Python 依赖管理与构建工具(CookieCutter, PyScaffold, PyBuilder, Poetry)

2021-09-30 | 阅读(1,195)

Python 历时这么久以来至今还未有一个事实上标准的项目管理及构建工具，以至于造成 Python 项目的结构与构建方式五花八门。这或许是体现了 Python 的自由意志。不像 Java 在经历了最初的手工构建，到半自动化的 Ant, 再到 Maven 基本就是事实上的标准了。其间 Maven 还接受了其他的 Gradle(Android 项目主推), SBT(主要是 Scala 项目), Ant+Ivy, Buildr 等的挑战，但都很难撼动 Maven 的江湖地位，而且其他的差不多遵循了 Maven 的目录布局。

回到 Python，产生过 pip, pipenv, conda 那样的包管理工具，但对项目的目录布局没有任何约定。关于构建很多还是延续了传统的 Makefile 的方式，再就是加上 setup.py 和 build.py 用程序代码来进行安装与构建。关于项目目录布局，有做成项目模板的，然后做成工具来应用项目模板。下面大概浏览一下四个工具的使用

CookieCutter
PyScaffold
PyBuilder
Poetry

CookieCutter 一个经典的 Python 项目目录结构

$ pip install cookiecutter
$ cookiecutter gh:audreyr/cookiecutter-pypackage # 以 github 上的 audreyr/cookiecutter-pypackage 为模板，再回答一堆的问题生成一个 Python 项目
......
project_name [Python Boilerplate]: sample
......

最后由 cookiecutter 生成的项目模板是下面的样子

$ tree sample
sample
├── AUTHORS.rst
├── CONTRIBUTING.rst
├── HISTORY.rst
├── LICENSE
├── MANIFEST.in
├── Makefile
├── README.rst
├── docs
│   ├── Makefile
│   ├── authors.rst
│   ├── conf.py
│   ├── contributing.rst
│   ├── history.rst
│   ├── index.rst
│   ├── installation.rst
│   ├── make.bat
│   ├── readme.rst
│   └── usage.rst
├── requirements_dev.txt
├── sample
│   ├── __init__.py
│   ├── cli.py
│   └── sample.py
├── setup.cfg
├── setup.py
├── tests
│   ├── __init__.py
│   └── test_sample.py
└── tox.ini

3 directories, 26 files

$ tree sample

sample

├── AUTHORS.rst

├── CONTRIBUTING.rst

├── HISTORY.rst

├── LICENSE

├── MANIFEST.in

├── Makefile

├── README.rst

├── docs

│ ├── Makefile

│ ├── authors.rst

│ ├── conf.py

│ ├── contributing.rst

│ ├── history.rst

│ ├── index.rst

│ ├── installation.rst

│ ├── make.bat

│ ├── readme.rst

│ └── usage.rst

├── requirements_dev.txt

├── sample

│ ├── __init__.py

│ ├── cli.py

│ └── sample.py

├── setup.cfg

├── setup.py

├── tests

│ ├── __init__.py

│ └── test_sample.py

└── tox.ini

3 directories, 26 files

这大概是当前比较流行的目录结构的主体框架，主要元素是

$ tree sample
sample
├── Makefile
├── README.rst
├── docs
│   └── index.rst
├── requirements.txt
├── sample
│   ├── __init__.py
│   └── sample.py
├── setup.cfg
├── setup.py
└── tests
    ├── __init__.py
    └── test_sample.py

$ tree sample

sample

├── Makefile

├── README.rst

├── docs

│ └── index.rst

├── requirements.txt

├── sample

│ ├── __init__.py

│ └── sample.py

├── setup.cfg

├── setup.py

└── tests

├── __init__.py

└── test_sample.py

项目 sample 目录中重复 sample 目录中放置 Python 源文件，tests 目录中是测试文件，再加一个 docs 目录放文档，README.rst, 其他的用于构建的 setup, setup.cfg 和 Makefile 文件。

这其实是一个很经典的 Python 项目结构，接下来的构建就用 make 命令了，输入 make 会看到定义在 Makefile 文件中的指令

$ make
clean                remove all build, test, coverage and Python artifacts
clean-build          remove build artifacts
clean-pyc            remove Python file artifacts
clean-test           remove test and coverage artifacts
lint                 check style
test                 run tests quickly with the default Python
test-all             run tests on every Python version with tox
coverage             check code coverage quickly with the default Python
docs                 generate Sphinx HTML documentation, including API docs
servedocs            compile the docs watching for changes
release              package and upload a release
dist                 builds source and wheel package
install              install the package to the active Python's site-packages

