[data]: default Ray Data log encoding to UTF-8 on Windows

[Sanskarzz]

Description

Addressed review comment for PR

Related issues

#59967

[gemini-code-assist]

Code Review

This pull request effectively resolves a UnicodeEncodeError in Ray Data logging on Windows by defaulting the log file encoding to UTF-8. The change is cleanly implemented in SessionFileHandler, applying conditionally for the Windows OS. The PR also commendably refactors SessionFileHandler to allow dependency injection for the log directory path, which enhances testability. A new, well-structured unit test leverages this refactoring to validate the fix by simulating a Windows environment. The code is of high quality, and the changes are correct and well-tested.

[bveeramani]

Nice. Just one minor comment about the test

[bveeramani]

Could we create a tmp_logger pytest fixture? I think it'll make this test easier to read, and fixtures are more conventional for Python than try-catch

@pytest.fixture
def tmp_logger():
    # 1. Unique name for total isolation
    name = f"test_{uuid.uuid4().hex}"
    logger = logging.getLogger(name)
    
    yield logger

    # 2. Close resources and clear references
    for handler in logger.handlers[:]:  # Slice copy to avoid mutation issues
        handler.close()
        logger.removeHandler(handler)
    
    # 3. Reset level to prevent any weird propagation
    logger.setLevel(logging.NOTSET)

[bveeramani]

@Sanskarzz looks like CI is failing. Would you mind taking a look?

[github-actions]

This pull request has been automatically marked as stale because it has not had
any activity for 14 days. It will be closed in another 14 days if no further activity occurs.
Thank you for your contributions.

You can always ask for help on our discussion forum or Ray's public slack channel.

If you'd like to keep this open, just leave any comment, and the stale label will be removed.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.