fix: add CLI LLM options

[he-yufeng]

Summary

• add CLI support for --llm-client openai, --llm-model, and --llm-prompt
• pass LLM options through to MarkItDown(...) so plugin converters can use the documented OCR workflow
• use sys.executable in CLI subprocess tests so they run under the active test interpreter on Windows

To verify

• .venv\Scripts\python.exe -m pytest packages\markitdown\tests\test_cli_misc.py -q
• .venv\Scripts\python.exe -m mypy --ignore-missing-imports packages\markitdown\src\markitdown\__main__.py packages\markitdown\tests\test_cli_misc.py
• .venv\Scripts\python.exe -m py_compile packages\markitdown\src\markitdown\__main__.py packages\markitdown\tests\test_cli_misc.py
• git diff --check -- packages/markitdown/src/markitdown/__main__.py packages/markitdown/tests/test_cli_misc.py

Fixes #1897

[he-yufeng]

Rebased onto current main; no conflicts.

Focused validation after the rebase:

python -m py_compile packages\markitdown\src\markitdown\__main__.py packages\markitdown\tests\test_cli_misc.py
$env:PYTHONPATH='C:\dev\GITHUB-clean\markitdown-1908\packages\markitdown\src'; python -m pytest packages\markitdown\tests\test_cli_misc.py -q
git diff --check origin/main..HEAD

Result: 4 passed for the CLI misc tests.