$ make

clean remove all build, test, coverage and Python artifacts

clean-build remove build artifacts

clean-pyc remove Python file artifacts

clean-test remove test and coverage artifacts

lint check style

test run tests quickly with the default Python

test-all run tests on every Python version with tox

coverage check code coverage quickly with the default Python

docs generate Sphinx HTML documentation, including API docs

servedocs compile the docs watching for changes

release package and upload a release

dist builds source and wheel package

install install the package to the active Python's site-packages

为使用上面的构建过程，需要安装相应的包，如 tox, wheel, coverage, sphinx, flake8, 它们都可以通过 pip 来安装。之后就可以 make test, make coverage, make docs，make dist 等。其中 make docs 可以生成一个很漂亮的 Web 文档。

另一个 Python 的 CookieCutter 模板 https://github.com/Kwpolska/python-project-template

PyScaffold 创建一个项目

PyScaffold 顾名思义，它是一个用来创建 Python 项目脚手架的工具，安装和使用

$ pip install pyscaffold
$ putup sample

这样创建了一个 Python 项目，目录结构与前面 cookiecutter 所选的模板差不多，只不过它把源文件放在了 src 目录，而非 sample 目录。

tree sample
sample
├── AUTHORS.rst
├── CHANGELOG.rst
├── CONTRIBUTING.rst
├── LICENSE.txt
├── README.rst
├── docs
│   ├── Makefile
│   ├── _static
│   ├── authors.rst
│   ├── changelog.rst
│   ├── conf.py
│   ├── contributing.rst
│   ├── index.rst
│   ├── license.rst
│   ├── readme.rst
│   └── requirements.txt
├── pyproject.toml
├── setup.cfg
├── setup.py
├── src
│   └── sample
│       ├── __init__.py
│       └── skeleton.py
├── tests
│   ├── conftest.py
│   └── test_skeleton.py
└── tox.ini

tree sample

sample

├── AUTHORS.rst

├── CHANGELOG.rst

├── CONTRIBUTING.rst

├── LICENSE.txt

├── README.rst

├── docs

│ ├── Makefile

│ ├── _static

│ ├── authors.rst

│ ├── changelog.rst

│ ├── conf.py

│ ├── contributing.rst

│ ├── index.rst

│ ├── license.rst

│ ├── readme.rst

│ └── requirements.txt

├── pyproject.toml

├── setup.cfg

├── setup.py

├── src

│ └── sample

│ ├── __init__.py

│ └── skeleton.py

├── tests

│ ├── conftest.py

│ └── test_skeleton.py

└── tox.ini

整个项目的构建就要用 tox 这个工具了。tox 是一个自动化测试和构建工具，它在构建过程中可创建 Python 虚拟环境，这让测试和构建能有一个干净的环境。

tox -av 能显示出定义在 tox.ini 中所有的任务

$ tox -av
default environments:
default   -> Invoke pytest to run automated tests

additional environments:
build     -> Build the package in isolation according to PEP517, see https://github.com/pypa/build
clean     -> Remove old distribution files and temporary build artifacts (./build and ./dist)
docs      -> Invoke sphinx-build to build the docs
doctests  -> Invoke sphinx-build to run doctests
linkcheck -> Check for broken links in the documentation
publish   -> Publish the package you have been developing to a package index server. By default, it uses testpypi. If you really want to publish your package to be publicly accessible in PyPI, use the `-- --repository pypi` option.

$ tox -av

default environments:

default -> Invoke pytest to run automated tests

additional environments:

build -> Build the package in isolation according to PEP517, see https://github.com/pypa/build

clean -> Remove old distribution files and temporary build artifacts (./build and ./dist)

docs -> Invoke sphinx-build to build the docs

doctests -> Invoke sphinx-build to run doctests

linkcheck -> Check for broken links in the documentation

publish -> Publish the package you have been developing to a package index server. By default, it uses testpypi. If you really want to publish your package to be publicly accessible in PyPI, use the `-- --repository pypi` option.

