The strip() operation in clean_text can cause incorrect hierarchy of lists

[shannanyinxiang]

Leading whitespaces are removed in https://github.com/rednote-hilab/dots.ocr/blob/75bd78eaca801585453c13e2672edc692c779d6c/dots_ocr/utils/format_transformer.py#L136

Thus, multi-level lists will be normalized to a single level during post-processing. For example

will be recognized as

• Treatments
• Control
• Traditional DDGS
• DDGS from fractionation process

rather than

• Treatments
  • Control
  • Traditional DDGS
  • DDGS from fractionation process

[ygfrancois]

Leading whitespaces are removed in

dots.ocr/dots_ocr/utils/format_transformer.py

Line 136 in 75bd78e

text = text.strip()
Thus, multi-level lists will be normalized to a single level during post-processing. For example

will be recognized as

• Treatments
• Control
• Traditional DDGS
• DDGS from fractionation process

rather than

• Treatments

  • Control
  • Traditional DDGS
  • DDGS from fractionation process

Thanks，this is a bug, you can submit a PR, or I will try to fix it.

[shannanyinxiang]

This issue has been fixed in PR #263.