| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266 |
- <?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 license@zend.com so we can send you a copy immediately.
- *
- * @category Zend
- * @package Zend_Search_Lucene
- * @subpackage UnitTests
- * @copyright Copyright (c) 2005-2015 Zend Technologies USA Inc. (http://www.zend.com)
- * @license http://framework.zend.com/license/new-bsd New BSD License
- * @version $Id$
- */
- /**
- * Zend_Search_Lucene
- */
- require_once 'Zend/Search/Lucene.php';
- /**
- * @category Zend
- * @package Zend_Search_Lucene
- * @subpackage UnitTests
- * @copyright Copyright (c) 2005-2015 Zend Technologies USA Inc. (http://www.zend.com)
- * @license http://framework.zend.com/license/new-bsd New BSD License
- * @group Zend_Search_Lucene
- */
- class Zend_Search_Lucene_SearchHighlightTest extends PHPUnit_Framework_TestCase
- {
- /**
- * Wildcard pattern minimum preffix
- *
- * @var integer
- */
- protected $_wildcardMinPrefix;
- /**
- * Fuzzy search default preffix length
- *
- * @var integer
- */
- protected $_defaultPrefixLength;
- public function setUp()
- {
- require_once 'Zend/Search/Lucene/Search/Query/Wildcard.php';
- $this->_wildcardMinPrefix = Zend_Search_Lucene_Search_Query_Wildcard::getMinPrefixLength();
- Zend_Search_Lucene_Search_Query_Wildcard::setMinPrefixLength(0);
- require_once 'Zend/Search/Lucene/Search/Query/Fuzzy.php';
- $this->_defaultPrefixLength = Zend_Search_Lucene_Search_Query_Fuzzy::getDefaultPrefixLength();
- Zend_Search_Lucene_Search_Query_Fuzzy::setDefaultPrefixLength(0);
- }
- public function tearDown()
- {
- Zend_Search_Lucene_Search_Query_Wildcard::setMinPrefixLength($this->_wildcardMinPrefix);
- Zend_Search_Lucene_Search_Query_Fuzzy::setDefaultPrefixLength($this->_defaultPrefixLength);
- }
- public function testHtmlFragmentHighlightMatches()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('title:"The Right Way" AND text:go');
- $highlightedHtmlFragment = $query->htmlFragmentHighlightMatches('Text highlighting using Zend_Search_Lucene is the right way to go!');
- $this->assertEquals($highlightedHtmlFragment,
- 'Text highlighting using Zend_Search_Lucene is <b style="color:black;background-color:#66ffff">the</b> <b style="color:black;background-color:#66ffff">right</b> <b style="color:black;background-color:#66ffff">way</b> to <b style="color:black;background-color:#ff66ff">go</b>!');
- }
- // public function testHtmlFragmentHighlightMatchesCyrillic()
- // {
- // $query = Zend_Search_Lucene_Search_QueryParser::parse('title:"некоторый текст" AND text:поехали');
- //
- // $highlightedHtmlFragment = $query->htmlFragmentHighlightMatches('Подсвечиваем некоторый текст с использованием Zend_Search_Lucene. Поехали!');
- //
- // $this->assertEquals($highlightedHtmlFragment,
- // 'Text highlighting using Zend_Search_Lucene is <b style="color:black;background-color:#66ffff">the</b> <b style="color:black;background-color:#66ffff">right</b> <b style="color:black;background-color:#66ffff">way</b> to <b style="color:black;background-color:#ff66ff">go</b>!');
- // }
- //
- // public function testHtmlFragmentHighlightMatchesCyrillicWindows()
- // {
- // $query = Zend_Search_Lucene_Search_QueryParser::parse('title:"Некоторый текст" AND text:поехали');
- //
- // $highlightedHtmlFragment =
- // $query->htmlFragmentHighlightMatches(iconv('UTF-8',
- // 'Windows-1251',
- // 'Подсвечиваем некоторый текст с использованием Zend_Search_Lucene. Поехали!'),
- // 'Windows-1251');
- //
- // $this->assertEquals($highlightedHtmlFragment,
- // 'Text highlighting using Zend_Search_Lucene is <b style="color:black;background-color:#66ffff">the</b> <b style="color:black;background-color:#66ffff">right</b> <b style="color:black;background-color:#66ffff">way</b> to <b style="color:black;background-color:#ff66ff">go</b>!');
- // }
- public function testHighlightPhrasePlusTerm()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('title:"The Right Way" AND text:go');
- $html = '<HTML>'
- . '<HEAD><TITLE>Page title</TITLE></HEAD>'
- . '<BODY>'
- . 'Text highlighting using Zend_Search_Lucene is the right way to go!'