要执行哪个命令便用 tox -e build, tox -e docs 等, 下面是如何使用 PyScaffold 的动图

在我体验 tox 命令过程中，每一步好像都比较慢，应该是创建虚拟机要花些时间。

PyBuilder

最好再看另一个构建工具 PyBuilder，它所创建出的目录结构很接近于 Maven, 下面来瞧瞧

$ pip install pybuilder
$ mkdir sample && cd sample # 项目目录需手工创建
$ pyb --start-project # 回答一些问题后创建所需的目录和文件

完后看下它的目录结构

sample $ tree
.
├── build.py
├── docs
├── pyproject.toml
├── setup.py
└── src
    ├── main
    │   ├── python
    │   └── scripts
    └── unittest
        └── python

sample $ tree

├── build.py

├── docs

├── pyproject.toml

├── setup.py

└── src

├── main

│ ├── python

│ └── scripts

└── unittest

└── python

构建过程仍然是用 pyb 命令，可用 pyb -h 查看帮助，pyb -t 列出所有的任务, PyBuilder 的任务是以插件的方式加入的，插件配置在 build.py 文件中。

sample$ pyb -t
Tasks found for project "sample":
                  analyze -  Execute analysis plugins.
                            depends on tasks: prepare run_unit_tests
                    clean - Cleans the generated output.
          compile_sources - Compiles source files that need compilation.
                            depends on tasks: prepare
                 coverage - <no description available>
                            depends on tasks: verify
                  install - Installs the published project.
                            depends on tasks: package publish(optional)
                  package - Packages the application. Package a python application.
                            depends on tasks: compile_sources run_unit_tests(optional)
                  prepare - Prepares the project for building. Creates target VEnvs
        print_module_path - Print the module path.
       print_scripts_path - Print the script path.
                  publish - Publishes the project.
                            depends on tasks: package verify(optional) coverage(optional)
    run_integration_tests - Runs integration tests on the packaged application.
                            depends on tasks: package
           run_unit_tests - Runs all unit tests. Runs unit tests based on Python's unittest module
                            depends on tasks: compile_sources
                   upload - Upload a project to PyPi.
                   verify - Verifies the project and possibly integration tests.
                            depends on tasks: run_integration_tests(optional)
sample$ pyb run_unit_tests

sample$ pyb -t

Tasks found for project "sample":

analyze - Execute analysis plugins.

depends on tasks: prepare run_unit_tests

clean - Cleans the generated output.

compile_sources - Compiles source files that need compilation.

depends on tasks: prepare

coverage - <no description available>

depends on tasks: verify

install - Installs the published project.

depends on tasks: package publish(optional)

package - Packages the application. Package a python application.

depends on tasks: compile_sources run_unit_tests(optional)

prepare - Prepares the project for building. Creates target VEnvs

print_module_path - Print the module path.

print_scripts_path - Print the script path.

publish - Publishes the project.

depends on tasks: package verify(optional) coverage(optional)

run_integration_tests - Runs integration tests on the packaged application.

depends on tasks: package

run_unit_tests - Runs all unit tests. Runs unit tests based on Python's unittest module

depends on tasks: compile_sources

upload - Upload a project to PyPi.

verify - Verifies the project and possibly integration tests.

depends on tasks: run_integration_tests(optional)

sample$ pyb run_unit_tests

PyBuilder 也是在构建或测试之前创建虚拟环境, 从 0.12.9 版开始可通过参数 --no-venvs 跳过创建虚拟环境这一步。使用了 --no-venvs 的话 Python 代码将会在运行 pyb 的当前 Python 环境中执行，所需的依赖将要手工安装。

项目的依赖也要定义在 build.py 文件中

@init
def set_properties(project):
    project.depends_on('boto3', '>=1.18.52')
    project.build_depends_on('mock')

@init

def set_properties(project):

project.depends_on('boto3', '>=1.18.52')

project.build_depends_on('mock')

随后在执行 pyb 创建虚拟环境时就会安装上面的依赖，并在其中运行测试与构建。

Poetry

最后一个 Poetry, 感觉这是一个更为成熟，项目活跃度也更高的 Python 构建，它有着更强大的信赖管理功能，用 poetry add boto3 就能添加依赖，poetry show --tree 显示出依赖树。看下如何安装及创建一个项目

