如何从pyproject.toml和conf.py向Docker Compose传递参数？

完全可以实现从指定文件提取元数据并传递给Docker Compose，下面提供两种实用的落地方案：

方法1：Shell脚本直接提取并传递环境变量 #

适合快速部署场景，可选择基础Shell工具或专业TOML解析工具实现。

1.1 用基础Shell工具（grep+sed）提取 #

编写脚本extract_meta.sh，通过正则匹配提取值并传递给Docker Compose：

#!/bin/bash

# 从pyproject.toml提取元数据
PROJECT_NAME=$(grep '^name =' pyproject.toml | sed 's/name = "\(.*\)"/\1/')
PROJECT_VERSION=$(grep '^version =' pyproject.toml | sed 's/version = "\(.*\)"/\1/')
PROJECT_DESCRIPTION=$(grep '^description =' pyproject.toml | sed 's/description = "\(.*\)"/\1/')
# 将authors数组转为逗号分隔字符串
AUTHORS=$(grep -A5 'authors =' pyproject.toml | grep -E '\["?.*"?\]' | sed 's/authors = \[\(.*\)\]/\1/' | sed 's/"//g' | sed 's/, /,/g')

# 可选：从docs/conf.py提取（替换上面的pyproject.toml提取逻辑即可）
# PROJECT_NAME=$(grep '^project =' docs/conf.py | sed "s/project = '\(.*\)'/\1/")
# PROJECT_VERSION=$(grep '^version =' docs/conf.py | sed "s/version = '\(.*\)'/\1/")
# AUTHORS=$(grep '^author =' docs/conf.py | sed "s/author = '\(.*\)'/\1/")

# 导出变量并启动服务
export PROJECT_NAME PROJECT_VERSION PROJECT_DESCRIPTION AUTHORS
docker-compose up -d

使用步骤：

1. 给脚本添加执行权限：chmod +x extract_meta.sh
2. 运行脚本：./extract_meta.sh

1.2 用tomlq（更可靠的TOML解析） #

如果依赖yq工具（含tomlq），可以避免正则匹配的局限性，更稳定地解析TOML结构：

#!/bin/bash

# 从pyproject.toml提取元数据
PROJECT_NAME=$(tomlq -r '.project.name' pyproject.toml)
PROJECT_VERSION=$(tomlq -r '.project.version' pyproject.toml)
PROJECT_DESCRIPTION=$(tomlq -r '.project.description' pyproject.toml)
AUTHORS=$(tomlq -r '.project.authors | join(", ")' pyproject.toml)

# 导出变量并启动服务
export PROJECT_NAME PROJECT_VERSION PROJECT_DESCRIPTION AUTHORS
docker-compose up -d

说明：tomlq可通过pip install yq或系统包管理器（如Ubuntu的apt install yq）安装。

方法2：Python脚本生成.env文件 #

Docker Compose会自动读取当前目录的.env文件，用Python解析元数据并生成该文件，适合复杂场景或需要统一数据源的情况：

编写脚本extract_meta.py：

import tomllib
import os

# 读取pyproject.toml作为数据源
with open('pyproject.toml', 'rb') as f:
    pyproject_data = tomllib.load(f)

project_name = pyproject_data['project']['name']
project_version = pyproject_data['project']['version']
project_description = pyproject_data['project']['description']
authors = ', '.join(pyproject_data['project']['authors'])

# 可选：从docs/conf.py提取（替换上面的pyproject.toml读取逻辑）
# conf_vars = {}
# with open('docs/conf.py', 'r') as f:
#     exec(f.read(), conf_vars)
# project_name = conf_vars['project']
# project_version = conf_vars['version']
# authors = conf_vars['author']
# project_description = pyproject_data['project']['description']  # 从pyproject补全描述

# 生成.env文件
env_content = f'''PROJECT_NAME="{project_name}"
PROJECT_VERSION="{project_version}"
PROJECT_DESCRIPTION="{project_description}"
AUTHORS="{authors}"
'''

with open('.env', 'w') as f:
    f.write(env_content)

# 启动Docker Compose服务
os.system('docker-compose up -d')

使用步骤：直接运行脚本python extract_meta.py，脚本会自动生成.env并启动服务。

补充说明 #

• 建议以pyproject.toml作为单一元数据源，避免pyproject.toml和conf.py的数据不一致问题
• 如果需要在镜像构建阶段使用这些变量，需在Dockerfile中用ARG声明，例如：

  ARG PROJECT_NAME
  ARG PROJECT_VERSION
  ENV PROJECT_NAME=$PROJECT_NAME
  ENV PROJECT_VERSION=$PROJECT_VERSION
  

内容的提问来源于stack exchange，提问作者Bosser445