fulltext_search::indexedWordQuery() - Code Metrics - albertlast/SMF2.1 - Measure and Improve Code Quality continuously with Scrutinizer

fulltext_search::indexedWordQuery() F
last analyzed 2025-07-20 06:27 UTC

↳ Parent: fulltext_search

Complexity

Conditions	35
Paths	> 20000

Size

Total Lines	141
Code Lines	89

Duplication

Lines	0
Ratio	0 %

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
cc	35
eloc	89
c	1
b	0
f	0
nc	32768
nop	2
dl	0
loc	141
rs	0

How to fix Long Method Complexity

<?php

/**
 * Simple Machines Forum (SMF)
 *
 * @package SMF
 * @author Simple Machines https://www.simplemachines.org
 * @copyright 2023 Simple Machines and individual contributors
 * @license https://www.simplemachines.org/about/smf/license.php BSD
 *
 * @version 2.1.4
 */

if (!defined('SMF'))
	die('No direct access...');

/**
 * Class fulltext_search
 * Used for fulltext index searching
 */
class fulltext_search extends search_api
{
	/**
	 * @var array Which words are banned
	 */
	protected $bannedWords = array();

	/**
	 * @var int The minimum word length
	 */
	protected $min_word_length = 4;

	/**
	 * @var array Which databases support this method?
	 */
	protected $supported_databases = array('mysql', 'postgresql');

	/**
	 * The constructor function
	 */
	public function __construct()
	{
		global $modSettings, $db_type;

		// Is this database supported?
		if (!in_array($db_type, $this->supported_databases))
		{
			$this->is_supported = false;
			return;
		}

		$this->bannedWords = empty($modSettings['search_banned_words']) ? array() : explode(',', $modSettings['search_banned_words']);
		$this->min_word_length = $this->_getMinWordLength();
	}

	/**
	 * {@inheritDoc}
	 */
	public function supportsMethod($methodName, $query_params = null)
	{
		$return = false;
		switch ($methodName)
		{
			case 'searchSort':
			case 'prepareIndexes':
			case 'indexedWordQuery':
				$return = true;
				break;

			// All other methods, too bad dunno you.
			default:
				$return = false;
				break;
		}

		// Maybe parent got support
		if (!$return)
			$return = parent::supportsMethod($methodName, $query_params);

		return $return;
	}

	/**
	 * fulltext_search::_getMinWordLength()
	 *
	 * What is the minimum word length full text supports?
	 *
	 * @return int The minimum word length
	 */
	protected function _getMinWordLength()
	{
		global $smcFunc, $db_type;

		if ($db_type == 'postgresql')
			return 0;
		// Try to determine the minimum number of letters for a fulltext search.
		$request = $smcFunc['db_search_query']('max_fulltext_length', '
			SHOW VARIABLES
			LIKE {string:fulltext_minimum_word_length}',
			array(
				'fulltext_minimum_word_length' => 'ft_min_word_len',
			)
		);
		if ($request !== false && $smcFunc['db_num_rows']($request) == 1)
		{
			list (, $min_word_length) = $smcFunc['db_fetch_row']($request);
			$smcFunc['db_free_result']($request);
		}
		// 4 is the MySQL default...
		else
			$min_word_length = 4;

		return $min_word_length;
	}

	/**
	 * {@inheritDoc}
	 */
	public function searchSort($a, $b)
	{
		global $excludedWords, $smcFunc;

		$x = $smcFunc['strlen']($a) - (in_array($a, $excludedWords) ? 1000 : 0);
		$y = $smcFunc['strlen']($b) - (in_array($b, $excludedWords) ? 1000 : 0);

		return $x < $y ? 1 : ($x > $y ? -1 : 0);
	}

	/**
	 * {@inheritDoc}
	 */
	public function prepareIndexes($word, array &$wordsSearch, array &$wordsExclude, $isExcluded)
	{
		global $modSettings, $smcFunc;

		$subwords = text2words($word, null, false);

		if (empty($modSettings['search_force_index']))
		{
			// A boolean capable search engine and not forced to only use an index, we may use a non indexed search
			// this is harder on the server so we are restrictive here
			if (count($subwords) > 1 && preg_match('~[.:@$]~', $word))
			{
				// using special characters that a full index would ignore and the remaining words are short which would also be ignored
				if (($smcFunc['strlen'](current($subwords)) < $this->min_word_length) && ($smcFunc['strlen'](next($subwords)) < $this->min_word_length))
				{
					$wordsSearch['words'][] = trim($word, "/*- ");
					$wordsSearch['complex_words'][] = count($subwords) === 1 ? $word : '"' . $word . '"';
				}
			}
			elseif ($smcFunc['strlen'](trim($word, "/*- ")) < $this->min_word_length)
			{
				// short words have feelings too
				$wordsSearch['words'][] = trim($word, "/*- ");
				$wordsSearch['complex_words'][] = count($subwords) === 1 ? $word : '"' . $word . '"';
			}
		}

