/* Convert pdf to text. 
 *
 * 	Last modified: TS, 24/6/2009
 *
 * 	Strips non-printable characters.
 * 	Checks that filename has extension .pdf and is actually a valid file
 *	HTML entities not encoded (as this is intended to be used with my search snippet which will encode the extract anyway)
 *
 * 	Requires pdftotext as found in xpdf-utils
 * 	Requires docmanager.class.php
 *
 * for OnDocFormSave
 */
 
/* ----- CONFIG ----- */
define('SHELL_COMMAND_FIRST', 'pdftotext '); // First part of shell command - before filename (includes following space)
define('SHELL_COMMAND_LAST', ' -'); // Second part of shell command - after filename (includes preceeding space)
define('TEXT_TV', 'downloadText'); // modx TV to hold text
define('PDF_FILE_TV', 'download'); // modx TV that points to PDF file
/* ----- ------ ----- */


require_once($modx->config['base_path'].'assets/libs/docmanager/document.class.inc.php');

$doc_tvs = $modx->getTemplateVarOutput(array(PDF_FILE_TV), $id);
$pdf_full_file_path = $modx->config['base_path'].$doc_tvs[PDF_FILE_TV];

if ($doc_tvs[PDF_FILE_TV] && strtolower(substr($pdf_full_file_path, -4)) == '.pdf' && is_file($pdf_full_file_path))
	{
	$doc = new Document($id);
	$doc->SetTV(TEXT_TV, mysql_real_escape_string(preg_replace('/[^[:print:]]/', '', shell_exec(SHELL_COMMAND_FIRST.'"'.$pdf_full_file_path.'"'.SHELL_COMMAND_LAST))));

	// FOLLOWING CODE IS VERY IMPORTANT FOR SECURITY!!!
	$unchanged_text_tv_fields = array('pagetitle', 'longtitle', 'menutitle', 'introtext', 'type', 'contentType', 'description', 'alias', 'content', 'link_attributes');
	foreach ($unchanged_text_tv_fields as $field)
		{
		$doc->Set($field, $modx->db->escape($doc->Get($field)));
		}
		
	$doc->Save();
	}