15 lines
1.2 KiB
Markdown
15 lines
1.2 KiB
Markdown
---
|
|
description: Ultra-dense Python code generation rules to save output tokens
|
|
globs: "*.py"
|
|
---
|
|
|
|
# Python Token Optimization
|
|
|
|
- **Use Syntactic Sugar:** Prioritize list comprehensions, dict comprehensions, and ternary operators (`x if condition else y`) to keep code on a single line.
|
|
- **Built-in Libraries First:** Use standard libraries (`pathlib`, `json`, `subprocess`, `asyncio`) instead of introducing heavy external dependencies unless already in `requirements.txt`.
|
|
- **Type Hinting:** Do NOT add type hints (`def func(x: int) -> str:`) unless the existing file strictly uses them. Type hints consume significant tokens.
|
|
- **No Format Duplication:** When modifying scripts (like video downloaders), provide only the modified function or class method. Never output the `if __name__ == "__main__":` block or argument parsing logic if they haven't changed.
|
|
- **No Docstrings:** Strictly forbid writing `"""docstrings"""` or `# comments` explaining the logic.
|
|
- **Preserve yt-dlp Options:** Never rewrite, duplicate, or expand the `ydl_opts` configuration dictionary or custom extraction options. If changes are unrelated to download options, use a placeholder comment like `# ... existing ydl_opts ...` instead of outputting the full dictionary block.
|
|
|