		$fulltextWord = count($subwords) === 1 ? $word : '"' . $word . '"';
		$wordsSearch['indexed_words'][] = $fulltextWord;
		if ($isExcluded)
			$wordsExclude[] = $fulltextWord;
	}

	/**
	 * {@inheritDoc}
	 */
	public function indexedWordQuery(array $words, array $search_data)
	{
		global $modSettings, $smcFunc;

		// Specify the function to search with. Regex is for word boundaries.
		$is_search_regex = !empty($modSettings['search_match_words']) && !$search_data['no_regexp'];
		$query_match_type = $is_search_regex ? 'RLIKE' : 'LIKE';
		$word_boundary_wrapper = function(string $str) use ($smcFunc): string
		{
			return sprintf($smcFunc['db_supports_pcre'] ? '\\b%s\\b' : '[[:<:]]%s[[:>:]]', $str);
		};
		$escape_sql_regex = function(string $str): string
		{
			return addcslashes(preg_replace('/[\[\]$.+*?&^|{}()]/', '[$0]', $str), '\\\'');
		};

		$query_select = array(
			'id_msg' => 'm.id_msg',
		);
		$query_where = array();
		$query_params = $search_data['params'];

		if ($smcFunc['db_title'] === POSTGRE_TITLE)
			$modSettings['search_simple_fulltext'] = true;

		if ($query_params['id_search'])
			$query_select['id_search'] = '{int:id_search}';

		$count = 0;
		if (empty($modSettings['search_simple_fulltext']))
			foreach ($words['words'] as $regularWord)
			{
				if (in_array($regularWord, $query_params['excluded_words']))
					$query_where[] = 'm.body NOT ' . $query_match_type . ' {string:complex_body_' . $count . '}';
				else
					$query_where[] = 'm.body ' . $query_match_type . ' {string:complex_body_' . $count . '}';

				if ($is_search_regex)
					$query_params['complex_body_' . $count++] = $word_boundary_wrapper($escape_sql_regex($regularWord));
				else
					$query_params['complex_body_' . $count++] = '%' . $smcFunc['db_escape_wildcard_string']($regularWord) . '%';
			}

		if ($query_params['user_query'])
			$query_where[] = '{raw:user_query}';
		if ($query_params['board_query'])
			$query_where[] = 'm.id_board {raw:board_query}';

		if ($query_params['topic'])
			$query_where[] = 'm.id_topic = {int:topic}';
		if ($query_params['min_msg_id'])
			$query_where[] = 'm.id_msg >= {int:min_msg_id}';
		if ($query_params['max_msg_id'])
			$query_where[] = 'm.id_msg <= {int:max_msg_id}';

		$count = 0;
		if (!empty($query_params['excluded_phrases']) && empty($modSettings['search_force_index']))
			foreach ($query_params['excluded_phrases'] as $phrase)
			{
				$query_where[] = 'subject NOT ' . $query_match_type . ' {string:exclude_subject_phrase_' . $count . '}';

				if ($is_search_regex)
					$query_params['exclude_subject_phrase_' . $count++] = $word_boundary_wrapper($escape_sql_regex($phrase));
				else
					$query_params['exclude_subject_phrase_' . $count++] = '%' . $smcFunc['db_escape_wildcard_string']($phrase) . '%';
			}
		$count = 0;
		if (!empty($query_params['excluded_subject_words']) && empty($modSettings['search_force_index']))
			foreach ($query_params['excluded_subject_words'] as $excludedWord)
			{
				$query_where[] = 'subject NOT ' . $query_match_type . ' {string:exclude_subject_words_' . $count . '}';

				if ($is_search_regex)
					$query_params['exclude_subject_words_' . $count++] = $word_boundary_wrapper($escape_sql_regex($excludedWord));
				else
					$query_params['exclude_subject_words_' . $count++] = '%' . $smcFunc['db_escape_wildcard_string']($excludedWord) . '%';
			}

		if (!empty($modSettings['search_simple_fulltext']))
		{
			if ($smcFunc['db_title'] === POSTGRE_TITLE)
			{
				$language_ftx = $smcFunc['db_search_language']();

