Feature request: add support for user-provided OCR model

[Vdaleke]

Requested feature

Similar to #661 and #930. Provide the user with the ability to pass a custom OcrModel class when there is no support for the required model in docling. There are many cloud-based OCR solutions that can be used in docling, but submitting pull requests and waiting for approval for all available models is difficult to support, and some unpopular or commercial models are generally undesirable in such a project.

Alternatives

I propose to provide the user with the ability to pass in the OCR model options for use and call it inside docling. ocr_model: Optional[Type[BaseOcrModel]] = None