Numbers in pdfs not converting

[kilgoretrout]

I have inserted an example of what Im getting..the "squares" are supposed to be numbers.

The Destruction of Dresden was first published in Great Britain by William Kimber & Co. Ltd on April , ; in a revised and updated edition by Corgi Books Ltd in ; and by Papermac, a division of Macmillan Publishers Ltd, in .

I am using Calibre to translate a pdf to epub. This is from epub. I have sigil bt dont have a clue how to isolate these occurrences of numbers throughout book and use a dif font

I amusing Macbook Pro, ipad.

[Jellby]

It's possible that the PDF is using some custom glyphs in the font for the numbers, instead of the standard number positions. For instance, some fonts include both "old-style" numbers and "normal" numbers, one set is placed in the usual positions for numbers, while the other occupies other positions, and are therefore not recognized as numbers.

[kilgoretrout]

I am sure there are a million possibilities of WHY the numbers aren't converting but is there a way to choose a default font for an entire doc or possibly every occurrence of say...numbers?

[Freeshadow]

the d be lower case numbers i assume
the problem is technically the same as removing ligatures

[frabjous]

Am I the only one who can see the numbers in this quote?

The Destruction of Dresden was first published in Great Britain by William Kimber & Co. Ltd on April , ; in a revised and updated edition by Corgi Books Ltd in ; and by Papermac, a division of Macmillan Publishers Ltd, in .

They're Unicode characters U+F730 throuth U+F739 (for 0 through 9), which is in the "private use area" of Unicode. I don't know which fonts support them, but on my system, the fallback seems to be the Norasi Unicode fonts (designed for Thai).

Yes, their appearance is of "lowercase" or "old style" numerals.

It should be possible to do a find and replace on them using this information. Rather than starting with calibre, you might want to use pdftohtml and pdfreflow to get an html file, and then find and replace in the html and then use calibre to convert to ePub. This is all scriptable. Something like:

#!/bin/bash
filename="$1"
pdftohtml -xml "$filename" 
pdfreflow "${filename%.pdf}.xml"
sed -i -e 'y//0123456789/' "${filename%.pdf}.html"
ebook-convert "${filename%.pdf}.html" "${filename%.pdf}.epub"

[kilgoretrout]

I sent in the exact view of what Im seeing. They are legit fonts. The publisher is aware that the numbers use a dif font.
All I wanna know is THIS:
How doyou isolate a font globally in a pdf or epub..and use a more generic font.
If you have a fix for this..I would like to hear it. Im sure its quite simple,

[kilgoretrout]

That sounds "do-able" think Ill try it!

[kilgoretrout]

OK
Heres what I did..for the Nuremberg.pdf file..

#!/bin/bash
filename="$1"
pdftohtml -xml "$NUREMBERG"
pdfreflow "${NUREMBERG%.pdf}.xml"
sed -i -e 'y//0123456789/' "${NUREMBERG%.pdf}.html"
ebook-convert "${NUREMBERG%.pdf}.html" "${NUREMBERG%.pdf}.epub"

But apparently I havent compiled the reflow thing right. I ran configure, then makefile.

Heres the output I got when I ran the above
-bash: ebook-convert: command not found
MacBookPro:~ wth$

My machine is called Macbookpro and wth of course is my home name.
Thanx for your help.

[Jellby]

"ebook-convert" is part of Calibre, do you have Calibre correctly installed? is it in your PATH?

[frabjous]

Originally Posted by kilgoretrout

OK
Heres what I did..for the Nuremberg.pdf file..

#!/bin/bash
filename="$1"
pdftohtml -xml "$NUREMBERG"
pdfreflow "${NUREMBERG%.pdf}.xml"
sed -i -e 'y//0123456789/' "${NUREMBERG%.pdf}.html"
ebook-convert "${NUREMBERG%.pdf}.html" "${NUREMBERG%.pdf}.epub"

Not sure what $NUREMBERG is, but I had it set up so that it would use whatever you added as first argument to the script. So if you saved the script as fixnumbers.sh you could put in:

fixnumbers.sh nuremberg.pdf

and it should work.

But apparently I havent compiled the reflow thing right. I ran configure, then makefile.

It's just a script. No compiling is necessary. Just save the script in a text file and make it executable.

Heres the output I got when I ran the above
-bash: ebook-convert: command not found
MacBookPro:~ wth$

My machine is called Macbookpro and wth of course is my home name.
Thanx for your help.

As Jellby says, ebook-convert is part of calibre. If it's in your path, it should find it. If not, well, if you remove that line, does it successfully create an html file? If so, you could try opening it in a browser and see if the numbers are fixed, and if so, manually import the html file into calibre for converting to whatever.