зеркало из
1
0
Форкнуть 0
azure-sdk-for-python/CONTRIBUTING.md

14 KiB

Contributing to Azure Python SDK

If you would like to become an active contributor to this project please follow the instructions provided in Microsoft Azure Projects Contribution Guidelines.

Generated code

If you want to contribute to a file that is generated (header contains Code generated by Microsoft (R) AutoRest Code Generator.), the best approach to open a PR on the initial Swagger specification, as we can NOT merge a PR on generated code (it would be replaced by next generation). See https://github.com/Azure/azure-rest-api-specs/ for details.

Tools Overview

We utilize a variety of tools to ensure smooth development, testing, and code quality for the Azure Python SDK. Below is a list of key tools and their purpose in the workflow:

  • Tox: Tox is our primary tool for managing test environments. It allows us to distribute tests to virtual environments, install dependencies, and maintain consistency between local and CI builds. Tox is configured to handle various testing scenarios, including linting, type checks, and running unit tests.

  • Virtualenv: Virtualenv is leveraged by Tox to create isolated environments for each test suite, ensuring consistent dependencies and reducing conflicts.

  • Pytest: Pytest is the test framework we use for writing and running our unit tests. It supports fixtures, parameterized tests, and other features that make testing more powerful and flexible.

  • Pylint: Pylint is used for code linting to enforce coding standards and catch potential issues early. Maintaining a consistent code style is important, and Pylint helps achieve that goal.

  • Mypy and Pyright: Both tools are used for type checking. Mypy helps verify that type annotations are correct, while Pyright provides additional support for type completeness and validation.

  • Sphinx: Sphinx is used for generating package documentation. This ensures that contributors and users alike have clear, accessible information about how to use the SDK.

  • Bandit: Bandit is employed to find common security issues in Python code. It performs static analysis and helps us secure the codebase.

  • Azure DevOps: Our CI/CD pipelines are managed using Azure DevOps, ensuring that builds, tests, and deployments are executed consistently and reliably across all contributions.

Building and Testing

The Azure SDK team's Python CI leverages the tool tox to distribute tests to virtual environments, handle test dependency installation, and coordinate tooling reporting during PR/CI builds. This means that a dev working locally can reproduce exactly what the build machine is doing.

A Brief Overview of Tox

A Monorepo and Tox in Harmony

Traditionally, the tox.ini file for a package sits alongside the setup.py in source code. The azure-sdk-for-python necessarily does not adhere to this policy. There are over one-hundred packages contained here-in. That's a lot of tox.ini files to maintain!

Instead, the CI system leverages the --root argument which is new to tox4. The --root argument allows tox to act as if the tox.ini is located in whatever directory you specify!

Tox Environments

A given tox.ini works on the concept of test environments. A given test environment is a combination of:

  1. An identifier (or identifiers)
  2. A targeted Python version
    1. tox will default to base python executing the tox command if no Python environment is specified
  3. (optionally) an OS platform

Internally tox leverages virtualenv to create each test environment's virtual environment.

This means that once the tox workflow is in place, all tests will be executed within a virtual environment.

You can use the command tox list to list all the environments provided by a tox.ini file. You can either use that command in the same directory as the file itself, or use the --conf argument to specify the path to it directly.

Sample output of tox list:

sdk-for-python/eng/tox> tox list
default environments:
whl              -> Builds a wheel and runs tests
sdist            -> Builds a source distribution and runs tests

additional environments:
pylint           -> Lints a package with a pinned version of pylint
next-pylint      -> Lints a package with pylint (version 2.15.8)
mypy             -> Typechecks a package with mypy (version 1.9.0)
next-mypy        -> Typechecks a package with the latest version of mypy
pyright          -> Typechecks a package with pyright (version 1.1.287)
next-pyright     -> Typechecks a package with the latest version of static type-checker pyright
verifytypes      -> Verifies the "type completeness" of a package with pyright
whl_no_aio       -> Builds a wheel without aio and runs tests
develop          -> Tests a package
sphinx           -> Builds a package's documentation with sphinx
depends          -> Ensures all modules in a target package can be successfully imported
verifywhl        -> Verify directories included in whl and contents in manifest file
verifysdist      -> Verify directories included in sdist and contents in manifest file. Also ensures that py.typed configuration is correct within the setup.py
devtest          -> Tests a package against dependencies installed from a dev index
latestdependency -> Tests a package against the released, upper-bound versions of its azure dependencies
mindependency    -> Tests a package against the released, lower-bound versions of its azure dependencies
apistub          -> Generate an api stub of a package ( for https://apiview.dev )
bandit           -> Runs bandit, a tool to find common security issues, against a package
samples          -> Runs a package's samples
breaking         -> Runs the breaking changes checker against a package

Example Usage of the common Azure SDK For Python tox.ini

Basic usage of tox within this monorepo is:

  1. pip install "tox<5"
  2. Run tox run -e ENV_NAME -c path/to/tox.ini --root path/to/python_package
  • Note: You can use environment variables to provide defaults for tox config values
    • With TOX_CONFIG_FILE set to the absolute path of tox.ini, you can avoid needing -c path/to/tox.ini in your tox invocations
    • With TOX_ROOT_DIR set to the absolute path to your python package, you can avoid needing --root path/to/python_package

The common tox.ini location is eng/tox/tox.ini within the repository.

