Question how to eliminate unwanted space in Sigil

[jimdays]

I have an ebook originally in doc format. I opened the ebook in Microsoft Word and saved as PDF. Then I used Abbyy Finereader 12 to convert to epub. In some cases (usually around pictures or author sidebar), unwanted spaces appeared. When I opened the epub in Sigil, I couldn't just easily delete the unwanted space. I had to do a lot of copy/paste and typing to eliminate the space. My question is: is there an easy way to delete the unwanted spaces.
The epub is here (search for "rolling" to find the part):
http://files.videohelp.com/u/61125/Travel2.epub
This first screenshot shows the unwanted space in Kindle Previewer.
The second screenshot shows the unwanted space in Sigil. I can't take the cursor and just delete the space.


Image violates Posting Guideline for size - MODERATOR]

Image violates Posting Guideline for size - MODERATOR]

[eschwartz]

Stop converting DOCX to PDF and then OCRing to EPUB. You are deliberately shooting yourself in the foot (and cutting off your nose for good measure).

Convert DOCX to EPUB.

End of story.

[jimdays]

I initially tried that (convert doc to epub with several different methods, Calibre/online), but around the pictures and author sidebars, the results were completely unacceptable.
Examples of sidebar/pics:

[eschwartz]

Is this Fixed-Format?

Regardless, choose your problem: horrifically mangled OCR spew, or improperly-converted sidebars.

Either one requires fixing by hand.

Alternatively, Kindle Previewer might be able to do directly from DOCX to dual-MOBI and preserve everything, and KindleUnpack can extract an EPUB from that, which *might* do the trick (or a trick, anyway).

[BetterRed]

@jimdays - have you tried

• converting a DOCX from Word (or Writer) with calibre
• importing a DOCX into the calibre book editor
• using Toxaris' e-Book Tools - a Word add-in

Any one of them should do a better job than exporting a PDF from Word and OCRing it. As eschwartz indicates - you'll have to do some editing, but not to get rid of the crud created by OCR scanning.

BR

[theducks]

Your original issue looked like a split (broken) paragraph.

I fix those in Codeview by using the proper REGEX to remove the trailing Paragraph ending (usually </p> but may include /Spans, /i, /b ... or it might use a Div)

I tend to do in 3 passes (note that some patterns include SELECTED punctuation marks
<lower to lower pass>
([a-z,"])</p>\s+([a-z])
\1 \2 ( after a few test replace, I use All )

<lower to upper pass>
([a-z,"])</p>\s+([A-Z"])
\1 \2 (I step through on this one and confirm each replace)

<dehyphen pass>
([a-z]-)</p>\s+([a-z])
\1\2 <no space in this replace. Also a step through pass

And you might want to do a emdash pass as well

[Turtle91]

Originally Posted by theducks

<dehyphen pass>
([a-z]-)</p>\s+([a-z])
\1\2 <no space in this replace. Also a step through pass

And you might want to do a emdash pass as well

Thanks Ducks!

I am nowhere close to an expert on regex... could you not do the third pass like such to get dash and emdash on the same pass?

<dehyphen pass>
([a-z])([-—])</p>\s+([a-z])
\1\3

The way I read yours is that it includes the "-" in the first parenthesis and would thus put it back on the replace.

[rubeus]

Instead of declaring wrong chars at the end of a paragraph ist better to declare allowed chars at the end of a paragraph... such as [^.,:!"] and some other quote chars.

[Hitch]

Originally Posted by jimdays

I initially tried that (convert doc to epub with several different methods, Calibre/online), but around the pictures and author sidebars, the results were completely unacceptable.
Examples of sidebar/pics:

I'll just cut this short. There is NO SUCH THING as an automagic conversion for a book that has wraparound articles, sidebars, floating images, etc. In trying to achieve this, you are causing yourself untold amounts of unnecessary work--all created in the false hope that you'll have LESS work. This is not the case.

You would have been FAR better off just going straight from Word. You will face this extra work one way or the other. It's not necessarily "less work" to do the ePUB from Word, but it's LESS CRUFT to clean up. When you try to invent a process, as you have, you are not seeing the iceberg. By which I mean, you think you are seeing progress, on the surface; but the ungodly mess that you're not seeing, under the hood, will come back and wreak havoc when you try to put this on any real device.

You're spinning your wheels, chasing ephemera that isn't real. What you think you're seeing, in BookView, isn't what you'll see down the road.

OH, and by the way: using freebie online converters or calibre isn't helping you, either. Not the way you're using them.

BUT, I'll shut up now. As sure as shooting, you'll keep trying the methods that SEEM to produce the result you want. I'm really glad I'm not the one cleaning up the undersides of that file.

Hitch