class.attachment.file.php

<?php

/**
 * This file was created to be used as an "include" in the remove_old_tickets.php mod.
 */

class TDJ_Attachment_File {
    public $uri, $str, $type, $ext;
    function TDJ_Attachment_File($attachment) {
        $this->uri = array ();
        $this->str = '';

        $id = $attachment ['attach_id'];
                                         // Make the folder if it doesn't exist.
        $dir = URI_FOLDER;
        if (! is_dir ( $dir )) {
            error_log ( "Making dir: $dir" );
            mkdir ( $dir );
        }
        //Locate osTickets attachment object for this id
        $a = Attachment::lookup ( $id );
        //Get the file object itself, which is also an osTicket object/db entity etc.
        $f = $a->getFile ();

        // Route by attachment I wish I could remember where I found this majestic bitty.. ;-)
        $this->ext = $ext = strtolower ( substr ( strrchr ( $f->getName (), '.' ), 1 ) );

        $filename = $dir . DIRECTORY_SEPARATOR . "t_$id.$ext";

        switch ($ext) {
            case 'pdf' :
                {
                    if (file_exists ( $filename ) != TRUE) {

                        file_put_contents ( $filename, $f->getData () );
                    }
                    $this->uri = $filename;
                    $this->type = 'pdf';
                    break;
                }
            case 'txt' :
            case 'htm' :
            case 'html' :
                {
                    $this->type = 'text';
                    $this->str = ($ext == 'txt') ? '<pre>' . $f->getData () . '</pre>' : Format::safe_html ( $f->getData () );
                    // we don't set the URI here, as we are using the string value itself. No need to create a file.
                    $this->uri = false;
                    break;
                }
            case 'png' :
                {
                    if(!GD_ENABLED)
                        break;
                    //PDF doesn't support alpha channels for some reason, so we need to quickly and painlessly remove them.
                    $tmp = tempnam('/tmp/osticket','_png_converter_');
                    $jpeg_version = $this->png2jpg($filename, $tmp);

                    //need to change its file-extension to match its new colorful format ;-)
                    $this->ext = $ext = 'jpg';
                    $filename = str_replace('.png', '.jpg', $filename);
                    rename($tmp, $filename);
                    //don't break, let the image section continue to work.
                }
            case 'jpg' :
            case 'gif' :
            case 'jpeg' :
            case 'tif' :
                {
                    $this->type = 'image';
                    if (file_exists ( $filename ) != TRUE) {
                        file_put_contents ( $filename, $f->getData () );
                    }
                    $this->uri = $filename;
                    break;
                }
            case 'doc' :
            case 'docx' : // stupid word docs
            case 'xls' :
            case 'xlsx' : // stupid excel docs
                {
                    $this->type = 'text';
                    if (file_exists ( $filename ) != TRUE) {
                        file_put_contents ( $filename, $f->getData () );
                    }
                    // We don't really have time to format this shite, but we can at least grok some of the contents.. ;-)
                    $docObj = new DocxConversion ( $filename );
                    $this->str = $docObj->convertToText ();
                    break;
                }
            default :
        }
    }

    /**
     * I'm starting to wonder how much is original code, and how much I just ripped off from StackOverflow!
     *
     * from: <a href="http://stackoverflow.com/a/1201823/276663">StackOverflow!</a>
     * @param string $originalFile The full path to the PNG file.
     * @param string $outputFile where does he go?
     * @param int $quality (default should be 75 on imagejpeg anyway)
     */
    function png2jpg($originalFile, $outputFile, $quality=75) {
            $image = imagecreatefrompng($originalFile);
            imagejpeg($image, $outputFile, $quality);
            imagedestroy($image);
    }
}


