Working on a new project its always exciting to jump straight in and get coding without any setup time. However spending a small amount of time to setup the project with the best tools and practices will lead to a standardised and aligned coding experience for developers.

In this article I will go through what I consider to be the best python project setup. Please follow along, or if you prefer to jump straight in, you can use cookiecutter to generate a new project following these standards, install poetry then create a new project.

Poetry: Dependency Management

Poetry is a Python dependency management and packaging system that makes package management easy!

Poetry comes with all the features you would require to manage a project’s packages, it removes the need to freeze and potentially include packages that are not required for the specific project. Poetry only adds the libraries that you require for that specific project.

No more need for the unmanagale requirements.txt file.

Poetry will also add a venv to ensure only the required packages are loaded. with one simple command poetry shell you enter the venv with all the required packages.

Lets get setup with poetry

pip install poetry
poetry init
poetry add <package>
poetry shell
poetry run python

So, thats a few commands! but what do they all do?

  • pip install poetry - Installs the poetry package to you macine
  • poetry init - Adds poetry to an existing project (for a new project use poetry new <projectName>)
  • poetry add <package> - Adds a single or multiple python packages
  • poetry shell - Activates the poetry venv
  • poetry run python - Runs the script within the poetry venv

Black: Code Formatting

black is an uncompromising code formatter in Python. If your code violates pep8 then Black will notify or resolve the issues

Lets get that installed as a development dependency:

poetry add black --dev

We also need to add some additional configuration for Black to the end of pyproject.toml

line-length = 88
target_version = ['py38']
include = '\.pyi?$'
exclude = '''
      \.eggs         # exclude a few common directories in the
    | \.git          # root of the project
    | \.hg
    | \.mypy_cache
    | \.tox
    | \.venv
    | _build
    | buck-out
    | build
    | dist

These settings can be changed to your preferences, for example I like the line length to be 88, but you may prefer this shorter / longer.

To use black you can run the following command:

black . --check

This will check the formatting of the files in the current directory and its subfolders, if you remove the --check option, it will automatically reformat your python code.

Example Black Output

isort: Import Sorting

isort is a Python library that automatically sorts imported libraries alphabetically and separates them into sections and types.

Lets get that installed as a development dependency:

poetry add isort --dev

isort and black dont get along, their configurations conflict with eachother, so to get around this issue we need to add some configuration to the end of pyproject.toml:

profile = "black"
force_sort_within_sections = true
known_first_party = [
forced_separate = [

Adding the profile = "black" option ensures that iSort respects changes made by Black, It is also advisable to add the folders for known_first_party files, this enables iSort to group those imports together in order.

To use isort run the following command:

isort . --diff

This will check the order of the imports and let you know if it is correct, if you would like isort to automatically fix the ordering, remove the --diff option.

flake8: Style Enforcement

flake8 is a python tool that checks the style and quality of your Python code. It checks for various issues not covered by black.

Lets get this added to our project:

poetry add flake8 --dev

flake8 also has some configuration that is recommended with black, create a new file called .flake8 and place the below configuration:

max-line-length = 88
max-complexity = 15
exclude = build/*
extend-ignore =
    # See
ignore = E203, E266, E501, W503, W605
select = B,C,E,F,W,T4

This configuration will ensure that stype errors that conflict with black will be ignored.

To use flake8 you can run the below command:

flake8 . --fix

This will run flake8 and fix any issues on all files in the current directory and subdirectories, if you just want to see the issues remove the --fix option.

Example Flake8 Output

MyPy: Static Types Checker

Mypy is an optional static type checker for Python that aims to combine the benefits of dynamic (or “duck”) typing and static typing.

MyPy does require that the static types are installed for each library, if a library has no static types that will cause mypy to error.

poetry add mypy --dev

Additional configuration can be added to the pyproject.toml file if required similar to below:

python_version = "3.8"
warn_return_any = true
warn_unused_configs = true

disallow_untyped_defs = true

To use mypy simply enter the following command:

mypy .

Interrogate: DocString standardisation

interrogate checks your codebase for missing docstrings.

Docstrings provide the ability to automatically document and also assist developers allowing then to quickly and easily see what a specific class or function is used for.

To install interrogate, type:

poetry add interrogate --dev

Additional configuration for the pyproject.toml file:

ignore-init-method = true
ignore-init-module = false
ignore-magic = false
ignore-semiprivate = false
ignore-private = false
ignore-property-decorators = false
ignore-module = true
ignore-nested-functions = false
ignore-nested-classes = true
ignore-setters = false
fail-under = 95
exclude = ["", "docs", "build"]
ignore-regex = ["^get$", "^mock_.*", ".*BaseClass.*"]
verbose = 0
quiet = false
whitelist-regex = []
color = true

To run interrogate use the following command:

interrogate -vv

Remove the -vv to just see a success or fail message without a list of files

Example Interrogate Output

Pre-Commit hooks

Now we have all of these tests, but we dont want to run them manually every time we make changes to code. This is where pre-commit hooks come into play.

Pre-commit hooks allow you to run multiple checks against code before git commit will be applied, if any of the tests fail, the commit will not apply until the issues raised are resolved.

This feature is great for a few reasons:

  1. You dont need to remember to run all of the above manually each time you wish to check code
  2. Github Actions based on code quality should continue to succeed

Setup pre-commit hooks

First we create the configuration file in root .pre-commit-config.yaml:

  - repo:
    rev: 5.8.0
      - id: isort
        name: isort (python)
  - repo: local
      - id: black
        name: black
        stages: [commit]
        language: system
        entry: black
        types: [python]
  - repo:
    rev: 4.0.1
      - id: flake8
        additional_dependencies: [flake8-bugbear]
  - repo:
    rev: v1.2.0
      - id: conventional-pre-commit
        stages: [commit-msg]
        args: [] # optional: list of Conventional Commits types to allow
  - repo:
    rev: 1.5.0
      - id: interrogate

Install pre-commit & tell pre-commit to register our config:

poetry install pre-commit --dev
pre-commit install -t pre-commit -t commit-msg

Now when you run a commit you will see each hook running, this will then show any errors prior to committing, you can then fix the issues and try the commit again.

You can also see I have conventional-pre-commit applied with the -t commit-msg tag this enforces the use of conventional commit messages for all commits, ensuring that our commit messages all follow the same standard.

Example pre-commit Output

Final Thoughts

This method of utilising cookie cutter, and pre-commit hooks has saved me a lot of time, I think there is more to be explored with pre-commit hooks such as adding tests for my code etc. that will come with time on my development journey.

With these methods I know my commit messages are tidy, and my code is cleaner than before its a great start with more to come.

I also execute these as github actions on my projects, that way anyone else who contributes but doesnt install the pre-commit hooks will be held accountable to resolve any issues prior to merging their pull requests.

Hopefully some of this information was useful for you, If you have any questions about this article and share your thoughts head over to my Discord.