python: tooling selection

[AlexanderLanin]

The python ecosystem consists of a bunch of different tools. We need to agree which ones we want to use.

Rough overview by ChatGPT

no guarantees of correctness or completeness

Linting & Formatting

• ruff – An extremely fast Python linter and code formatter. [GitHub Stars: 13.2K]
• flake8 – A traditional linter with extensive plugin support. [GitHub Stars: 2.6K]
• black – An uncompromising code formatter. [GitHub Stars: 32.5K]

Type Checking

• pyright – A static type checker for Python. [GitHub Stars: 10.5K]
• mypy – Optional static typing for Python. [GitHub Stars: 15.2K]

Runtime Type Checking

• beartype – A runtime type checker for Python. [GitHub Stars: 2.3K]

Security & Dependency Analysis

• bandit – A tool designed to find common security issues in Python code. [GitHub Stars: 5.9K]
• safety – Checks Python dependencies for known security vulnerabilities. [GitHub Stars: 1.6K]

Test Coverage & Complexity Analysis

• coverage.py – Measures code coverage of Python programs. [GitHub Stars: 2.2K]
• radon – A Python tool that computes various metrics from the source code. [GitHub Stars: 1.8K]
• pytest-cov – Pytest plugin for measuring coverage. [GitHub Stars: 1.1K]

Testing Frameworks

• pytest – A framework that makes building simple and scalable test cases easy. [GitHub Stars: 10.5K]
• unittest – (Part of Python's standard library; no separate GitHub repository)
• hypothesis – A library for property-based testing. [GitHub Stars: 6.6K]

Static Analysis & Dependency Management

• pylint – A static code analyzer for Python. [GitHub Stars: 4.7K]
• isort – A Python utility to sort imports. [GitHub Stars: 5.6K]
• pip-audit – A tool for scanning Python environments for packages with known vulnerabilities. [GitHub Stars: 1.1K]

Documentation Quality

• docformatter – Formats docstrings to follow PEP 257. [GitHub Stars: 520]
• pydocstyle – A docstring style checker. [GitHub Stars: 1.4K]

Performance Profiling

• scalene – A high-performance, high-precision CPU, GPU, and memory profiler for Python. [GitHub Stars: 6.2K]

[AlexanderLanin]

Deeper analysis by @sameer6989: #184 (comment)

[AlexanderLanin]

We'll start a discussion in Chat-Access, #score-python-community before moving to concrete ideas and comments here.

[AlexanderLanin]

First quick conclusion:

• Everyone is familiar with a different toolstack
• All the tools are in use by someone
• We don't want a lengthy evaluation of the "best" tools

[AlexanderLanin]

My personal opinionated proposal:

1. pytest
2. ruff
3. pyright
4. pytest-cov
5. pylint
6. beartype

[MaximilianSoerenPollak]

I think this can be closed as the tooling selection has been made, now we can move towards implementing it which is tracked in #242 .

[AlexanderLanin]

@MaximilianSoerenPollak did we document the decision anywhere?
#242 contains some sub-issues which look almost random.