python

khalido 的头像作者 khalido

Python development preferences. Use when writing Python code, CLI tools, APIs, or scripts.

📖 中文导读:Python development preferences. Use when writing Python code, CLI tools, APIs, or scripts. 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

name: python description: Python development preferences. Use when writing Python code, CLI tools, APIs, or scripts.

Python Development

Philosophy

Readable, concise code. Function names and type hints should make purpose clear without verbose docstrings.

Tooling

  • Package manager: uv - new projects: uv init --python 3.14
  • Linting/formatting: ruff
  • Testing: pytest
  • Logging: loguru

Preferred Libraries

Purpose Library
CLI tools typer
CLI output rich (when needed)
HTTP clients httpx
Web APIs fastapi + uvicorn
DataFrames polars
ORM sqlmodel (Pydantic + SQLAlchemy)

Code Style

  • Type hints on function signatures
  • Small, focused functions
  • pathlib for paths
  • dataclasses or Pydantic for structured data
  • SQLite for prototyping, Postgres for prod

Docstrings

Minimal but useful. Skip verbose Google/NumPy style. Answer: "what does this return in practice?"

def fetch_weather(city: str) -> dict[str, float]:
    """Get current weather from OpenWeatherMap API.

    Returns {"temp": 24.5, "humidity": 65.0, "wind_speed": 12.3}
    """

Include: what it does (if name isn't obvious), what's actually returned (not just the type), optional example for complex functions.

Google APIs

Use Google SDKs directly (gspread and similar wrappers are outdated/unmaintained).

uv add google-api-python-client google-auth-httplib2 google-auth-oauthlib

Refs: Sheets quickstart, reading values

Local Testing

Expose local servers via Cloudflare tunnel:

cloudflared tunnel --url http://localhost:8000

Deployment

Railway for FastAPI projects. Volumes required for persistent files (SQLite db, credentials.json).

Ref: Railway LLM docs