Re: METS/XML Search facility

Burhan Khalid wrote:

Dan Field wrote:

I'm trying to write an XML Document search engine specifically for METS encoded data ( http://www.loc.gov/standards/*mets*/ ). I've tried knocking together some proof of concept code in PHP4 using both DOM and xml_parser but I've run into dificulties with both. Can anyone point me at some code/class which enables me to search within an XML document for the existance of $string within a set list of tags eg. <METS:title>Some Title Here</METS:title> <METS:description>This is a description of this document</METS:description> It seems like a very straightforward task.

You can do searching from the function that is designated as the character handler (if you are using the xml_* functions). You would just have to keep track of what tag you are checking by setting flags, etc. This is what I tried first off, but I found that I was getting forced breaks within the data for some reason. I believe certain characters caused the parser to stop and parse in multiple passes. I'm probably not explaining it very well, heres the code I was using:

$xml_repository = "/var/www/html/mets/sample/"; $searchable_tags = array("MODS:TITLE"); $current_data = ""; $dir_array = array(); $hits = 0; $MAXHITS = 100; $search_term = "William"; $validtag = false; $found = false;

function startElement($parser, $name, $attrs) {

       global $depth, $searchable_tags, $current_data, $hits, $colour, $found, $search_term, $validtag;
                                                                                                                                              
       if (in_array($name,$searchable_tags)) {
               $validtag = true;
       }

}

function characterData($parser, $data) {

       global $current_data, $found, $validtag, $hits;
       if ($validtag) {
               if (preg_match("/".$search_term."/i",$current_data)) {
                       $found = true;
                       $hits++;
                       $data = trim($data);
                       $current_data .= $data;
               }
       }

}

function endElement($parser, $name) {

       global $depth, $current_data, $found, $hits;
       if ($found) {
               echo "<div>hits:$hits current_data: $current_data</div><br />\n";
               $found = false;
               $current_data = "";
       }

} function search_data($term,$data) {

       if (preg_match($term,$data)) { return true; } else { return false; }

}

// read all the xml files and parse through search engine if ($handle = opendir($xml_repository)) {

       $i = 0;
       while (false !== ($file = readdir($handle))) {
               $dir_array[$i] = $file;
               $i ++;
       }
       closedir($handle);

} for ($i = 2; $i < count($dir_array); $i++) {

       $xml_parser = xml_parser_create("UTF-8");
       xml_parser_set_option ($xml_parser,$XML_OPTION_SKIP_WHITE ,1);
       xml_set_element_handler($xml_parser, "startElement", "endElement");
       xml_set_character_data_handler($xml_parser, "characterData");
                                                                                                                                              
       if (!($fp = fopen($xml_repository.$dir_array[$i], "r"))) {
               die("could not open XML input");
       }
                                                                                                                                              
       while ($data = fread($fp, 4096)) {
               if ($hits >= $MAXHITS) { echo "<p><b>over 100 hits!\n</b> Stopping here.</p>"; exit; }
               if (!xml_parse($xml_parser, $data, feof($fp))) {
                       die(sprintf("XML error: %s at line %d",
                               xml_error_string(xml_get_error_code($xml_parser)),
                               xml_get_current_line_number($xml_parser)));
               }
       }
                                                                                                                                              
       xml_parser_free($xml_parser);

} I tried handing values out from characterData() into a global but this just plain didn't work. Only the concatenation operator seemed to work here and this wasn't what I wanted. -- Dan Field <dof@llgc.org.uk> Support Programmer: Cymru ar y we / Wales on the Web cy_GB: http://www.cymruarywe.org en_GB: http://www.walesontheweb.org