These tools help SDK Code Example developers' integration with internal docs tooling. This is mainly in the form of .doc_gen SOS metadata validation, with other tools as necessary.

SDK Code Examples tools team uses Python for our tools because of its cross-platform runtime and broad knowledge base.

1. Use Python latest for tooling (this is not the same rule as AWS SDK for Python, be aware).
2. Our “standard library” adds flake8, black, pyyaml, yamale, and requests.
   • Current versions are in base_requirements.txt.
3. Run .tools/verify_python.py to ensure python tools are configured as expected.
   • This will install our “standard library” to site_packages and ensure that the python version is at least 3.(current - 1).
4. Use a venv in .venv if the project needs packages beyond our standard library.
   1. Include those libraries in the requirements.txt.
   2. Copy .tools/base-requirements.txt to the tool’s folder to seed requirements.txt if you like.
   3. With an active venv, run pip freeze > requirements.txt to update the tool’s requirements.
5. Keep tools self executable.
   1. Mark them executable (chmod a+x <script.py> in *nix).
   2. Add a shebang #!/usr/bin/env python3.
   3. Use __name__ == "__main__" check that delegates to a main function.
6. Use f-strings for string building.
7. Use logging to write diagnostics to the console.
8. Format using Black with default settings.
9. Lint using flake8 with overrides for Black default settings. Treat Warnings as errors.
10. Verify type annotations using mypy.
    • Type annotations are strongly recommended.
11. Run tests using pytest.
12. Parse arguments using argparse.
    • All scripts must run headless, without user interaction.
13. Use pathlib for files and paths.
    • Prefer os.scandir to os.listdir.
14. Parse & dump yaml using PyYAML (imports from yaml).
    • Validate yaml schemas using yamale.
15. When working with dates, import from datetime and always include a timezone (usually timezone.utc).
16. Prefer data classes or immutable tuples for base data types.
17. Use requests for web calls, but prefer a native wrapper (like boto3 or pygithub) when available.
    • If you http/2, you might consider httpx.
18. Use subprocess for system calls, but prefer a native wrapper (like gitpython) when available.
19. When reading project files, use encoding="utf-8-sig", as many .cs files have a BOM.

Python 3.12:

• PEP 701 -- More flexible f-string parsing.
• PEP 695 -- New type annotations syntax for generic classes.

Python 3.11:

• PEP 654 -- Exception Groups and except*.

Python 3.10:

• PEP 604 -- Allow writing union types as X | Y.
• PEP 636 -- Structural Pattern Matching: Tutorial.

Python 3.9:

• PEP 585, Type Hinting Generics In Standard Collections.
• PEP 602, Python adopts a stable annual release cadence.

Python 3.8

• PEP 572, Assignment expressions.
• PEP 586, Literal types.
• PEP 589, TypedDict.

Python 3.7

• PEP 557, Data Classes.

• https://marketplace.visualstudio.com/items?itemName=ms-python.mypy-type-checker
• https://marketplace.visualstudio.com/items?itemName=ms-python.flake8
• https://marketplace.visualstudio.com/items?itemName=ms-python.black-formatter