How to recognize the soft return and hard return?

[fitexmage]

I would like to combine the texts separated by soft return but keep the hard return.

For example, in this PDF, I want to extract the text to
额外参数：\n-b : extra_post_body 额外增加的请求内容,如-b "isp:001#002",抓包会用到，平时不会用到，如果value值为数组值，则用 # 进行分隔元素，也可以使用 # 强制生成数组类型如：001#
which will keep the line break between line 1 and 2 but ignore the line break between line 2 and 3.
Is there any way to handle this?

[jsvine]

Hi @fitexmage, and thank you for you interest in pdfplumber. Generally speaking, PDFs don't have a concept of line breaks; instead, text is laid out positionally. If one line appears below another, it is because the PDF instructions indicate that lower position, rather than the PDF instructions including a newline command.

However, it is possible that, for your specific PDF, there is enough information to distinguish between these different transitions between lines. The first thing I'd suggest is examining the results of page.extract_text_lines(...), and seeing whether there is a difference in the vertical distances between those lines (by examining the top and/or bottom values).

[fitexmage]

Thank you for your response! I have tried page.extract_text_lines(...) and found that the distance between line 1 and 2 is 2.349999999999909. The distance between line 2 and 3 is 2.3700000000000045. I also tried other lines and found the distances are around 2.36 for both soft return and hard return. It seems not have any patterns.