Modern Python Tooling in 2026: Package Managers, Type Checkers, and Speed
A complete guide to modern Python tooling in 2026: uv, Poetry, and pip; mypy, Pyright, ty, and Pyrefly; plus free-threading, t-strings, and Polars vs Pandas.
A complete guide to modern Python tooling in 2026: uv, Poetry, and pip; mypy, Pyright, ty, and Pyrefly; plus free-threading, t-strings, and Polars vs Pandas.
An asyncio Queue worker pool ran 100,000 jobs with 100 tasks and flat memory, while gather() added 163 MB. Code for task_done, join, backpressure and shutdown.
Pydantic AI review: v2.68 run offline with real output. Validation retries, docstring tool schemas and hooks work well, but importing it alone takes 1.1 s.
httpx vs aiohttp vs requests, benchmarked locally: aiohttp ran 9-19x faster than httpx async. Plus httpx2, the Pydantic fork the OpenAI SDK now ships.
Python tenacity tutorial with runnable code: exponential backoff, jitter that survives the max cap, Retry-After on 429s, async retries and fast pytest tests.
asyncio TaskGroup vs gather, run side by side: gather leaves siblings running after a failure and silently drops later errors. TaskGroup cancels them all.
FastAPI rate limiting with slowapi, a token-bucket dependency and Redis. Real runs show a fixed window letting 19 requests through a 10-per-minute limit.
Python 3.15's built-in sampling profiler (Tachyon) attaches to a running process by PID, no code changes. Hands-on: attach, flamegraph, wall vs CPU vs GIL.
Ruff vs Flake8, Black, and Pylint in 2026: Astral's speed numbers, what it replaces, where it still loses, and a migration you can finish in an afternoon.
DuckDB in Python turns SQL into a zero-setup analytics engine: query CSV, Parquet, and Pandas DataFrames in-process, join across sources, and add UDFs.