If at any time you want to blow away the tox created virtual environments and start over, simply append -r to any tox invocation!

Example azure-core mypy

  1. Run tox run -e mypy -c ./eng/tox/tox.ini --root sdk/core/azure-core

Example azure-storage-blob tests

  1. Execute tox run -c ./eng/tox/tox.ini --root sdk/storage/azure-storage-blob

Note that we didn't provide an environment argument for this example. Reason here is that the default environment selected by our common tox.ini file is one that runs pytest.

whl environment

Used for test execution across the spectrum of all the platforms we want to support. Maintained at a platform specific level just in case we run into platform-specific bugs.

  • Installs the wheel, runs tests using the wheel
\> tox run -e whl -c <path to tox.ini> --root <path to python package>

sdist environment

Used for the local dev loop.

  • Installs package in editable mode
  • Runs tests using the editable mode installation, not the wheel

\> tox run -e sdist -c <path to tox.ini> --root <path to python package>

pylint environment

Pylint install and run.

\> tox run -e pylint -c <path to tox.ini> --root <path to python package>

mypy environment

Mypy install and run.

\> tox run -e mypy -c <path to tox.ini> --root <path to python package>

sphinx environment

Generate sphinx doc for this package.

\> tox run -e sphinx -c <path to tox.ini> --root <path to python package>

Custom Pytest Arguments

tox supports custom arguments, and the defined pytest environments within the common tox.ini also allow these. Essentially, separate the arguments you want passed to pytest by a -- in your tox invocation.

Tox Documentation on Positional Arguments

Example: Invoke tox, breaking into the debugger on failure tox run -e whl -c <path to tox.ini> --root <path to python package> -- --pdb

Performance Testing

SDK performance testing is supported via the custom perfstress framework. For full details on this framework, and how to write and run tests for an SDK - see the perfstress tests documentation.

More Reading

We maintain an additional document that has a ton of detail as to what is actually happening in these executions.

Dev Feed

Daily dev build version of Azure sdk packages for python are available and are uploaded to Azure devops feed daily. We have also created a tox environment to test a package against dev built version of dependent packages. Below is the link to Azure devops feed. https://dev.azure.com/azure-sdk/public/_packaging?_a=feed&feed=azure-sdk-for-python

To install latest dev build version of a package
pip install <package-name> --extra-index-url https://pkgs.dev.azure.com/azure-sdk/public/_packaging/azure-sdk-for-python/pypi/simple --pre

To Install a specific dev build version of a package

For e.g.

pip install azure-appconfiguration==1.0.0b6.dev20191205001 --extra-index-url https://pkgs.dev.azure.com/azure-sdk/public/_packaging/azure-sdk-for-python/pypi/simple

To test a package being developed against latest dev build version of dependent packages: a. cd to package root folder b. run tox environment devtest

\> tox run -e devtest -c <path to tox.ini> --root <path to python package>

This tox test( devtest) will fail if installed dependent packages are not dev build version.

Samples

Third-party dependencies

Third party libraries should only be included in samples when necessary to demonstrate usage of an Azure SDK package; they should not be suggested or endorsed as alternatives to the Azure SDK.

When code samples take dependencies, readers should be able to use the material without significant license burden or research on terms. This goal requires restricting dependencies to certain types of open source or commercial licenses.

Samples may take the following categories of dependencies:

  • Open-source : Open source offerings that use an Open Source Initiative (OSI) approved license. Any component whose license isn't OSI-approved is considered a commercial offering. Prefer OSS projects that are members of any of the OSS foundations that Microsoft is part of. Prefer permissive licenses for libraries, like MIT and Apache 2. Copy-left licenses like GPLv3 are acceptable for tools, and OSs. Kubernetes, Linux, and Newtonsoft.Json are examples of this license type. Links to open source components should be to where the source is hosted, including any applicable license, such as a GitHub repository (or similar).

  • Commercial: Commercial offerings that enable readers to learn from our content without unnecessary extra costs. Typically, the offering has some form of a community edition, or a free trial sufficient for its use in content. A commercial license may be a form of dual-license, or tiered license. Links to commercial components should be to the commercial site for the software, even if the source software is hosted publicly on GitHub (or similar).

  • Dual licensed: Commercial offerings that enable readers to choose either license based on their needs. For example, if the offering has an OSS and commercial license, readers can choose between them. MySql is an example of this license type.

  • Tiered licensed: Offerings that enable readers to use the license tier that corresponds to their characteristics. For example, tiers may be available for students, hobbyists, or companies with defined revenue thresholds. For offerings with tiered licenses, strive to limit our use in tutorials to the features available in the lowest tier. This policy enables the widest audience for the article. Docker, IdentityServer, ImageSharp, and Visual Studio are examples of this license type.

In general, we prefer taking dependencies on licensed components in the order of the listed categories. In cases where the category may not be well known, we'll document the category so that readers understand the choice that they're making by using that dependency.

Code of Conduct

This project's code of conduct can be found in the CODE_OF_CONDUCT.md file (v1.4.0 of the https://contributor-covenant.org/ CoC).