Unrecognized Arguments Error in markitdown CLI for undocumented arguments

Unrecognized Arguments Error in markitdown CLI for Undocumented Arguments

Description

The README.md file in packages/markitdown-ocr provides a CLI example:

markitdown document.pdf --use-plugins --llm-client openai --llm-model gpt-4o

However, running this command results in an unrecognized arguments error. This discrepancy occurs because the arguments --use-plugins, --llm-client, and --llm-model are not defined in the CLI parser (__main__.py). The documentation suggests these arguments are supported, but they are not implemented in the codebase.

Steps to Reproduce

1. Navigate to the packages/markitdown-ocr directory.
2. Run the following command:

   markitdown document.pdf --use-plugins --llm-client openai --llm-model gpt-4o
   

3. Observe the error:

   error: unrecognized arguments: --use-plugins --llm-client openai --llm-model gpt-4o
   

Expected Behavior

The CLI should recognize and process the arguments --use-plugins, --llm-client, and --llm-model as documented in the README.md. These arguments should be passed to the MarkItDown constructor to ensure the CLI behaves as described.

Suggested Fix

Update the CLI parser in __main__.py to include the following arguments:

• --use-plugins: A flag to enable plugin support.
• --llm-client: A string argument to specify the LLM client (e.g., openai).
• --llm-model: A string argument to specify the LLM model (e.g., gpt-4o).

These arguments should be passed as keyword arguments to the MarkItDown constructor to ensure compatibility with the documented behavior.

Compare the documented command in packages/markitdown-ocr/README.md with the argument definitions in main.py, then inspect how the MarkItDown constructor receives CLI options. Re-run the documented command from packages/markitdown-ocr; done means all three arguments are accepted and passed through as described without an unrecognized-arguments error.