[RESOLVED] shell_exec and pdftotext does not work?

[said_fox]

Hi,
I don't know why the following does not work:

echo shell_exec('/usr/bin/pdftotext'); 

It does not output anything, where in terminal it outputs the usage of pdftotext. On the other hand, some command such as dir give output correctly.

The path of pdftotext is correct because I got it by the command "which". The PHP I have is XAMPP on Ubuntu 11.10.

Another thing, If I inserted wrong command, it does not output any error message too!

[rtrethewey]

I believe you need to pass a filename and path to pdftotext. Your code just causes the program to run without any input data to convert. It should probably be something like:

echo shell_exec('/usr/bin/pdftotext path/filename.pdf'); 

[said_fox]

Unfortunately, this does not solve the problem.

[007Julien]

I use pdftotext on a local server with this function

function pdfTxt($path){
  $o=shell_exec('pdftotext -enc UTF-8 '.$path.' pdf.txt');
    $c=file_get_contents('pdf.txt');
// Some corrections depending on your pdfs
    return preg_replace("@\x0D\x0A\x0D\x0A\.\x0D\x0A\x0D\x0A\.\x0D\x0A\x0D\x0A[^\x0D]+\x0D\x0A\x0D\x0A[^\x0D]+\x0D\x0A\x0D\x0A[^\x0D]+\x0D\x0A\x0D\x0A@"," ",$c);
} 

pdftotext is in same directory as the running php file, $path is the pdf file relative path. Without pdf.txt the new txt file take the name of your file.
Download it on this page and read the pdftotext.txt file for other options...
Good luck !

[said_fox]

I use pdftotext on a local server with this
pdftotext is in same directory as the running php file, $path is the pdf file relative path. Without pdf.txt the new txt file take the name of your file.
Download it on this page and read the pdftotext.txt file for other options...
Good luck !

I may misunderstand you. I already have on my local Linux machine pdftotext and it run from terminal. As I think it is belong to package called poppler.

Do you mean that xpdf is another package than poppler and it do its functionality?

[007Julien]

I work on a local server with windows. I have fund pdftotext.exe in the package xpdfbin-win-3.03.zip with other files like pdftoppm.exe, pdfimages.exe, ... and her descriptions, options in pdftotext.txt, pdftoppm.txt, pdfimages.txt ...

I do not know poppler but read on this page that Poppler is a PDF rendering library based on the xpdf-3.0 code base.

[eval(BadCode)]

In bash/"the terminal" there are things called file pointers. When you run any given command, there are at least THREE file pointers, called 1 2 3.

1 is standard output "STDOUT"
2 is error output "STDERR"
3 is standard input "STDIN"

When you run this:

$o=shell_exec('pdftotext -enc UTF-8 '.$path.' pdf.txt'); 

you are getting ONLY standard output. But what if there is no standard output?

Maybe you have an ERROR. And you really want to be looking at both STDOUT + STDERR.

You can redirect the pointer for STDERR to STDOUT by adding this to your bash/sh command:
2>&1

$o=shell_exec('pdftotext -enc UTF-8 '.$path.' pdf.txt 2>&1'); 

Now when you run this:

$path = 'asd9ajs09dj039jadad'; //doesn't exist
echo shell_exec('pdftotext -enc UTF-8 '.$path.' pdf.txt 2>&1'); 

pdf2text: command not found

But, for you it will probably be something like
Could not open input file: asd9ajs09dj039jadad

Taken from my own code:

$cmdStr = CONFIG::PATH_TO_PYTHON2.' '.CONFIG::PATH_TO_PYTHON_PARSER.' '.$file->getDiskPath().' 2>&1'; //also send stderr to stdout
    exec($cmdStr, $buffer); //send command to terminal, capture output in a buffer 

You can also make your own custom file pointers or perhaps you are familiar with "piping" standard output into another programs standard input:


echo ASDF_ASDF_ASDF_ASDF | tr 'S' "D"
outputs:
ADDF_ADDF_ADDF_ADDF

echo ASDF_ASDF_ASDF_ASDF | tr 'S' "D" | tr 'D' "F"
outputs:
AFFF_AFFF_AFFF_AFFF

hope this helps some

[said_fox]

Thank you for your valuable hint.
Now I able to see the error:

/usr/bin/pdftotext: /opt/lampp/lib/libjpeg.so.62: no version information available (required by /usr/lib/libpoppler.so.13) /usr/bin/pdftotext: /opt/lampp/lib/libgcc_s.so.1: version `GCC_4.2.0' not found (required by /usr/lib/i386-linux-gnu/libstdc++.so.6) 

I think it's something related to XAMPP for linux, If you know how to solve it, I will be appreciated.

[eval(BadCode)]

I am not entirely sure what .so is, but I can tell you that the error is complaining about GCC missing- which is REALLY ODD considering you're running a program calld pdf2text and not compiling anything.

I would send the error to the system admin, it doesn't look like something you can fix without being root.

GCC is a `C` compiler
PHP is written in C also (it's a programming language).

Edit:
but first, make sure you can't run the actual program pdf2text from a real shell. Use putty if you're on windows, otherwise open a terminal and use ssh user@host.com or just man ssh and it will explain it.

[said_fox]

pdftotext is running well from shell (Terminal).

[criterion9]

pdftotext is running well from shell (Terminal).

Have you tried running it as the Apache/www-data/php user (whichever your system has configured? It is possible that the user running the php doesn't have the same access as your user account.

[said_fox]

Have you tried running it as the Apache/www-data/php user (whichever your system has configured? It is possible that the user running the php doesn't have the same access as your user account.

I think that the problem is not related to the access, the error message shows that it is about something missing. However, how could I know the user that running PHP?

[eval(BadCode)]

exec("whoami",$buffer);
echo implode($buffer); 

[criterion9]

I think that the problem is not related to the access, the error message shows that it is about something missing. However, how could I know the user that running PHP?

Is is possible that your user account has different classpath (or whatever it is called in your environment) than the account that runs PHP.

[said_fox]

Is is possible that your user account has different classpath (or whatever it is called in your environment) than the account that runs PHP.

How could I determine this and how could I fix it?