convert_image.rst 1.0 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647
  1. Convert Image
  2. ===============
  3. Command Line
  4. ^^^^^^^^^^^^^
  5. .. code:: python
  6. # make sure the file have correct suffix
  7. magic-pdf -p a.png -o output -m auto
  8. API
  9. ^^^^^^
  10. .. code:: python
  11. import os
  12. from magic_pdf.data.data_reader_writer import FileBasedDataWriter
  13. from magic_pdf.model.doc_analyze_by_custom_model import doc_analyze
  14. from magic_pdf.data.read_api import read_local_images
  15. # prepare env
  16. local_image_dir, local_md_dir = "output/images", "output"
  17. image_dir = str(os.path.basename(local_image_dir))
  18. os.makedirs(local_image_dir, exist_ok=True)
  19. image_writer, md_writer = FileBasedDataWriter(local_image_dir), FileBasedDataWriter(
  20. local_md_dir
  21. )
  22. # proc
  23. ## Create Dataset Instance
  24. input_file = "some_image.jpg" # replace with real image file
  25. input_file_name = input_file.split(".")[0]
  26. ds = read_local_images(input_file)[0]
  27. # ocr mode
  28. ds.apply(doc_analyze, ocr=True).pipe_ocr_mode(image_writer).dump_md(
  29. md_writer, f"{input_file_name}.md", image_dir
  30. )