Pystub FAQs

Question 1

How does Pystub determine which packages are stubbable?

Accepted Answer

Pystub employs a sophisticated AST analysis algorithm. It parses all import statements, builds a comprehensive import graph from your entry point, and tracks actual package usage. It then provides verdicts: 'stubbable' (safe to replace), 'required' (actively used at runtime), or 'nofollow' (protected imports).

Question 2

What is Pystub and what problem does it solve?

Accepted Answer

Pystub is an MCP server designed to optimize Python executable builds (e.g., with PyInstaller, Nuitka, cx_Freeze). It automatically detects 'stubbable' packages – heavy dependencies that aren't actually used in your application's runtime path – and generates minimal stub code to replace them, significantly reducing your executable's final size.

Question 3

What core features does Pystub offer for analysis and generation?

Accepted Answer

Pystub includes several powerful tools: `analyze` to auto-detect stub candidates, `graph` to visualize your project's import graph, `check` for deep analysis of specific package usage, and `generate` to create minimal stub code directly (without writing to files).

Question 4

How much executable size reduction can I expect with Pystub?

Accepted Answer

By replacing unused, heavy dependency packages with minimal stubs, Pystub can reduce executable size by 50-80%. For example, in a Rich CLI app, it achieved a 13% reduction (1.52 MB) by optimizing pygments and markdown_it.

Question 5

Are there any limitations to Pystub's analysis?

Accepted Answer

Yes, Pystub cannot analyze C extension packages (.pyd/.so) due to their binary nature, nor can it statically analyze dynamic imports (e.g., `importlib.import_module(variable)`). Users may also need to manage PyInstaller hooks separately for stubbed packages, and some stubbable packages might be conservatively marked as 'required' to ensure safety.

Pystub

Pystub

主要功能

使用案例

主要功能

使用案例