Awesome Python

Libraries for handling visual and auditory media:

Image Processing

• Core Manipulation: pillow (PIL fork), pyvips (fast/low memory), scikit-image (scientific processing), wand (ImageMagick bindings).
• Specialized: pymatting (alpha matting), python-barcode (barcodes), python-qrcode (QR codes), thumbor (on-demand resizing/cropping).

Tools for managing versions, environments, and dependencies:

Environment Management

• KillPy: Detect and clean unused environments/pipx packages.
• pyenv: Python version management.
• pyenv-win: Pyenv for Windows.
• uv: Extremely fast version, package, and project manager (Rust-based).
• virtualenv: Create isolated environments.

Package Management

• conda: Cross-platform, Python-agnostic binary manager.
• hatch: Modern project manager for environments, builds, and publishing.
• pip: Standard package installer.
• pipx: Install and run Python applications in isolated environments (like npx).
• poetry: Dependency management and packaging.
• uv: (Also listed under environment management) Fast Rust-based manager.

For real-time features, use these libraries:

• autobahn-python: WebSocket & WAMP for Twisted and asyncio.
• channels: Asynchronous capabilities specifically for Django.
• flask-socketio: Socket.IO integration for Flask.
• picows: High-performance WebSocket clients/servers with frame-level interface.
• websockets: Simple and correct library for building WebSocket servers and clients.

Frameworks for building and training neural networks:

• jax: High-performance numerical computing with JIT and auto-diff.
• keras: High-level API supporting JAX, TensorFlow, and PyTorch.
• pytorch: Dynamic neural networks with GPU acceleration.
• pytorch-lightning: High-level framework for scaling PyTorch training.
• tensorflow: Google's widely-used deep learning framework.
• stable-baselines3: Reinforcement learning algorithms implemented in PyTorch.

Depending on your existing stack, you can build RESTful or GraphQL APIs using:

Django-based APIs

• django-modern-rest: Modern REST with support for msgspec and pydantic.
• django-ninja: Fast, type-hint based REST framework using Pydantic.
• django-rest-framework: A powerful, flexible toolkit for web APIs.
• strawberry-django: Integration for Strawberry GraphQL.

Flask-based APIs

• apiflask: Lightweight API framework based on Flask and Marshmallow.

Framework Agnostic APIs

• connexion: Spec-first framework using OpenAPI specifications.
• falcon: High-performance framework for cloud APIs.
• fastapi: Modern, fast framework using standard Python type hints.
• sanic: High-performance web server and framework.
• strawberry: GraphQL library leveraging Python type annotations.
• webargs: Library for parsing HTTP request arguments.

For building games in Python, the following frameworks are recommended:

• arcade: Modern framework for graphics and sound.
• panda3d: 3D game engine developed by Disney.
• py-sdl2: ctypes wrapper for SDL2.
• pygame: Set of modules for general game writing.
• pyopengl: ctypes bindings for OpenGL.
• renpy: Visual Novel engine.

For non-blocking and parallel execution, use these libraries:

• anyio: High-level async framework (works with asyncio or trio).
• asyncio: (Standard Library) Core async I/O and event loop.
• concurrent.futures: (Standard Library) High-level interface for asynchronous execution.
• gevent: Coroutine-based networking using greenlet.
• multiprocessing: (Standard Library) Process-based parallelism.
• trio: Friendly async concurrency and I/O.
• twisted: Event-driven networking engine.
• uvloop: Ultra-fast asyncio event loop.

Several libraries are available for generating documentation for Python projects:

• diagrams: Uses 'Diagram as Code' approach.
• mkdocs: A Markdown-friendly documentation generator.
• pdoc: An Epydoc replacement that automatically generates API documentation for Python libraries.
• sphinx: A widely used Python documentation generator (see also awesome-sphinxdoc).