boolean.php

来自「PHP 知识管理系统（基于树结构的知识管理系统）, 英文原版的PHP源码。」· PHP 代码 · 共 710 行 · 第 1/2 页

<?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) 2005-2007 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_Boolean */
require_once 'Zend/Search/Lucene/Search/Weight/Boolean.php';


/**
 * @category   Zend
 * @package    Zend_Search_Lucene
 * @subpackage Search
 * @copyright  Copyright (c) 2005-2007 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_Boolean extends Zend_Search_Lucene_Search_Query
{

    /**
     * Subqueries
     * Array of Zend_Search_Lucene_Query
     *
     * @var array
     */
    private $_subqueries = array();

    /**
     * Subqueries signs.
     * If true then subquery is required.
     * If false then subquery is prohibited.
     * If null then subquery is neither prohibited, nor required
     *
     * If array is null then all subqueries are required
     *
     * @var array
     */
    private $_signs = array();

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

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


    /**
     * Class constructor.  Create a new Boolean query object.
     *
     * if $signs array is omitted then all subqueries are required
     * it differs from addSubquery() behavior, but should never be used
     *
     * @param array $subqueries    Array of Zend_Search_Search_Query objects
     * @param array $signs    Array of signs.  Sign is boolean|null.
     * @return void
     */
    public function __construct($subqueries = null, $signs = null)
    {
        if (is_array($subqueries)) {
            $this->_subqueries = $subqueries;

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


    /**
     * Add a $subquery (Zend_Search_Lucene_Query) to this query.
     *
     * The sign is specified as:
     *     TRUE  - subquery is required
     *     FALSE - subquery is prohibited
     *     NULL  - subquery is neither prohibited, nor required
     *
     * @param  Zend_Search_Lucene_Search_Query $subquery
     * @param  boolean|null $sign
     * @return void
     */
    public function addSubquery(Zend_Search_Lucene_Search_Query $subquery, $sign=null) {
        if ($sign !== true || $this->_signs !== null) {       // Skip, if all subqueries are required
            if ($this->_signs === null) {                     // Check, If all previous subqueries are required
                foreach ($this->_subqueries as $prevSubquery) {
                    $this->_signs[] = true;
                }
            }
            $this->_signs[] = $sign;
        }

        $this->_subqueries[] = $subquery;
    }

    /**
     * Re-write queries into primitive queries
     *
     * @param Zend_Search_Lucene $index
     * @return Zend_Search_Lucene_Search_Query
     */
    public function rewrite(Zend_Search_Lucene $index)
    {
        $query = new Zend_Search_Lucene_Search_Query_Boolean();
        $query->setBoost($this->getBoost());

        foreach ($this->_subqueries as $subqueryId => $subquery) {
            $query->addSubquery($subquery->rewrite($index),
                                ($this->_signs === null)?  true : $this->_signs[$subqueryId]);
        }

        return $query;
    }

    /**
     * Optimize query in the context of specified index
     *
     * @param Zend_Search_Lucene $index
     * @return Zend_Search_Lucene_Search_Query
     */
    public function optimize(Zend_Search_Lucene $index)
    {
        $subqueries = array();
        $signs      = array();

        // Optimize all subqueries
        foreach ($this->_subqueries as $id => $subquery) {
            $subqueries[] = $subquery->optimize($index);
            $signs[]      = ($this->_signs === null)? true : $this->_signs[$id];
        }

        // Check for empty subqueries
        foreach ($subqueries as $id => $subquery) {
            if ($subquery instanceof Zend_Search_Lucene_Search_Query_Empty) {
                if ($signs[$id] === true) {
                    // Matching is required, but is actually empty
                    return new Zend_Search_Lucene_Search_Query_Empty();
                } else {
                    // Matching is optional or prohibited, but is empty
                    // Remove it from subqueries and signs list
                    unset($subqueries[$id]);
                    unset($signs[$id]);
                }
            }
        }


        // Check if all non-empty subqueries are prohibited
        $allProhibited = true;
        foreach ($signs as $sign) {
            if ($sign !== false) {
                $allProhibited = false;
                break;
            }
        }
        if ($allProhibited) {
            return new Zend_Search_Lucene_Search_Query_Empty();
        }


        // Check, if reduced subqueries list has only one entry
        if (count($subqueries) == 1) {
            // It's a query with only one required or optional clause
            // (it's already checked, that it's not a prohibited clause)

            if ($this->getBoost() == 1) {
                return reset($subqueries);
            }

