multiterm.php

Bug tracker, and reporter.

<?php
/**
 * Zend Framework
 *
 * LICENSE
 *
 * This source file is subject to the new BSD license that is bundled
 * with this package in the file LICENSE.txt.
 * It is also available through the world-wide-web at this URL:
 * http://framework.zend.com/license/new-bsd
 * If you did not receive a copy of the license and are unable to
 * obtain it through the world-wide-web, please send an email
 * to [e-mail] so we can send you a copy immediately.
 *
 * @category   Zend
 * @package    Zend_Search_Lucene
 * @subpackage Search
 * @copyright  Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
 * @license    http://framework.zend.com/license/new-bsd     New BSD License
 */


/** Zend_Search_Lucene_Search_Query */
require_once 'Zend/Search/Lucene/Search/Query.php';

/** Zend_Search_Lucene_Search_Weight_MultiTerm */
require_once 'Zend/Search/Lucene/Search/Weight/MultiTerm.php';


/**
 * @category   Zend
 * @package    Zend_Search_Lucene
 * @subpackage Search
 * @copyright  Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
 * @license    http://framework.zend.com/license/new-bsd     New BSD License
 */
class Zend_Search_Lucene_Search_Query_MultiTerm extends Zend_Search_Lucene_Search_Query
{

    /**
     * Terms to find.
     * Array of Zend_Search_Lucene_Index_Term
     *
     * @var array
     */
    private $_terms = array();

    /**
     * Term signs.
     * If true then term is required.
     * If false then term is prohibited.
     * If null then term is neither prohibited, nor required
     *
     * If array is null then all terms are required
     *
     * @var array
     */
    private $_signs;

    /**
     * Result vector.
     *
     * @var array
     */
    private $_resVector = null;

    /**
     * Terms positions vectors.
     * Array of Arrays:
     * term1Id => (docId => freq, ...)
     * term2Id => (docId => freq, ...)
     *
     * @var array
     */
    private $_termsFreqs = array();


    /**
     * A score factor based on the fraction of all query terms
     * that a document contains.
     * float for conjunction queries
     * array of float for non conjunction queries
     *
     * @var mixed
     */
    private $_coord = null;


    /**
     * Terms weights
     * array of Zend_Search_Lucene_Search_Weight
     *
     * @var array
     */
    private $_weights = array();


    /**
     * Class constructor.  Create a new multi-term query object.
     *
     * if $signs array is omitted then all terms are required
     * it differs from addTerm() behavior, but should never be used
     *
     * @param array $terms    Array of Zend_Search_Lucene_Index_Term objects
     * @param array $signs    Array of signs.  Sign is boolean|null.
     */
    public function __construct($terms = null, $signs = null)
    {
        if (is_array($terms)) {
            $this->_terms = $terms;

            $this->_signs = null;
            // Check if all terms are required
            if (is_array($signs)) {
                foreach ($signs as $sign ) {
                    if ($sign !== true) {
                        $this->_signs = $signs;
                        break;
                    }
                }
            }
        }
    }


    /**
     * Add a $term (Zend_Search_Lucene_Index_Term) to this query.
     *
     * The sign is specified as:
     *     TRUE  - term is required
     *     FALSE - term is prohibited
     *     NULL  - term is neither prohibited, nor required
     *
     * @param  Zend_Search_Lucene_Index_Term $term
     * @param  boolean|null $sign
     * @return void
     */
    public function addTerm(Zend_Search_Lucene_Index_Term $term, $sign = null) {
        if ($sign !== true || $this->_signs !== null) {       // Skip, if all terms are required
            if ($this->_signs === null) {                     // Check, If all previous terms are required
                $this->_signs = array();
                foreach ($this->_terms as $prevTerm) {
                    $this->_signs[] = true;
                }
            }
            $this->_signs[] = $sign;
        }

        $this->_terms[] = $term;
    }


    /**
     * Re-write query into primitive queries in the context of specified index
     *
     * @param Zend_Search_Lucene_Interface $index
     * @return Zend_Search_Lucene_Search_Query
     */
    public function rewrite(Zend_Search_Lucene_Interface $index)
    {
        if (count($this->_terms) == 0) {
            return new Zend_Search_Lucene_Search_Query_Empty();
        }

