pdf documents

pdf documents

[Daniel Dalton]

Hi,

Does anyone know if it is possible to read pdf documents in a console with 
speakup?
I don't want to have to use a gui.
If so what application should I use?

Thanks for any help,

--
Daniel Dalton

http://members.iinet.net.au/~ddalton/
d.dalton@iinet.net.au

Re: pdf documents

[Tony Baechler]

Daniel Dalton wrote:
> Does anyone know if it is possible to read pdf documents in a console 
> with speakup?

Yes.

> I don't want to have to use a gui.

So don't use one.


> If so what application should I use?


Try the pdftotext utility from the xpdf package.  You didn't say what 
distro you run but I know it's prepackaged for Debian and probably in 
.rpm.  You could of course compile from source as well.  Note that 
pdftotext won't read files with the no copy and no print bits set.  It 
dumps the pdf file to the same name with a .txt extension which you can 
read with less(1) or more(1).  It might overwrite already existing files 
with the same name and .txt extension as the pdf file you convert.  You 
don't need the other programs in the package.  You used to get it from 
http://www.foolabs.com/ but this might have changed.  You can always get 
source from Debian ftp mirrors.

Re: pdf documents

[Matt Barnes]

[-- Attachment #1: Type: text/plain, Size: 795 bytes --]

Tesseract is an OCR and can convert pdf's and images to text. I haven't
gotten around to installing it and trying it out, but it seems like the OCR
of choice, located here:
http://sourceforge.net/project/showfiles.php?group_id=158586



On Dec 27, 2007 4:54 AM, Daniel Dalton <d.dalton@iinet.net.au> wrote:

> Hi,
>
> Does anyone know if it is possible to read pdf documents in a console with
> speakup?
> I don't want to have to use a gui.
> If so what application should I use?
>
> Thanks for any help,
>
> --
> Daniel Dalton
>
> http://members.iinet.net.au/~ddalton/<http://members.iinet.net.au/%7Eddalton/>
> d.dalton@iinet.net.au
>
> _______________________________________________
> Blinux-list mailing list
> Blinux-list@redhat.com
> https://www.redhat.com/mailman/listinfo/blinux-list
>

[-- Attachment #2: Type: text/html, Size: 1351 bytes --]

Re: pdf documents

[Tony Baechler]

Yes, but unless I'm badly mistaken, it is very old and doesn't support 
directly extracting images from pdf files.  You would still need to 
install the xpdf package to get the pdfimages utility so you can process 
the images as single files.  I read about the OCR package you describe 
but I'm fairly sure it's old and unmaintained.  Maybe someone was going 
to take over development, I'm not sure.  I've noticed that most pdf 
files are text and don't have page images, or if they do, the images are 
pictures so would be useless anyway.  Also, what is the accuracy rate 
for this OCR package?  What about accessibility?

Matt Barnes wrote:
> Tesseract is an OCR and can convert pdf's and images to text. I 
> haven't gotten around to installing it and trying it out, but it seems 
> like the OCR of choice, located here:
> http://sourceforge.net/project/showfiles.php?group_id=158586

Re: pdf documents

[Geoff Shang]

Hi,

In addition to pdftotext, I can also recommend pstotext (in Debian it's in 
its own package).

pstotext outputs to standard output by default.   You can save the output 
by either using > to redirect to a file, or by using the -output command 
line 
option.

I have both pstotext and pdftotext installed here.  Results seem to vary as 
to which is better and you may want to try both if a document is proving 
difficult to read and see which gives the best results.

Geoff.


-- 

Re: pdf documents

[Tony Baechler]

Hi,

Have you experimented with the pdftotext -layout and -raw options?  I've 
noticed that usually no command line options produces usable output but 
sometimes using the -raw command line optin works better.  Then again, 
sometimes it makes it worse.  There's generally a decrease in the output 
file size when using the -raw option.  Also, since presumably the 
interest of most people here is in making pdf documents accessible or 
somehow finding a way to use the text from them, do you know of a way to 
get around the protection bits?  I'm not trying to pirate or anything 
else, but I know of at least one major audio editing package which ships 
manuals that can't be read with pdftotext because of the no-print and 
no-copy bits.  I know the text is there in the pdf file because they 
sent me an unprotected copy upon request, but they have been bought out 
by a major media company now.

Geoff Shang wrote:
> I have both pstotext and pdftotext installed here.  Results seem to 
> vary as to which is better and you may want to try both if a document 
> is proving difficult to read and see which gives the best results.

Re: pdf documents

[Mike Gorse]

pdftotext has a few lines of code that simply check the permission bits, 
and, if they are set to disallow copying, print an error message and exit.

-- Mike Gorse / AIM:linvortex / http://mgorse.freeshell.org --

Re: pdf documents

[Tony Baechler]

Fine, but that doesn't help us non-programmers like me.  I admit that I 
haven't looked at the source but I wouldn't know what I'm looking at 
anyway.  I took a brief look but it meant nothing to me.  Again, I'm not 
trying to pirate or anything else, I just want equal access to pdf 
documents without using an OCR package.  The OCR package I used in the 
past created errors that weren't in the original document.

Mike Gorse wrote:
> pdftotext has a few lines of code that simply check the permission 
> bits, and, if they are set to disallow copying, print an error message 
> and exit.

Re: pdf documents

[Mike Gorse]

If you know the error string that gets displayed, then using grep -r 
should point you to the relevant code.  Actually, it doesn't look as 
though my version of pdftotext enforces permissions (I have the version 
included with poppler, which is a fork of xpdf made into a library and 
used by evince and possibly other projects).  It will enforce the 
permissions if ENFORCE_PERMISSIONS is defined at compile time, but it is 
not by default.

-- Mike Gorse / AIM:linvortex / http://mgorse.freeshell.org --

On Sat, 29 Dec 2007, Tony Baechler wrote:

> Fine, but that doesn't help us non-programmers like me.  I admit that I 
> haven't looked at the source but I wouldn't know what I'm looking at anyway. 
> I took a brief look but it meant nothing to me.  Again, I'm not trying to 
> pirate or anything else, I just want equal access to pdf documents without 
> using an OCR package.  The OCR package I used in the past created errors that 
> weren't in the original document.