Release v0.1.1 (What’s new?).

---

Welcome to shai_py Documentation

In the AI-driven development era, tools must be designed for machine consumption first. shai-py (Sanhe AI Python tools) embraces this philosophy by packaging Python development utilities as a CLI-first library, making it effortless for AI agents like Claude Code to invoke sophisticated workflows. Instead of writing one-off scripts scattered across agent skills, we consolidate battle-tested logic into a versioned Python package that AI can invoke with a single uvx command—no installation, no environment pollution, just instant execution.

This approach solves the fundamental challenge of AI tool integration: how to provide powerful, testable, and maintainable utilities without cluttering agent contexts with implementation details. By exposing functionality through clean CLI interfaces (uvx shai-py project-info, uvx shai-py test-path), AI agents can focus on orchestration while developers maintain business logic in a single, version-controlled codebase. The result is elegant, reproducible, and scales beautifully from simple project introspection to complex development automation.

Architecture: Subcommand Design Pattern

This project uses a Subcommand Delegation Pattern that cleanly separates business logic from CLI interface, enabling independent testing and maintainability.

Key Components:

1. Subcommand Modules (shai_py/subcmd/<subcommand>.py): Each module implements a single CLI subcommand. The module must define a main() function containing all business logic. The module-level __doc__ string serves as CLI help text.

2. CLI Aggregator (shai_py/cli.py): The Cli class exposes each subcommand as a method that delegates to the corresponding module’s main() function. Method docstrings are inherited from the subcommand module’s __doc__.

3. Test Files (tests/subcmd/test_subcmd_<subcommand>.py): Tests import and invoke the main() function directly, enabling unit testing without CLI overhead.

Pattern Benefits:

• Business logic is testable without CLI framework involvement

• Documentation lives with implementation (single source of truth)

• Adding new subcommands requires only: create module with main(), add method to Cli class

Reference Implementation:

• Subcommand: shai_py/subcmd/detect_python_project_metadata.py (see main() function)

• CLI integration: shai_py/cli.py (see Cli.project_info() method)

• Test example: tests/subcmd/test_subcmd_detect_python_project_metadata.py

Install

shai_py is released on PyPI, so all you need is to:

$ pip install shai-py

To upgrade to latest version:

$ pip install --upgrade shai-py

Table of Content

• Commands

About the Author

(\ (\
( -.-)o
o_(")(")

Sanhe Hu is a seasoned software engineer with a deep passion for Python development since 2010. As an author and maintainer of 150+ open-source Python projects, with over 15 million monthly downloads, I bring a wealth of experience to the table. As a Senior Solution Architect and Subject Matter Expert in AI, Data, Amazon Web Services, Cloud Engineering, DevOps, I thrive on helping clients with platform design, enterprise architecture, and strategic roadmaps.

Talk is cheap, show me the code:

• My Github: https://github.com/MacHu-GWU

API Document

• by Name

• by Structure