Zend_Search_Lucene question

iroy2000 · #1 (**permalink**) 03-17-2008, 11:03 PM

I was testing on the Zend_Search_Lucene, and I can't get this work properly, here is some of my testing code

I call this function in my indexAction

PHP Code:


			
    protected function buildDocument() {
        $feed = 'http://feeds.feedburner.com/ZendDeveloperZone';                
        $this->_channel = Zend_Feed::import($feed);        
        
        foreach($this->_channel->items as $item) {
            
            echo $item->link();
            
            $doc = new Zend_Search_Lucene_Document();
            
            $doc->addField(
                Zend_Search_Lucene_Field::Keyword('link', $this->_injectFilter->filter($item->link()))
            );
                
            $doc->addField(
                Zend_Search_Lucene_Field::Text('title', $this->_injectFilter->filter($item->title()))
            );    
            
            $doc->addField(
                Zend_Search_Lucene_Field::UnStored('contents', $this->_injectFilter->filter($item->description()))
            );
            
            $this->_index->addDocument($doc);
        }    

        $this->_index->commit();
    }

And then I try to fetch the result from my searchAction()

PHP Code:


			
    public function searchAction() {
         $index = Zend_Search_Lucene::create('/path/to/index');
         $this->_helper->viewRenderer->setNoRender(true);
         
         echo "Index contains {$index->count()} documents.\n";
         
         $query = 'framework'; // I try several other keywords as well, but still not working
        
          $hits = $index->find($query);
        
         echo "Search for '".$query."' returned " .count($hits). " hits<br /><br />";
        
         foreach($hits as $hit) {
             echo $hit->title . '<br />';
             echo $hit->link . '<br /><br />';
         }         
    }

But I got
Index contains 0 documents.
Search for 'framework' returned 0 hits

When I check my indexing folder, I saw some files being created
segements.gen
_2.cfs and some other segements

Can someone tell me what have I done wrong or any suggestion, thanks for any help!

[ Note: the $this->_index is from Zend_Search_Lucene::create('/path/to/index'); ]

xentek · #2 (**permalink**) 03-18-2008, 05:56 PM

Quote:

$index = Zend_Search_Lucene::create('/path/to/index');

You need to create/open your index before you can start indexing documents. Your index path should also be a real system path on your system that the web server can write to.

I created a static method that I can call to create/open my index:

Replace $GLOBALS['index_path'] with your real system path (I'm loading mine up from my config file).

PHP Code:


			
    public static function getIndex()
    {
        if (!file_exists($GLOBALS['index_path'])) {
            $index = Zend_Search_Lucene::create($GLOBALS['index_path']);
        } else {
            $index = Zend_Search_Lucene::open($GLOBALS['index_path']);
        }
        Zend_Search_Lucene_Analysis_Analyzer::setDefault(new Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8Num());
        return $index;
    }

Then when ever you need to add a document or perform a query, you can call this method first and set it to your $index var, for all index operations (find, addDocument, etc)

Also, this tool can be handy to test your index out, though I think there are some differences between the two (result sets may vary, but is a decent enough way to test some things - just test them in your Zend_Search_Lucene as well): http://www.getopt.org/luke/

iroy2000 · #3 (**permalink**) 03-19-2008, 12:20 AM

Thanks for the reply, very helpful.

I tried to make a getSearchIndex() and it works. But now there's another problem...

The first time I run, it calls the "create()", and after that it calls "open()". It all make sense here.

So I tried to reload several times for testing, but each time of reload, it adds more "same" queries into the indexing.

Do I need to delete them before updating, or there is a better way to solve that problem??

xentek · #4 (**permalink**) 03-19-2008, 01:52 PM

It's probably a good idea to delete what got started, and then run the getSearchIndex() method to create it the first time.

Then start to index your documents, and get your index filled with data. Use the Luke tool to ensure the index is readable and you can run a couple of queries against it for known data to ensure it's working.

Once you have your index populated, move on to creating your front end search methods and use them to use Zend_Search_Lucene to query your index. Once you're getting some hits back, dress up your results page, and you should be good to go.

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode