Implement isort's default-section setting

[mmerickel]

Summary

This fixes #7868.

Support isort's default-section feature which allows any imports that match sections that are not in section-order to be mapped to a specifically named section.

https://pycqa.github.io/isort/docs/configuration/options.html#default-section

This has a few implications:

• It is no longer required that all known sections are defined in section-order.
• This is technically a bw-incompat change because currently if folks define custom groups, and do not define a section-order, the code used to add all known sections to section-order while emitting warnings. However, when this happened, users would be seeing warnings so I do not think it should count as a bw-incompat change.

Test Plan

• Added a new test.
• Did not break any existing tests.

Finally, I ran the following config against Pyramid's complex codebase that was previously using isort and this change worked there.

pyramid's previous isort config

https://github.com/Pylons/pyramid/blob/5f7e286b0629b0a5f1225fe51172cba77eb8fda1/pyproject.toml#L22-L37

[tool.isort]
profile = "black"
multi_line_output = 3
src_paths = ["src", "tests"]
skip_glob = ["docs/*"]
include_trailing_comma = true
force_grid_wrap = false
combine_as_imports = true
line_length = 79
force_sort_within_sections = true
no_lines_before = "THIRDPARTY"
sections = "FUTURE,THIRDPARTY,FIRSTPARTY,LOCALFOLDER"
default_section = "THIRDPARTY"
known_first_party = "pyramid"

tested with ruff isort config

[tool.ruff.lint.isort]
case-sensitive = true
combine-as-imports = true
force-sort-within-sections = true
section-order = [
    "future",
    "third-party",
    "first-party",
    "local-folder",
]
default-section = "third-party"
known-first-party = [
    "pyramid",
]

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

Formatter (stable)

ℹ️ ecosystem check encountered format errors. (no format changes; 1 project error)

openai/openai-cookbook (error)

warning: Detected debug build without --no-cache.
error: Failed to read examples/How_to_handle_rate_limits.ipynb: Expected a Jupyter Notebook, which must be internally stored as JSON, but this file isn't valid JSON: trailing comma at line 47 column 4

Formatter (preview)

ℹ️ ecosystem check encountered format errors. (no format changes; 1 project error)

openai/openai-cookbook (error)

ruff format --preview

warning: Detected debug build without --no-cache.
error: Failed to read examples/How_to_handle_rate_limits.ipynb: Expected a Jupyter Notebook, which must be internally stored as JSON, but this file isn't valid JSON: trailing comma at line 47 column 4

[archoversight]

This will allow me to drop my dependency on isort in projects and just use ruff for it all. 👍

[charliermarsh]

This is good work -- thanks @mmerickel!

I'll admit that I have some hesitancy to support this setting. I think it's possible to come up with a more intuitive and user-friendly configuration scheme for the use-case that motivated the change (e.g., if we want users to be able to merge standard library and third-party imports, perhaps we should focus on how to enable that?). At the same time, I understand why it's helpful to follow isort here.

[mmerickel]

Thank you @charliermarsh!

[charliermarsh]

Thank you! Nice PR.