AgentSkillsCN

python-packaging

打包并发布 Python 项目。

SKILL.md
--- frontmatter
name: python-packaging
description: 打包和发布 Python 项目。

Python Packaging

打包和发布 Python 项目。

When to Invoke

  • •创建可安装的包
  • •发布到 PyPI
  • •设置开发环境
  • •管理项目依赖
  • •版本发布

Input Format

yaml
project_path: "./myproject"
package_name: "my-awesome-package"
build_tool: "poetry"

Output Format

yaml
setup_config: |
  [build-system]
  requires = ["poetry-core>=1.0.0"]
  build-backend = "poetry.core.masonry.api"
  
  [tool.poetry]
  name = "my-awesome-package"
  version = "0.1.0"
  description = "A awesome package"
  authors = ["Your Name <you@example.com>"]
  readme = "README.md"
  license = "MIT"
  
  [tool.poetry.dependencies]
  python = "^3.9"
  requests = "^2.28.0"
  
  [tool.poetry.group.dev.dependencies]
  pytest = "^7.0"
  black = "^22.0"
  
  [tool.poetry.scripts]
  my-cli = "my_package.cli:main"

build_script: |
  #!/bin/bash
  # 构建和发布
  poetry build
  poetry publish

publish_guide:
  - "注册 PyPI 账号"
  - "配置 API Token"
  - "运行 poetry publish"
  - "验证包安装"

check_list:
  - "README.md 完整"
  - "LICENSE 文件存在"
  - "版本号已更新"
  - "测试通过"
  - "CHANGELOG 已更新"

Examples

Example 1: 库包

Input: 工具库

Output:

  • •pyproject.toml 配置
  • •入口点定义
  • •依赖管理
  • •发布流程

Example 2: 应用包

Input: CLI 工具

Output:

  • •控制台脚本配置
  • •数据文件包含
  • •平台特定配置
  • •安装指南

Best Practices

  1. •语义化版本: 遵循 SemVer 规范
  2. •明确依赖: 指定版本范围
  3. •包含测试: 发布前运行测试
  4. •文档完整: README 和文档齐全