<?php
class pdf2txt
{
  var $src;
  var $dest;
  var $data;
  
  // constructor
  function pdf2txt($_src = false, $_dest = false)
  {
    $this->setSource($_src);
    $this->setDestination($_dest);
  }
  
  // set data if no conversion from file nescessary
  function setInput($_data)
  {
    $this->data = $_data;
  }
  
  // sets the source-file
  function setSource($_src)
  {
    $this->src = $_src;
  }
  
  // sets the destination-file
  function setDestination($_dest)
  {
    $this->dest = $_dest;
  }
  
  function directConvert($_data)
  {
    $pdf2txt = new pdf2txt;
    return $pdf2txt->convert($_data);
  }
  
  // convert to pdf
  function convert($_data = false)
  {
    if(false !== $_data)
      $this->data = $_data;
    
    
    if(
      // load from file?
      (false !== $this->src) AND
      // file exists?
      (false === $this->data = file_get_contents($this->src))
    )
    {
      // [ ERROR ]
      // file does not exist
      return false;
    }
    
    if($this->data === false)
    {
      // [ ERROR ]
      // nothing to convert
      return false;
    }
    
    
    // ###############################
    // data available -> start parsing
    // ###############################
    
    // parse encoding
    preg_match('~/Encoding\s*/(\w+)~ism', $this->data, $encoding);
    
    // detect encoding and assume that there is only a single charset for the hole document
    $fromEncoding = 'windows-1252';
    switch($encoding[1])
    {
      case 'MacRomanEncoding':
        $fromEncoding = 'macintosh';
      break;
      
      case 'WinAnsiEncoding':
        // standard encoding
      break;
    }
    
    // parse data
    // the following code ignores the keyword "stream" and "endstream" if they are in a string
    $isStream = false;
    $stream = '';
    $streams = array();
    $openBracketCount = 0;
    $encodedStream = false;
    foreach(preg_split('~(<<\s*/.*?>>\s*stream\s*)|(\s*endstream\s*)|(\()|(\))~ism', $this->data, -1, PREG_SPLIT_DELIM_CAPTURE | PREG_SPLIT_NO_EMPTY) as $k => $part)
    {
      if(preg_match('~<<\s*/(.*?)>>\s*stream\s*~ism', $part, $match))
      {
        $switch = 'stream';
        if(false !== strpos($match[1], '/Filter'))
          $encodedStream = true;
      }
      else
        $switch = trim($part);

      switch($switch)
      {
        case '(':
          if($isStream AND !$encodedStream)
            $openBracketCount++;
        break;
        
        case ')':
          if($isStream AND !$encodedStream)
            $openBracketCount--;
        break;
        
        case 'endstream':
          if($isStream AND $openBracketCount <= 0)
          {
            $isStream = false;
            $streams[] = $stream;
            $stream = '';
            $encodedStream = false;
          }
        break;
      }
      
      if($isStream)
      {
        $stream .= $part;
      }
      
      if($switch == 'stream')
      {
        if($isStream)
          $stream .= $part;
        else
          $isStream = true;
      }
    }
    
    $textObjects = array();
    foreach($streams as $k => $stream)
    {
      // uncompress the stream
      if(false === $uncompressed = @gzuncompress($stream))
        // if nothing to uncompress, assume that the stream is already uncompressed
        $uncompressed = $stream;
      
      // convert to internal encoding UTF-8
      $uncompressed = iconv($fromEncoding, 'UTF-8', $uncompressed);
      
      // replace escaped brackets with placeholders
      $text = str_replace(array('\(','\)','\[','\]'), array('##STARTBRACKET##','##ENDBRACKET##','##STARTSBRACKET##','##ENDSBRACKET##'), $uncompressed);
      
      // parse streams
      // the following code ignores the keyword "BT" and "ET" if they are in a string
      $isTextObj = false;
      $textObject = '';
      $openBracketCount = 0;
      foreach(preg_split('~(\s*BT\s+)|(\s+ET\s+)|(\()|(\))~ism', $text, -1, PREG_SPLIT_DELIM_CAPTURE | PREG_SPLIT_NO_EMPTY) as $k => $part)
      {
        $switch = trim($part);
        switch($switch)
        {
          case '(':
            if($isTextObj)
              $openBracketCount++;
          break;
          
          case ')':
            if($isTextObj)
              $openBracketCount--;
          break;
          
          case 'ET':
            if($isTextObj AND $openBracketCount <= 0)
            {
              $isTextObj = false;
              $textObjects[] = $textObject;
              $textObject = '';
            }
          break;
        }
        
        if($isTextObj)
        {
          $textObject .= $part;
        }
        
        if($switch == 'BT')
        {
          if($isTextObj)
            $textObject .= $part;
          else
            $isTextObj = true;
        }
      }
    }
    
    $return = '';
    foreach($textObjects as $textObject)
    {
      // parse text-objects
      // the following code ignores PDF-keywords if they are in a string
      $isString = false;
      $openBracketCount = 0;
      foreach(preg_split('~(?:\s+(Td|TD|T\*|"|\')\s+)|(\()|(\))~ism', $textObject, -1, PREG_SPLIT_DELIM_CAPTURE | PREG_SPLIT_NO_EMPTY) as $k => $part)
      {
        switch($part)
        {
          // new line
          case 'Td':
          case 'TD':
          case 'T*':
          case '"':
          case "'":
            if(!$isString)
              $return .= "\n";
          break;
          
          case ')':
            if($isString AND $openBracketCount <= 0)
            {
              $isString = false;
              $return .= $string;
              $string = '';
            }
            elseif($isString)
              $openBracketCount--;
          break;
        }
        
        if($isString)
        {
          $string .= $part;
        }
        
        if($part == '(')
        {
          if($isString)
          {
            $openBracketCount++;
          }
          else
          {
            $isString = true;
          }
        }
      }
      
      $return .= "\n";
    }
    
    // substitute the placeholders for the brackets and escape sequences
    $convert = array(
      '##STARTBRACKET##' => '(',
      '##ENDBRACKET##' => ')',
      '##STARTSBRACKET##' => '[',
      '##ENDSBRACKET##' => ']',
      "\\\n" => "\n",
      "\\\r" => "\n",
      "\\\n\r" => "\n",
      "\\\t" => "\t",
      "\\\b" => "\b",
      "\\\f" => "\f",
      '\\\\' => '\\'
    );
    
    // replace octal character codes
    $text = preg_replace_callback(
      '~\\\\([0-8]{3})~',
      create_function(
        '$matches',
        ' if(octdec($matches[1]) > 32)  
            return utf8_encode(chr(octdec($matches[1])));
          else
            return "";
        '
      ),
      $return
    );
    
    // execute conversion with $convert
    $text = strtr(($text), $convert);
    
    if(false !== $this->dest)
      // store $text into the specified destination file
      // and return true on success or false on error
      return false !== file_put_contents($this->dest);
    else
      // return $text
      return $text;
  }
}
?>