convert_image.rst 906 B

123456789101112131415161718192021222324252627282930313233
  1. Convert Image
  2. ===============
  3. .. code:: python
  4. import os
  5. from magic_pdf.data.data_reader_writer import FileBasedDataWriter
  6. from magic_pdf.model.doc_analyze_by_custom_model import doc_analyze
  7. from magic_pdf.data.read_api import read_local_images
  8. # prepare env
  9. local_image_dir, local_md_dir = "output/images", "output"
  10. image_dir = str(os.path.basename(local_image_dir))
  11. os.makedirs(local_image_dir, exist_ok=True)
  12. image_writer, md_writer = FileBasedDataWriter(local_image_dir), FileBasedDataWriter(
  13. local_md_dir
  14. )
  15. # proc
  16. ## Create Dataset Instance
  17. input_file = "some_image.jpg" # replace with real image file
  18. input_file_name = input_file.split(".")[0]
  19. ds = read_local_images(input_file)[0]
  20. ds.apply(doc_analyze, ocr=True).pipe_ocr_mode(image_writer).dump_md(
  21. md_writer, f"{input_file_name}.md", image_dir
  22. )