pdf->epub - using idents as a cue to line-unwrapping

[VcSaJen]

I'm using latest Calibre to convert pdf to epub. Using default settings, all of my paragraphs are split into multiple paragraphs. But all paragraphs are clearly indented, so I'm not sure why Calibre is having troubles. I don't want to adjust "unwrap factor", as it's relies on line ending early and therefore inexact. How to force Calibre into taking indents into account when determining paragraph breaks? If it's not possible, are there any other tool that have that option?

[retiredbiker]

Originally Posted by VcSaJen

If it's not possible, are there any other tool that have that option?

Read the sticky post about pdf conversion. As to other tools, I use "pdftotext" from the Poppler utilities. The --layout option will give you a text file with all the leading spaces and any extra linefeeds between paragraphs intact. Then a little regex will easily get you to the "real" paragraphs you want. If you are really lucky, there may be, say, 5 linefeeds at each chapter break, so you can get those with regex as well.

Assuming, of course, that what you want exists to start with. As with anything pdf, success depends on what is inside the source file. Pdftotext will at least show you what is there, and it may vary from excellent to impossible. Simple books like novels often work well with this, but if you have double columns or something complex like a science textbook, its a lot more work.

[VcSaJen]

Solved it with opening pdf with MS Word, then importing docx into calibre, thus converting it into epub and then manually fixing giant margins, or wrong text-align on rare paragraphs. That preserved all italics, which are all lost if I convert pdf to plain text. Bonus points for inline images which are actually inline.

[BetterRed]

@VcSaJen - There are a couple of Word addins that have tools to help deal with PDF conversions ==>> eBook Tools and TransTools.

There's some discussion on the latter in this thread from post #19 onwards.

BR