modern-python-toolset.md

The Modern Python Toolset

A collection of tools that are so much better than their predecessors they should replace them. Everyone should use these tools as part of their development process.

• uv (environment manager, better than pyenv pip pipenv poetry conda etc)
• pyright (or maybe mypy, for type checking)
• ruff (linting and formatting, better than pylint, black, flake8, isort)
• pytest (unit testing, better than unittest)
• pydantic (data validation that leverages type hints)
• pyproject.toml (for project organization, better than setup.py, setup.cfg)
• sqlalchemy (abstraction for database calls)
• github actions (to define automatic continuous deployment (CI/CD))
• pre-commit (framework to automate steps before commiting code)
• typer (command line arguments that leverages type hints)

uv

(environment manager)

• 10x-100x FASTER. Replaces pip, pyenv, pipenv, pipx, poetry, conda, etc. Works with pyproject.toml
• Generates and updates pyproject.toml and uv.lock lockfile (versions) and .venv uv init; uv init myproject; uv init --lib mylibrary; uv init --package mypackage uv add httpx; uv add 'httpx>0.1.5' uv run uv python install (eg: uv python install 3.6 3.9 3.12) uv run --python 3.12 uv python pin 3.12 uv venv --python 3.12 uv lock (not needed, uv run will do it automatically) uv sync (not needed, uv run will do it automatically) uv venv (create venv); source .venv/bin/activate uv pip install, uv pip sync requirements.txt (not needed, use uv add) uv add --script main.py "requests<3" "rich" (isolated scripts); uv run main.py uv tool install (install command line tools like ruff); uv tool list uv tool run (aka uvx, eg: uvx ruff)

pyright

(or maybe mypy, for type checking)

ruff

(linting and formatting, better than pylint, black, flake8, isort) ruff check ruff check --fix ruff format

pytest

(unit testing, better than unittest) test/test_sample.py:

import my_code
def test_test1():
  # val=call mycode with test data
  assert(val==5)

pydantic

• data validation & serialization, integrates well with dataclasses, type hints (mypy/pyright), SQLAlchemy and FastAPI

from datetime import datetime
from typing import List, Optional
from pydantic import BaseModel

class User(BaseModel):
    id: int
    name: str = 'John Doe'
    signup_ts: Optional[datetime] = None
    friends: List[int] = []

https://docs.pydantic.dev/latest/#why-use-pydantic

pyproject.toml

(for project organization, better than setup.py, setup.cfg)

[project]
name = "spam-eggs"
version = "2020.0.0"
dependencies = [
  "httpx",
  "gidgethub[httpx]>4.0.0",
  "django>2.1; os_name != 'nt'",
  "django>2.0; os_name == 'nt'",
]
requires-python = ">= 3.8"

sqlalchemy

(abstraction for database calls)

• abstract different dialects of sql
• Core: schema, command-oriented
• ORM (Object Relational Mapper) classes <-> tables: domain, state-oriented
• Core vs ORM: You choose how much abstraction is done for you

Alembic

github actions

• to define automatic continuous deployment (CI/CD)
• build, test & deploy
• on PUSH, run series of tests and commands to test package & deploy

pre-commit

• framework to automate steps before commiting code
• like github actions, can run tests and formatters but locally every time you commit

typer

(command line arguments that leverages type hints)

Others...

There are many other excellent modern libraries and frameworks that have made development a delight compare to 'the old way', but none that have definitely overthrown what came before to make it on this list. Some notable ones are:

• For web backends: FastAPI for Rest APIs; Flask or Django for more 'batteries included' backends.
• HTMX for web frontends: Easy way to generate dynamic front ends without much Javascript coding. Very opinionated and different than a traditional React or Vue front-end.
• Polar instead of Panda for data management. Polar has a better interface, but Panda is compatible with more related libraries.