$ pip install poetry
$ poetry new sample

注：Poetry 推荐的安装方式是

$ curl https://raw.githubusercontent.com/python-poetry/poetry/master/install-poetry.py | python3 -

因为 pip install poetry 会往当前 Python 环境中安装它自己的许多依赖，可能会造成包版本的冲突。

它创建的项目比上面都简单

$ tree sample
sample
├── README.rst
├── pyproject.toml
├── sample
│   └── __init__.py
└── tests
    ├── __init__.py
    └── test_sample.py

$ tree sample

sample

├── README.rst

├── pyproject.toml

├── sample

│ └── __init__.py

└── tests

├── __init__.py

└── test_sample.py

如果给 poetry new 带上 --src 参数，那么源文件目录 sample 会放在 src 目录下，即 sample/src/sample.

poetry init 会在当前目录中生成 pyproject.toml 文件，目录等的生成需手动完成。

它不关注文档的生成，代码规范的检查，代码覆盖率都没有。它的项目配置更集中，全部在 pyproject.toml 文件中，toml 是什么呢？它是一种配置文件的格式 Tom's Obvious, Minimal Language.

pyproject.toml 有些类似 NodeJS 的 package.json 文件，比如 poetry add, poetry install 命令的行

poetry add boto3 # 往 pyproject.toml 中添加对 boto3 的依赖并安装(add 还能从本地或 git 来安装依赖 ), 用 --dev 参数是给开发时用的
poetry install # 将依照 pyproject.toml 文件中定义安装相应的依赖到当前的 Python 虚拟环境中，比如在 <test-venv>/lib/python3.9/site-packages 目录中，安装好模块后也可让测试用例使用

其他主要的

poetry build # 构建可安装的 *.whl 和 tar.gz 文件
poetry shell # 会根据定义在 pyproject.toml 文件中的依赖创建并使用虚拟环境
poetry env use 3.11 # 使用某个特定的 Python 版本创建虚拟环境
poetry env list --full-path # 列出虚拟环境的完整路径
poetry run pytest # 运行使用 pytest 的测试用例，如 tests/test_sample.py
poetry run python -m unittest tests/sample_tests.py # 运行 unittest 测试用例
poetry export --without-hashes --output requirements.txt # 导出 requirements.txt 文件, --dev 导出含 dev 的依赖
或者用 poetry export --without-hashes > requirements.txt

poetry run 能执行任何系统命令，只是它会在它要的虚拟环境中执行。所以可以想见，poetry 的项目要生成文档或覆盖率都必须用 poetry run ... 命令来支持 sphinx, coverage 或 flake8。

在 sample 目录(与 pyproject.toml 文件平级)中创建文件 my_module.py, 内容为

def main():
    print('hello poetry')

1 2	def main(): print('hello poetry')

然后在 pyproject.toml 中写上

[tool.poetry.scripts]
my-script="sample.my_module:main"

1 2	[tool.poetry.scripts] my-script="sample.my_module:main"

再执行

$ poetry run my-script

就会输出 "hello poetry"。

通过对以上四个工具的认识，项目结构的复杂度由 cookiecutter-pyproject -> PyScaffold -> PyBuilder -> Poetry 依次降低，使用的难度大略也是相同的顺序。

链接：

Set up tests, linters and type checking in Python projects in 2020(介绍了 poetry 项目如何支持 coverage, lint 和 type checking)
Dependency management tools for Python

本文链接 https://yanbin.blog/python-dependency-management-build-tools/, 来自隔叶黄莺 Yanbin Blog

best coffee on SciPy 最优化之最小化I wanted to take a moment to commend you on the outstanding quality of...
seetimee on 体验 Python FastAPI 的并发能力及线, 进程模型感谢
Yanbin on Mockito 3.4.0 开始可 Mock 静态方法有一个补救，新写了一篇 https://yanbin.blog/mockito-mock-static-method-in-multiple...
Yanbin on 升级到 Spring Boot 3 后 javax.inject.Named 不可用怎么，被抄袭了！算是被机器翻译引用的？
oleksi on 升级到 Spring Boot 3 后 javax.inject.Named 不可用https://www.springcloud.io/post/2023-02/springboot-3-javax-inject-name...

M	T	W	T	F	S	S
	1	2	3	4	5	6
7	8	9	10	11	12	13
14	15	16	17	18	19	20
21	22	23	24	25	26	27
28	29	30