- . '</BODY>'
- . '</HTML>';
- $highlightedHTML = $query->highlightMatches($html);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">the</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">right</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">way</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#ff66ff">go</b>') !== false);
- }
- public function testHighlightMultitermWithProhibitedTerms()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('+text +highlighting -using -right +go');
- $html = '<HTML>'
- . '<HEAD><TITLE>Page title</TITLE></HEAD>'
- . '<BODY>'
- . 'Text highlighting using Zend_Search_Lucene is the right way to go!'
- . '</BODY>'
- . '</HTML>';
- $highlightedHTML = $query->highlightMatches($html);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">Text</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#ff66ff">highlighting</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, 'using Zend_Search_Lucene is the right way to') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#ffff66">go</b>') !== false);
- }
- public function testHighlightWildcard1()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('te?t');
- $html = '<HTML>'
- . '<HEAD><TITLE>Page title</TITLE></HEAD>'
- . '<BODY>'
- . 'Test of text highlighting using wildcard query with question mark. Testing...'
- . '</BODY>'
- . '</HTML>';
- $highlightedHTML = $query->highlightMatches($html);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">Test</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">text</b>') !== false);
- // Check that 'Testing' word is not highlighted
- $this->assertTrue(strpos($highlightedHTML, 'mark. Testing...') !== false);
- }
- public function testHighlightWildcard2()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('te?t*');
- $html = '<HTML>'
- . '<HEAD><TITLE>Page title</TITLE></HEAD>'
- . '<BODY>'
- . 'Test of text highlighting using wildcard query with question mark. Testing...'
- . '</BODY>'
- . '</HTML>';
- $highlightedHTML = $query->highlightMatches($html);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">Test</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">text</b>') !== false);
- // Check that 'Testing' word is also highlighted
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">Testing</b>') !== false);
- }
- public function testHighlightFuzzy1()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('test~');
- $html = '<HTML>'
- . '<HEAD><TITLE>Page title</TITLE></HEAD>'
- . '<BODY>'
- . 'Test of text fuzzy search terms highlighting. '
- . 'Words: test, text, latest, left, list, next, ...'
- . '</BODY>'
- . '</HTML>';
- $highlightedHTML = $query->highlightMatches($html);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">Test</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">test</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">text</b>') !== false);
- // Check that other words are not highlighted
- $this->assertTrue(strpos($highlightedHTML, 'latest, left, list, next, ...') !== false);
- }
- public function testHighlightFuzzy2()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('test~0.4');
- $html = '<HTML>'
- . '<HEAD><TITLE>Page title</TITLE></HEAD>'
- . '<BODY>'
- . 'Test of text fuzzy search terms highlighting. '
- . 'Words: test, text, latest, left, list, next, ...'
- . '</BODY>'
- . '</HTML>';
- $highlightedHTML = $query->highlightMatches($html);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">Test</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">test</b>') !== false);
- // Check that other words are also highlighted
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">text</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">latest</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">left</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">list</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">next</b>') !== false);
- }
- public function testHighlightRangeInclusive()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('[business TO by]');
- $html = '<HTML>'
- . '<HEAD><TITLE>Page title</TITLE></HEAD>'
- . '<BODY>'
- . 'Test of text using range query. '
- . 'It has to match "business", "by", "buss" and "but" words, but has to skip "bus"'
- . '</BODY>'
- . '</HTML>';
- $highlightedHTML = $query->highlightMatches($html);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">business</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">by</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">buss</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">but</b>') !== false);
- // Check that "bus" word is skipped
- $this->assertTrue(strpos($highlightedHTML, 'has to skip "bus"') !== false);
- }
- public function testHighlightRangeNonInclusive()
- {
- $query = Zend_Search_Lucene_Search_QueryParser::parse('{business TO by}');
- $html = '<HTML>'
- . '<HEAD><TITLE>Page title</TITLE></HEAD>'
- . '<BODY>'
- . 'Test of text using range query. '
- . 'It has to match "buss" and "but" words, but has to skip "business", "by" and "bus"'
- . '</BODY>'
- . '</HTML>';
- $highlightedHTML = $query->highlightMatches($html);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">buss</b>') !== false);
- $this->assertTrue(strpos($highlightedHTML, '<b style="color:black;background-color:#66ffff">but</b>') !== false);
- // Check that "bus" word is skipped
- $this->assertTrue(strpos($highlightedHTML, 'has to skip "business", "by" and "bus"') !== false);
- }
- }
|