependency is meant to be used in development mode only. Development dependencies are not installed by default.

• For example: poetry add flask, this will download and install Flask inside the virtual environment managed by Poetry. It will add the dependecy to poetry.lock and will add it to dependency list (only top-level dependency) inside pyproject.toml

3. You can also install a package only on development environement: poetry add --dev pytest

4. You can run a command inside de virtual environement with the command: poetry run python -m pytest

5. You can activate the virtual environment with: poetry shell

To deactivate simple run deactivate or exit

• Poetry is great and works well with pyenv too!

Pipenv

@TODO