How to repair broken lines?

[Jane Eyre]

Good morning.

I have a long text to export to epub, but..unfortunately, it often starts a new paragraph in the wrong way 🥲
So, in epub I would get this:

<p>My sentence starts here, but it should not</p>
<p>keep on here</p>

Is there any software or something that could help me?

[JSWolf]

It looks like a PDF conversion. Do you happen to have the original source?

[Jane Eyre]

Yes, I do..but, do you have any solution?

[JSWolf]

Originally Posted by Jane Eyre

Yes, I do..but, do you have any solution?

Is the source like this?

[JSWolf]

The problem is that while you can use regex to fix broken lines so you don't have broken sentences, you cannot fix it so you get the paragraphs correct without doing it by hand.

How would you know where one paragraph ends and the next begins?

[Jane Eyre]

Originally Posted by JSWolf

The problem is that while you can use regex to fix broken lines so you don't have broken sentences, you cannot fix it so you get the paragraphs correct without doing it by hand.

How would you know where one paragraph ends and the next begins?

Because the broken line re-starts by minuscule initial.

[gbm]

Originally Posted by Jane Eyre

Because the broken line re-starts by minuscule initial.

See this thread for some suggestions,

https://www.mobileread.com/forums/sh...d.php?t=237181

this is what I use in the calibre editor

search

([a-z])</p>\s*<p[^>]+>([a-z])

replace

\1 \2

bernie

P.s. make a backup copy before you start

[icearch]

You guys are lucky enough to have captical to determine whether a sentence is broken. Fix this in Chinese is really pain in the a$$.

[JSWolf]

Originally Posted by icearch

You guys are lucky enough to have captical to determine whether a sentence is broken. Fix this in Chinese is really pain in the a$$.

A capital letter is not only used at the beginning of a sentence. So that's cannot be relied on alone. You would sentence ended punctuation followed by a capitol. But there are cases when even that won't work.

And then there's the issue of where does the paragraph end/start which where is no way to automate that.

[JSWolf]

Originally Posted by Jane Eyre

Because the broken line re-starts by minuscule initial.

This cannot be automated. You will have to do some manual fixing. Automation cannot find where the paragraphs begin or end. Yiu may be able to fix most sentences, but you may not be able to fix every sentence. And you may not be able to fix the paragraphs.

This is a lose-lose situation unless you manually do some of the fixing.

[BetterRed]

If you have MS Word 2016 or later, install the TransTools addon (payware), open the PDF in Word, save as DOCX, use the Transtools "Unbreaker" feature then convert the DOCX to EPUB.

BR

[JSWolf]

Originally Posted by BetterRed

If you have MS Word 2016 or later, install the TransTools addon (payware), open the PDF in Word, save as DOCX, use the Transtools "Unbreaker" feature then convert the DOCX to EPUB.

BR

Will that have all paragraphs correct?

[BetterRed]

IIRC it has a limited time free trial. So, why don't you try it and judge for yourself.

[Jane Eyre]

Thank you all. Finally, I decided to give in. Such a waste of time!

[Karellen]

Originally Posted by Jane Eyre

Thank you all. Finally, I decided to give in. Such a waste of time!

Why? Yes there are quite a few off-putting and distracting comments in this thread, but it is do-able regardless of what others say.

I've come across this problem a few times before. Easy to fix... But there is a bit of follow up edits. So do your major fixes, then read the book and fix the remaining errors when you come across them.

I use a simpler regex than @gbm

My first pass is the following which will catch a good 90% of the splits...
Find... (\w)</p>\s+<p>(\w)
Replace... \1 \2

Then this will catch most of the remaining...
Find... ([;:,-–—])</p>\s+<p>(\w)
Replace... \1 \2