fix: use utf-8 encoding for subprocess to handle non-ascii chars

[yuxiaoli]

This PR forces UTF-8 encoding when running git subprocess commands to prevent UnicodeDecodeError on systems with different default encodings (e.g., GBK on Windows).

[tomice]

Thank you for your contribution! It looks like the unit tests fail because of the new args you are adding not matching up with the expected args in the tests:

AssertionError: expected call not found.
Expected: run(['git', 'status'], stdout=-1, stderr=-1, text=True, check=True)
Actual: run(['git', 'status'], stdout=-1, stderr=-1, text=True, check=True, encoding='utf-8', errors='replace')

https://github.com/git-quick-stats/git-py-stats/blob/main/git_py_stats/tests/test_git_operations.py

Should we be doing errors=replace here, or do we look towards PEP323 for guidance and instead use errors=surrogateescape? I am fine with either/or, but the argument for surrogateescape seems compelling: https://docs.python.org/3/library/codecs.html