Enhance Conversion Result by Specifying Default OCR Option (EasyOCR)?

[haoshan98]

Based on the documentation, EasyOCR is the default ocr_options, which the custom conversion pipeline would work without specifying ocr_options.
However when parsing a Chinese Language PDF, the result is not acceptable. After a random trying to clearly specify the ocr_options as EasyOcrOptions(), the result suddenly became very good. Is this an intended design?

def parse_document_content(file_path: str) -> ConversionResult:
    pipeline_options = PdfPipelineOptions()
    pipeline_options.do_ocr = True
    pipeline_options.do_table_structure = True
    pipeline_options.table_structure_options.do_cell_matching = True
    pipeline_options.ocr_options.force_full_page_ocr = True
    pipeline_options.accelerator_options = AcceleratorOptions(
        num_threads=8, device=AcceleratorDevice.CUDA
    )
    
    # clearly specify ocr_options yield better result
    pipeline_options.ocr_options = EasyOcrOptions()  # if comment out this line the result doesn't meet the expectation
    
    doc_converter = DocumentConverter(
        format_options={InputFormat.PDF: PdfFormatOption(pipeline_options=pipeline_options)}
    )
   conv_result = doc_converter.convert(file_path)
   return conv_result.document.export_to_markdown()

[haoshan98]

The sequence of specifications made a difference. No need to set force_full_page_ocr =True for Chinese language. Set pipeline_options.ocr_options = EasyOcrOptions() after pipeline_options.ocr_options.force_full_page_ocr = True overwrite it and made force_full_page_ocr = False, which yield better result.

[AdityaMannu1709]

Can u tell more about pipeline_options.ocr_options.force_full_page_ocr = True especially for English?