Poetry：现代Python依赖管理工具的实践指南

curl -sSL https://install.python-poetry.org | python3 -

poetry --version

poetry config --list

# 是否自动创建虚拟环境
poetry config virtualenvs.create true
# 是否将虚拟环境保存在项目目录中
poetry config virtualenvs.in-project true

# 禁用Poetry自动创建虚拟环境
poetry config virtualenvs.create false

# 禁用项目内虚拟环境
poetry config virtualenvs.in-project false

poetry new poetry-demo

poetry-demo
├── pyproject.toml  # 项目配置文件
├── README.md       # 项目说明文档
├── src             # 源代码目录
│   └── poetry_demo
│       └── __init__.py
└── tests           # 测试目录
    └── __init__.py

[tool.poetry]
name = "poetry-demo"
version = "0.1.0"
description = "Poetry示例项目"
authors = ["Your Name <your.email@example.com>"]
readme = "README.md"
packages = [{include = "poetry_demo", from = "src"}]

[tool.poetry.dependencies]
python = ">=3.9,<4.0"

[tool.poetry.group.dev.dependencies]
pytest = "^7.0.0"

[build-system]
requires = ["poetry-core>=1.0.0"]
build-backend = "poetry.core.masonry.api"

# 添加运行时依赖
poetry add requests

# 添加特定版本的依赖
poetry add numpy@^1.24.0

# 添加开发环境依赖
poetry add pytest --group dev

# 安装所有声明的依赖
poetry install

# 安装时排除开发依赖
poetry install --without dev

# 更新项目的所有依赖
poetry update

# 只更新特定的依赖包
poetry update requests

# 查看依赖关系树
poetry show --tree

# 查看已安装的包详情
poetry show

poetry shell

# 运行Python脚本
poetry run python script.py

# 运行测试套件
poetry run pytest

# 显示当前环境的详细信息
poetry env info

# 列出所有可用的虚拟环境
poetry env list

# 首先激活Conda环境
conda activate myenv

# 验证当前Python解释器路径
which python

# 在Conda环境中使用Poetry管理依赖
poetry install

poetry build

# 发布到官方PyPI仓库
poetry publish

# 发布到自定义私有仓库
poetry publish --repository my-repo

# 一步完成构建并发布
poetry publish --build

# 导出标准requirements.txt
poetry export -f requirements.txt --output requirements.txt

# 包含开发依赖的完整导出
poetry export --with dev -f requirements.txt --output requirements-dev.txt

# 创建项目目录
mkdir my-data-science-project
cd my-data-science-project

# 交互式初始化Poetry项目
poetry init

# 添加数据科学核心库
poetry add pandas numpy matplotlib scikit-learn

# 添加开发和测试工具
poetry add pytest black isort mypy --group dev

# 创建源代码和测试目录
mkdir -p src/my_project tests

# 创建必要的初始文件
touch src/my_project/__init__.py
touch tests/__init__.py

# src/my_project/data_processor.py
import pandas as pd
from pandas import DataFrame

def process_data(data: DataFrame) -> DataFrame:
    """Clean the input dataframe by removing missing values.
    
    Args:
        data: Input DataFrame to be processed
        
    Returns:
        Cleaned DataFrame with no missing values
    """
    return data.dropna()

# tests/test_data_processor.py
import pandas as pd
import pytest
from my_project.data_processor import process_data

def test_process_data():
    # 准备测试数据，包含空值
    data = pd.DataFrame({'A': [1, 2, None], 'B': [4, None, 6]})
    
    # 调用处理函数
    result = process_data(data)
    
    # 验证结果
    assert len(result) == 1
    assert result.iloc[0, 0] == 1
    assert result.iloc[0, 1] == 4

================================================= test session starts ==================================================
platform darwin -- Python 3.13.2, pytest-8.3.5, pluggy-1.5.0
rootdir: /Users/shoukai/source/python-project/my-data-science-project
configfile: pyproject.toml
plugins: anyio-4.9.0
collected 1 item

tests/test_data_processor.py .                                                                                   [100%]

================================================== 1 passed in 0.81s ===================================================

reformatted /Users/shoukai/source/python-project/my-data-science-project/src/my_project/data_processor.py
reformatted /Users/shoukai/source/python-project/my-data-science-project/tests/test_data_processor.py

All done! ✨ 🍰 ✨
2 files reformatted, 2 files left unchanged.