OCR processing for incoming faxes (patch included)

[GoogleCodeExporter]

I wanted to be able to get an idea of the content of the fax before opening an 
email attachment, or downloading the PDF from the webinterface.

You need to have tesseract installed:

apt-get install tesseract-ocr tesseract-ocr-eng
(apt-get install tesseract-ocr-deu) ...

Index: fax_to_email.php
===================================================================
--- fax_to_email.php    (revision 2492)
+++ fax_to_email.php    (working copy)
@@ -49,6 +49,9 @@
    ob_end_clean();
    ob_start();

+// OCR "configuration"
+$ocr_lang = "eng";
+
 //add a delimeter to the log
    echo "\n---------------------------------\n";

@@ -155,6 +158,15 @@
                exec($cmd);
            }
        }
+       // if tesseract is installed, create ocr representation of received fax
+       if (!file_exists($dir_fax.'/'.$fax_name.".txt")) {
+           $tmp_tesseract = exec("which tesseract");
+           if (strlen($tmp_tesseract) > 0) {
+               $cmd = "cd ".$dir_fax."; nice ".$tmp_tesseract." 
".$dir_fax."/".$fax_name.".tif ".$dir_fax."/".$fax_name." -l 
".$ocr_lang.">".sys_get_temp_dir()."/lastfax.log 2>&1";
+               echo $cmd."\n";
+               exec($cmd);
+           }
+       }
    }
    else {
        $fax_file_warning = " Fax image not available on server.";
@@ -225,6 +237,13 @@
            if ($fax_retry == 'yes') {
                $tmp_text_plain .= "This message arrived earlier and has been queued until now due to email server issues.\n";
            }
+           if (file_exists($dir_fax."/".$fax_name.".txt")) {
+               // An OCR representation has been found
+               // Include it in the mail body
+               $tmp_text_plain .= "\n ================= My guess on the content of this 
fax: =================\n";
+               $tmp_text_plain .= file_get_contents($dir_fax."/".$fax_name.".txt", NULL, 
NULL, 0, 100000);
+               $tmp_text_plain .= "\n 
========================================================================\n";
+           }
            $tmp_text_html = $tmp_text_plain;

        //prepare the mail object
@@ -271,6 +290,9 @@
                if (file_exists($dir_fax.'/'.$fax_name.".pdf")) {
                    $mail->AddAttachment($dir_fax.'/'.$fax_name.'.pdf'); // pdf attachment
                }
+               if (file_exists($dir_fax.'/'.$fax_name.".txt")) {
+                   $mail->AddAttachment($dir_fax.'/'.$fax_name.'.txt'); // OCR 
representation as TXT file
+               }
                //$filename='fax.tif'; $encoding = "base64"; $type = "image/tif";
                //$mail->AddStringAttachment(base64_decode($strfax),$filename,$encoding,$type);
            }
@@ -337,4 +359,4 @@
 //write the contents of the buffer
    fwrite($fp, $content);
    fclose($fp);

I know, tesseract by far isn't perfect, but for most of our faxes I get an idea 
if it's a signed contract or just a spam fax.

Maybe there is an intelligent way to add this generated text to the 
webinterface (mouseover-popups or something like that).

Regards,

  Johannes Jakob

Original issue reported on code.google.com by lists...@googlemail.com on 3 May 2012 at 6:05

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

Setting this as a feature request.

Original comment by markjcrane@gmail.com on 18 Jul 2014 at 12:46

• Added labels: Type-Enhancement