/**
 *
 * Here is simple and easily understandable class which does the right job for .doc/.docx , PHP docx reader: Convert MS Word Docx files to text.
 * http://www.phpclasses.org/package/7934-PHP-Convert-MS-Word-Docx-files-to-text.html
 * http://stackoverflow.com/a/19503654
 *
 */
class DocxConversion{
    private $filename;

    public function __construct($filePath) {
        $this->filename = $filePath;
    }

    private function read_doc() {
        $fileHandle = fopen($this->filename, "r");
        $line = @fread($fileHandle, filesize($this->filename));
        $lines = explode(chr(0x0D),$line);
        $outtext = "";
        foreach($lines as $thisline)
        {
            $pos = strpos($thisline, chr(0x00));
            if (($pos !== FALSE)||(strlen($thisline)==0))
            {
            } else {
                $outtext .= $thisline." ";
            }
        }
        $outtext = preg_replace("/[^a-zA-Z0-9\s\,\.\-\n\r\t@\/\_\(\)]/","",$outtext);
        return $outtext;
    }

    private function read_docx(){

        $striped_content = '';
        $content = '';

        $zip = zip_open($this->filename);

        if (!$zip || is_numeric($zip)) return false;

        while ($zip_entry = zip_read($zip)) {

            if (zip_entry_open($zip, $zip_entry) == FALSE) continue;

            if (zip_entry_name($zip_entry) != "word/document.xml") continue;

            $content .= zip_entry_read($zip_entry, zip_entry_filesize($zip_entry));

            zip_entry_close($zip_entry);
        }// end while

        zip_close($zip);

        $content = str_replace('</w:r></w:p></w:tc><w:tc>', " ", $content);
        $content = str_replace('</w:r></w:p>', "\r\n", $content);
        $striped_content = strip_tags($content);

        return $striped_content;
    }

    /************************excel sheet************************************/

    function xlsx_to_text($input_file){
        $xml_filename = "xl/sharedStrings.xml"; //content file name
        $zip_handle = new ZipArchive;
        $output_text = "";
        if(true === $zip_handle->open($input_file)){
            if(($xml_index = $zip_handle->locateName($xml_filename)) !== false){
                $xml_datas = $zip_handle->getFromIndex($xml_index);
                $xml_handle = DOMDocument::loadXML($xml_datas, LIBXML_NOENT | LIBXML_XINCLUDE | LIBXML_NOERROR | LIBXML_NOWARNING);
                $output_text = strip_tags($xml_handle->saveXML());
            }else{
                $output_text .="";
            }
            $zip_handle->close();
        }else{
            $output_text .="";
        }
        return $output_text;
    }

    /*************************power point files*****************************/
    function pptx_to_text($input_file){
        $zip_handle = new ZipArchive;
        $output_text = "";
        if(true === $zip_handle->open($input_file)){
            $slide_number = 1; //loop through slide files
            while(($xml_index = $zip_handle->locateName("ppt/slides/slide".$slide_number.".xml")) !== false){
                $xml_datas = $zip_handle->getFromIndex($xml_index);
                $xml_handle = DOMDocument::loadXML($xml_datas, LIBXML_NOENT | LIBXML_XINCLUDE | LIBXML_NOERROR | LIBXML_NOWARNING);
                $output_text .= strip_tags($xml_handle->saveXML());
                $slide_number++;
            }
            if($slide_number == 1){
                $output_text .="";
            }
            $zip_handle->close();
        }else{
            $output_text .="";
        }
        return $output_text;
    }


    public function convertToText() {

        if(isset($this->filename) && !file_exists($this->filename)) {
            return "File Not exists";
        }

        $fileArray = pathinfo($this->filename);
        $file_ext  = $fileArray['extension'];
        if($file_ext == "doc" || $file_ext == "docx" || $file_ext == "xlsx" || $file_ext == "pptx")
        {
            if($file_ext == "doc") {
                return $this->read_doc();
            } elseif($file_ext == "docx") {
                return $this->read_docx();
            } elseif($file_ext == "xlsx") {
                return $this->xlsx_to_text();
            }elseif($file_ext == "pptx") {
                return $this->pptx_to_text();
            }
        } else {
            return "Invalid File Type";
        }
    }

}
?>