Introduce uv for Project Management

[KanchiShimono]

This Pull Request introduces uv to improve project management. The key reasons for adopting uv are as follows:

1. Unified Environment Management
   uv allows managing Python versions, virtual environments, and dependencies with a single tool, reducing the overhead of environment setup.

2. Strict Dependency Management
   uv precisely manages dependencies based on OS, CPU architecture, and Python versions. In contrast, the current approach using pip-compile generates lock files tied to the developer's Python version, which can lead to unexpected issues (e.g., failing to resolve dependencies for libraries that have dropped support for older Python versions).

3. High Performance
   With optimized dependency resolution and caching mechanisms, uv provides a fast and efficient package management experience.

4. Consistency Across the Organization
   Since Pydantic has already adopted uv, standardizing project management tools within the same organization reduces cognitive load for developers.

[KanchiShimono]

The test failures are caused by changes introduced in Pydantic 2.10 and are unrelated to the changes in this PR.

However, it was found that in Python 3.9 or earlier combined with Pydantic 2.10, CliSettingsSource does not correctly generate metavar.

pydantic-settings/pydantic_settings/sources.py

Lines 2001 to 2005 in 4b6fd3d

	elif isinstance(obj, WithArgsTypes):
	return self._metavar_format_choices(
	list(map(self._metavar_format_recurse, self._get_modified_args(obj))),
	obj_qualname=obj.__qualname__ if hasattr(obj, '__qualname__') else str(obj),
	)

As a workaround, applying the following change allows the tests to pass, but I am unsure if this is the appropriate solution.

        elif isinstance(obj, WithArgsTypes) and sys.version_info >= (3, 9):

I believe it would be best to handle the test failures separately from this PR. What do you think?

[hramezani]

The test failures are caused by changes introduced in Pydantic 2.10 and are unrelated to the changes in this PR.

However, it was found that in Python 3.9 or earlier combined with Pydantic 2.10, CliSettingsSource does not correctly generate metavar.

pydantic-settings/pydantic_settings/sources.py

Lines 2001 to 2005 in 4b6fd3d

	elif isinstance(obj, WithArgsTypes):
	return self._metavar_format_choices(
	list(map(self._metavar_format_recurse, self._get_modified_args(obj))),
	obj_qualname=obj.__qualname__ if hasattr(obj, '__qualname__') else str(obj),
	)

As a workaround, applying the following change allows the tests to pass, but I am unsure if this is the appropriate solution.

        elif isinstance(obj, WithArgsTypes) and sys.version_info >= (3, 9):

I believe it would be best to handle the test failures separately from this PR. What do you think?

Thanks @KanchiShimono for this PR.

Yes, it is better to handle the test failure in separate PR.

I think @kschwab can fix the problem because he wrote most of the code there and he is the CLI expert

[KanchiShimono]

@hramezani @kschwab
Thank you for your reply.

1. I have created a new issue Incorrect metavar generation in CliSettingsSource with Python 3.9 or earlier and Pydantic 2.10 #551 regarding the incorrect generation of metavar.
2. As a workaround, I added a constraint on the Pydantic version used during development with Python 3.9 or earlier. This constraint is applied only when managing the development environment with uv and does not affect installations from PyPI as a library.
3. Tests for Python 3.9 and earlier use an older version of Pydantic, causing some print tests using pytest-examples to fail. To address this, I modified the tests to skip test_docs.py when an older version of Pydantic is detected.

[hramezani]

@KanchiShimono you need to rebase.

we removed support for Python 3.8 and fixed the bug

[KanchiShimono]

@hramezani

I have rebased the branch to incorporate the latest changes from main.
All tests are now passing.

[hramezani]

Let's not change the release part. as we are using the same thing in pydantic.

[hramezani]

@KanchiShimono I requested a small change, other than that LGTM.

[KanchiShimono]

@hramezani
Thank you for your review.

I have borrowed the release part from the latest version of Pydantic. Additionally, I have upgraded the default Python version used in GitHub Actions to 3.12.

[hramezani]

Thanks @KanchiShimono