Erase Page? Numbers after heuristic corrections

[msshain]

The heuristic processing worked great to unify paragraphs converting PDF to ePub. I am getting numbers at various intervals though. Please see example (24 & 25) below:

nonrealistic view suggested by quantum theory. 24 Einstein protested: “I cannot seriously believe in [the quantum theory] because it cannot be reconciled with the idea that physics should represent a reality in time and space, free from spooky actions at a distance.” 25 It was in a discussion of the EPR paper that Erwin Schrödinger first coined the term “entanglement.”

Any ideas how to omit these, thanks.

[eschwartz]

Sounds like page numbers. You will need to add a regex under Search and Replace to get rid of those.

[theducks]

Originally Posted by msshain

The heuristic processing worked great to unify paragraphs converting PDF to ePub. I am getting numbers at various intervals though. Please see example (24 & 25) below:

nonrealistic view suggested by quantum theory. 24 Einstein protested: “I cannot seriously believe in [the quantum theory] because it cannot be reconciled with the idea that physics should represent a reality in time and space, free from spooky actions at a distance.” 25 It was in a discussion of the EPR paper that Erwin Schrödinger first coined the term “entanglement.”

Any ideas how to omit these, thanks.

Your example is page numbers embedded within normal text (A very bad OCR).

This is a slightly tedious EDITOR job, not a conversion job.

REGEX in a conversion expects a FIXED pattern to the Page # appearance.
Long Winded 56
57 Short Story
Long Winded 103
104 Short Story

When it is (semi) random, you need to step through each find (there will be many patterns to find. you create a unique REGEX for each pattern you discover.

BTW This is probably a case to NOT have Heuristics clean up. The page pattern might have been easier to discover before the attempt to join lines. Every PDF is unique in the issues presented (see the sticky about PDF)

[eschwartz]

theducks -- if you use S&R in the conversion settings, it operates before line unwrapping. Handy.

Of course, you lose the ability to step through each match and confirm.

There are advantages either way.

[theducks]

Originally Posted by eschwartz

theducks -- if you use S&R in the conversion settings, it operates before line unwrapping. Handy.

Of course, you lose the ability to step through each match and confirm.

There are advantages either way.

I am a WYSWYG make a REGEX kind of guy, so I use a Editor and forgo the

[eschwartz]

You do actully get to preview the parsed xhtml that is extracted from the PDF -- it is part of the S&R wizard. So it isn't as dangerous as it could be.

[cybmole]

Zapping numbers without manual checks is bad for dates, and guns, and times
E.g. He was shot in '66 with a colt 45. At 11am
We called 911 but.....

[BetterRed]

Actually they could be reference link numbers - is there a numbered reference list at the back of the book, and do the numbers embedded in the text bear any relationship to the reference with the same number.

I suspect those quotes may be from the correspondence between E and S on the latter's thought experiments on cats in boxes and all that, and E's statement claiming God doesn't play dice etc.

If they are such - then you might want to 'fix' them in the editor by recreating the links.

BR

[eschwartz]

Originally Posted by cybmole

Zapping numbers without manual checks is bad for dates, and guns, and times
E.g. He was shot in '66 with a colt 45. At 11am
We called 911 but.....

The right regex would look for numbers wrapped in the pre-tested footer XHTML.