            $optimizedQuery = clone reset($subqueries);
            $optimizedQuery->setBoost($optimizedQuery->getBoost()*$this->getBoost());

            return $optimizedQuery;
        }


        // Check, if reduced subqueries list is empty
        if (count($subqueries) == 0) {
            return new Zend_Search_Lucene_Search_Query_Empty();
        }


        // Prepare first candidate for optimized query
        $optimizedQuery = new Zend_Search_Lucene_Search_Query_Boolean($subqueries, $signs);
        $optimizedQuery->setBoost($this->getBoost());


        $terms        = array();
        $tsigns       = array();
        $boostFactors = array();

        // Try to decompose term and multi-term subqueries
        foreach ($subqueries as $id => $subquery) {
            if ($subquery instanceof Zend_Search_Lucene_Search_Query_Term) {
                $terms[]        = $subquery->getTerm();
                $tsigns[]       = $signs[$id];
                $boostFactors[] = $subquery->getBoost();

                // remove subquery from a subqueries list
                unset($subqueries[$id]);
                unset($signs[$id]);
           } else if ($subquery instanceof Zend_Search_Lucene_Search_Query_MultiTerm) {
                $subTerms = $subquery->getTerms();
                $subSigns = $subquery->getSigns();

                if ($signs[$id] === true) {
                    // It's a required multi-term subquery.
                    // Something like '... +(+term1 -term2 term3 ...) ...'

                    // Multi-term required subquery can be decomposed only if it contains
                    // required terms and doesn't contain prohibited terms:
                    // ... +(+term1 term2 ...) ... => ... +term1 term2 ...
                    //
                    // Check this
                    $hasRequired   = false;
                    $hasProhibited = false;
                    if ($subSigns === null) {
                        // All subterms are required
                        $hasRequired = true;
                    } else {
                        foreach ($subSigns as $sign) {
                            if ($sign === true) {
                                $hasRequired   = true;
                            } else if ($sign === false) {
                                $hasProhibited = true;
                                break;
                            }
                        }
                    }
                    // Continue if subquery has prohibited terms or doesn't have required terms
                    if ($hasProhibited  ||  !$hasRequired) {
                        continue;
                    }

                    foreach ($subTerms as $termId => $term) {
                        $terms[]        = $term;
                        $tsigns[]       = ($subSigns === null)? true : $subSigns[$termId];
                        $boostFactors[] = $subquery->getBoost();
                    }

                    // remove subquery from a subqueries list
                    unset($subqueries[$id]);
                    unset($signs[$id]);

                } else { // $signs[$id] === null  ||  $signs[$id] === false
                    // It's an optional or prohibited multi-term subquery.
                    // Something like '... (+term1 -term2 term3 ...) ...'
                    // or
                    // something like '... -(+term1 -term2 term3 ...) ...'

                    // Multi-term optional and required subqueries can be decomposed
                    // only if all terms are optional.
                    //
                    // Check if all terms are optional.
                    $onlyOptional = true;
                    if ($subSigns === null) {
                        // All subterms are required
                        $onlyOptional = false;
                    } else {
                        foreach ($subSigns as $sign) {
                            if ($sign !== null) {
                                $onlyOptional = false;
                                break;
                            }
                        }
                    }

                    // Continue if non-optional terms are presented in this multi-term subquery
                    if (!$onlyOptional) {
                        continue;
                    }

                    foreach ($subTerms as $termId => $term) {
                        $terms[]  = $term;
                        $tsigns[] = ($signs[$id] === null)? null  /* optional */ :
                                                            false /* prohibited */;
                        $boostFactors[] = $subquery->getBoost();
                    }

                    // remove subquery from a subqueries list
                    unset($subqueries[$id]);
                    unset($signs[$id]);
                }
            }
        }


        // Check, if there are no decomposed subqueries
        if (count($terms) == 0 ) {
            // return prepared candidate
            return $optimizedQuery;
        }


        // Check, if all subqueries have been decomposed and all terms has the same boost factor
        if (count($subqueries) == 0  &&  count(array_unique($boostFactors)) == 1) {
            $optimizedQuery = new Zend_Search_Lucene_Search_Query_MultiTerm($terms, $tsigns);
            $optimizedQuery->setBoost(reset($boostFactors)*$this->getBoost());

            return $optimizedQuery;
        }


        // This boolean query can't be transformed to Term/MultiTerm query and still contains
        // several subqueries

        // Separate prohibited terms
        $prohibitedTerms        = array();
        foreach ($terms as $id => $term) {
            if ($tsigns[$id] === false) {
                $prohibitedTerms[]        = $term;

                unset($terms[$id]);
                unset($tsigns[$id]);
                unset($boostFactors[$id]);
            }
        }

        if (count($terms) == 1) {