fix: parse charset parameters case-insensitively

[guyua9]

Summary

• Treat charset parameters case-insensitively when parsing Content-Type response headers.
• Normalize data URI parameter names so Charset=utf-8 is handled the same as charset=utf-8.
• Add regression coverage for both paths.

Why

MIME parameter names are commonly handled case-insensitively, and real HTTP responses may spell charset as Charset=UTF-8. MarkItDown already uses the parsed charset as a conversion hint, so preserving it avoids falling back to charset guessing for otherwise explicit inputs.

Tests

• uv run --with '.[all]' --with pytest --directory packages/markitdown python -m pytest tests/test_module_misc.py::test_data_uris tests/test_module_misc.py::test_response_content_type_charset_is_case_insensitive -q

[guyua9]

@microsoft-github-policy-service agree