XML Reference Guide

Now that we've discussed just what it is we can do with Yahoo! Web Services, it's time to actually do it. In this guide section, we'll build, from scratch, a PHP form that builds a Yahoo! query, sends it off, and displays the results based on the type of search.

Our intent is to build a small "module," possibly intended to be displayed on a second page, that displays the top 10 results for a given keyword. To keep things simple, we won't cover local searches, but the ideas are the same.

We'll start with a form that takes the keyword and service information:

   <h3>Yahoo! Results</h3>

   <form action="yahoo.php" method="get">

      Choose a keyword: <input type="text" name="keyword" /><br />
      Choose a service: <select name="service">
         <option value="web">web</option>
         <option value="image">images</option>

         <option value="video">video</option>
         <option value="news">news</option>
      </select>
      <input type="submit" />      
   </form>

The form creates a page much like figure 1:

When we submit the form, the first thing we need to do is extract the submitted information:

<?php

$keyword=$HTTP_GET_VARS["keyword"];
if ($keyword == ""){
?>
   <h3>Yahoo! Results</h3>

   <form action="yahoo.php" method="get">
      Choose a keyword: <input type="text" name="keyword" /><br />
      Choose a service: <select name="service">
         <option value="web">web</option>

         <option value="image">images</option>
         <option value="video">video</option>
         <option value="news">news</option>
      </select>

      <input type="submit" />      
   </form>

<?php
} else {

   $searchService = $HTTP_GET_VARS["service"];
   if ($searchService == "") {
       $searchService = "web";
   }
}
?>

First, we're checking to see whether the form's been submitted, based on whether or not there's a keyword value. If not, we're displaying the form. If so, we're getting the service value. Just to be on the safe side, if one wasn't submitted, we'll use the web search service.

Next we need to build the URL for the request. To do that, we need to take into account that each service has its own "base" URL:

<?php

function getRequestURL($baseURL, $keyword){

   $otherArgs="&appid=InformITXML&results=10";
   return $baseURL."?query=".$keyword.$otherArgs;

}


$keyword=$HTTP_GET_VARS["keyword"];
if ($keyword == ""){
...
} else {

   $searchService = $HTTP_GET_VARS["service"];
   if ($searchService == "") {
       $searchService = "web";
   }
   $baseURLs = array(
       "web" => "http://api.search.yahoo.com/WebSearchService/V1/webSearch", 
       "video" => "http://api.search.yahoo.com/VideoSearchService/V1/videoSearch", 
       "news" => "http://api.search.yahoo.com/NewsSearchService/V1/newsSearch", 
       "image" => "http://api.search.yahoo.com/ImageSearchService/V1/imageSearch");

   $fullURL = getRequestURL($baseURLs[$searchService], $keyword);

}

Let's start at the bottom. The actual request consists of a URL. The URL is made up of a "base" url and various parameters. We can pull that base URL from an array based on the search service. From there we move up to the getRequestURL() function.

The getRequestURL() function builds the complete URL based on the base URL, which we're passing in, the query, or keyword, which we're also passing in, and any additional arguments. In this case, we're passing only the (required) appid and the results parameter, specifying that we only want 10 results. If we were including the Local Search service in our form, we would also have to accomodate the radius and other location-based parameters. Finally, we combine them all and send back the results.

Now we're ready to actually send the request. Fortunately, PHP's Document Object Model implementation makes that part easy:

...
   $fullURL = getRequestURL($baseURLs[$searchService], $keyword);

   $document = domxml_open_file($fullURL);
   $root = $document->document_element();

   $totalResults = 0;
   $thisPageResults = 0;

   if ($root->has_attributes()) {
       $attributesArray = $root->attributes();
       foreach ($attributesArray AS $thisAttribute) {
          $attributeName = $thisAttribute->name();
          if ($attributeName == "totalResultsAvailable") {
              $totalResults = $thisAttribute->value();
          } elseif ($attributeName == "totalResultsReturned"){
              $thisPageResults = $thisAttribute->value();
          }
       }
   }

   echo ("<h3>Top ".$thisPageResults." Yahoo! results for \"".$keyword."\"</h3>");
   echo ("<p>Total Results: ".$totalResults."</p>");

}

PHP's DOM implementation let's us parse a document at an external URL, so we don't need to fool around with making and reading HTTP requests. Instead, we get a reference to the root element, in this case ResultSet. From there, we'll extract the totalResultsAvailable and totalResultsReturned attributes of the ResultSet element so we know whether there are any results to analyze.

Assuming there are, we'll handle that next:

...
   echo ("<h3>Top ".$thisPageResults." Yahoo! results for \"".$keyword."\"</h3>");
   echo ("<p>Total Results: ".$totalResults."</p>");
   
   $thisResult = $root->first_child();
   while ($thisResult){

      $titleElement = $thisResult->get_elements_by_tagname("Title");
      $titleValue = $titleElement[0]->first_child();
      $title = $titleValue->node_value();

      $click_urlElement = $thisResult->get_elements_by_tagname("ClickUrl");
      $click_urlValue = $click_urlElement[0]->first_child();
      $click_url = $click_urlValue->node_value();

      $display_urlElement = $thisResult->get_elements_by_tagname("Url");
      $display_urlValue = $display_urlElement[0]->first_child();
      $display_url = $display_urlValue->node_value();

      $summaryElement = $thisResult->get_elements_by_tagname("Summary");
      $summaryValue = $summaryElement[0]->first_child();
      if ($summaryValue){
           $summary = $summaryValue->node_value();
      }
...
      $thisResult = $thisResult->next_sibling();

   }

}

The actual results come back as Result element children of the ResultSet element, so we'll start by getting the first child of the ResultSet element and analyzing it to pull the Title, ClickUrl (the URL the user should be directed to), Url (the actual URL of the resource, and not a Yahoo! redirect), and the Summary. Some results don't actually have a summary, so we'll check that before we try to access the value.

Finally, we'll move to the next sibling so we don't create an infinite loop; when we run out of children, the while() loop stops.

Once we have the information, we can display it on the page:

...
      $summaryValue = $summaryElement[0]->first_child();
      if ($summaryValue){
           $summary = $summaryValue->node_value();
      }

      if ($searchService == "image"){
          $thumbnailElement = $thisResult->get_elements_by_tagname("Thumbnail");
          $thumbnailURLElement = $thumbnailElement[0]->first_child();
          $thumbnailURLValue = $thumbnailURLElement->first_child();
          $thumbnail = $thumbnailURLValue->node_value();

          echo ("<a href=\"".$click_url."\">");
          echo ("<img src=\"".$thumbnail."\" />");
          echo ("</a><br /><br />");
      } else {

          echo ($title);
          echo ("<br />");
      
          echo ("<a href=\"".$click_url."\">");
          echo ($display_url);
          echo ("</a><br />");

          echo ($summary."<br /><br />");
      }

      $thisResult = $thisResult->next_sibling();

   }
}

How we actually display the information depends on what it is. If we're dealing with an Image Search, we'll want to display the thumbnail, which also means extracting extra information from the Thumbnail element. Otherwise, we're simply displaying the title, URL (both the click and display variety) and the summary.

The results look like Figures 2 and 3:

The implementation in other languages will, of course, be different, but the concept is the same: get the query information, build the URL for the request, and pull the important information out of the results.