        // Check, that all fields are qualified
        $allQualified = true;
        foreach ($this->_terms as $term) {
            if ($term->field === null) {
                $allQualified = false;
                break;
            }
        }

        if ($allQualified) {
            return $this;
        } else {
            /** transform multiterm query to boolean and apply rewrite() method to subqueries. */
            $query = new Zend_Search_Lucene_Search_Query_Boolean();
            $query->setBoost($this->getBoost());

            foreach ($this->_terms as $termId => $term) {
                $subquery = new Zend_Search_Lucene_Search_Query_Term($term);

                $query->addSubquery($subquery->rewrite($index),
                                    ($this->_signs === null)?  true : $this->_signs[$termId]);
            }

            return $query;
        }
    }

    /**
     * Optimize query in the context of specified index
     *
     * @param Zend_Search_Lucene_Interface $index
     * @return Zend_Search_Lucene_Search_Query
     */
    public function optimize(Zend_Search_Lucene_Interface $index)
    {
        $terms = $this->_terms;
        $signs = $this->_signs;

        foreach ($terms as $id => $term) {
            if (!$index->hasTerm($term)) {
                if ($signs === null  ||  $signs[$id] === true) {
                    // Term is required
                    return new Zend_Search_Lucene_Search_Query_Empty();
                } else {
                    // Term is optional or prohibited
                    // Remove it from terms and signs list
                    unset($terms[$id]);
                    unset($signs[$id]);
                }
            }
        }

        // Check if all presented terms are prohibited
        $allProhibited = true;
        if ($signs === null) {
            $allProhibited = false;
        } else {
            foreach ($signs as $sign) {
                if ($sign !== false) {
                    $allProhibited = false;
                    break;
                }
            }
        }
        if ($allProhibited) {
            return new Zend_Search_Lucene_Search_Query_Empty();
        }

        /**
         * @todo make an optimization for repeated terms
         * (they may have different signs)
         */

        if (count($terms) == 1) {
            // It's already checked, that it's not a prohibited term

            // It's one term query with one required or optional element
            $optimizedQuery = new Zend_Search_Lucene_Search_Query_Term(reset($terms));
            $optimizedQuery->setBoost($this->getBoost());

            return $optimizedQuery;
        }

        if (count($terms) == 0) {
            return new Zend_Search_Lucene_Search_Query_Empty();
        }

        $optimizedQuery = new Zend_Search_Lucene_Search_Query_MultiTerm($terms, $signs);
        $optimizedQuery->setBoost($this->getBoost());
        return $optimizedQuery;
    }


    /**
     * Returns query term
     *
     * @return array
     */
    public function getTerms()
    {
        return $this->_terms;
    }


    /**
     * Return terms signs
     *
     * @return array
     */
    public function getSigns()
    {
        return $this->_signs;
    }


    /**
     * Set weight for specified term
     *
     * @param integer $num
     * @param Zend_Search_Lucene_Search_Weight_Term $weight
     */
    public function setWeight($num, $weight)
    {
        $this->_weights[$num] = $weight;
    }


    /**
     * Constructs an appropriate Weight implementation for this query.
     *
     * @param Zend_Search_Lucene_Interface $reader
     * @return Zend_Search_Lucene_Search_Weight
     */
    public function createWeight(Zend_Search_Lucene_Interface $reader)
    {
        $this->_weight = new Zend_Search_Lucene_Search_Weight_MultiTerm($this, $reader);
        return $this->_weight;
    }


    /**
     * Calculate result vector for Conjunction query
     * (like '+something +another')
     *
     * @param Zend_Search_Lucene_Interface $reader
     */
    private function _calculateConjunctionResult(Zend_Search_Lucene_Interface $reader)
    {
        $this->_resVector = null;

        if (count($this->_terms) == 0) {
            $this->_resVector = array();
        }

        $resVectors      = array();
        $resVectorsSizes = array();
        $resVectorsIds   = array(); // is used to prevent arrays comparison
        foreach ($this->_terms as $termId => $term) {
            $resVectors[]      = array_flip($reader->termDocs($term));
            $resVectorsSizes[] = count(end($resVectors));
            $resVectorsIds[]   = $termId;

            $this->_termsFreqs[$termId] = $reader->termFreqs($term);
        }
        // sort resvectors in order of subquery cardinality increasing
        array_multisort($resVectorsSizes, SORT_ASC, SORT_NUMERIC,
                        $resVectorsIds,   SORT_ASC, SORT_NUMERIC,
                        $resVectors);

        foreach ($resVectors as $nextResVector) {
            if($this->_resVector === null) {
                $this->_resVector = $nextResVector;