				$query_where[] = 'to_tsvector({string:language_ftx},body) @@ plainto_tsquery({string:language_ftx},{string:body_match})';
				$query_params['language_ftx'] = $language_ftx;
			}
			else
				$query_where[] = 'MATCH (body) AGAINST ({string:body_match})';
			$query_params['body_match'] = implode(' ', array_diff($words['indexed_words'], $query_params['excluded_index_words']));
		}
		else
		{
			$query_params['boolean_match'] = '';

			// remove any indexed words that are used in the complex body search terms
			$words['indexed_words'] = array_diff($words['indexed_words'], $words['complex_words']);

			if ($smcFunc['db_title'] === POSTGRE_TITLE)
			{
				$row = 0;
				foreach ($words['indexed_words'] as $fulltextWord)
				{
					$query_params['boolean_match'] .= ($row <> 0 ? '&' : '');
					$query_params['boolean_match'] .= (in_array($fulltextWord, $query_params['excluded_index_words']) ? '!' : '') . $fulltextWord . ' ';
					$row++;
				}
			}
			else
				foreach ($words['indexed_words'] as $fulltextWord)
					$query_params['boolean_match'] .= (in_array($fulltextWord, $query_params['excluded_index_words']) ? '-' : '+') . $fulltextWord . ' ';

			$query_params['boolean_match'] = substr($query_params['boolean_match'], 0, -1);

			// if we have bool terms to search, add them in
			if ($query_params['boolean_match'])
			{
				if ($smcFunc['db_title'] === POSTGRE_TITLE)
				{
					$language_ftx = $smcFunc['db_search_language']();

					$query_where[] = 'to_tsvector({string:language_ftx},body) @@ plainto_tsquery({string:language_ftx},{string:boolean_match})';
					$query_params['language_ftx'] = $language_ftx;
				}
				else
					$query_where[] = 'MATCH (body) AGAINST ({string:boolean_match} IN BOOLEAN MODE)';
			}
		}

		$ignoreRequest = $smcFunc['db_search_query']('insert_into_log_messages_fulltext', ($smcFunc['db_support_ignore'] ? ('
			INSERT IGNORE INTO {db_prefix}' . $search_data['insert_into'] . '
				(' . implode(', ', array_keys($query_select)) . ')') : '') . '
			SELECT ' . implode(', ', $query_select) . '
			FROM {db_prefix}messages AS m
			WHERE ' . implode('
				AND ', $query_where) . (empty($search_data['max_results']) ? '' : '
			LIMIT ' . ($search_data['max_results'] - $search_data['indexed_results'])),
			$query_params
		);

		return $ignoreRequest;
	}
}

?>

1			<?php
2
3			/**
4			* Simple Machines Forum (SMF)
5			*
6			* @package SMF
7			* @author Simple Machines https://www.simplemachines.org
8			* @copyright 2023 Simple Machines and individual contributors
9			* @license https://www.simplemachines.org/about/smf/license.php BSD
10			*
11			* @version 2.1.4
12			*/
13
14			if (!defined('SMF'))
15			die('No direct access...');
16
17			/**
18			* Class fulltext_search
19			* Used for fulltext index searching
20			*/
21			class fulltext_search extends search_api
22			{
23			/**
24			* @var array Which words are banned
25			*/
26			protected $bannedWords = array();
27
28			/**
29			* @var int The minimum word length
30			*/
31			protected $min_word_length = 4;
32
33			/**
34			* @var array Which databases support this method?
35			*/
36			protected $supported_databases = array('mysql', 'postgresql');
37
38			/**
39			* The constructor function
40			*/
41			public function __construct()
42			{
43			global $modSettings, $db_type;
44
45			// Is this database supported?
46			if (!in_array($db_type, $this->supported_databases))
47			{
48			$this->is_supported = false;
49			return;
50			}
51
52			$this->bannedWords = empty($modSettings['search_banned_words']) ? array() : explode(',', $modSettings['search_banned_words']);
53			$this->min_word_length = $this->_getMinWordLength();
54			}
55
56			/**
57			* {@inheritDoc}
58			*/
59			public function supportsMethod($methodName, $query_params = null)
60			{
61			$return = false;
62			switch ($methodName)
63			{
64			case 'searchSort':
65			case 'prepareIndexes':
66			case 'indexedWordQuery':
67			$return = true;
68			break;
69
70			// All other methods, too bad dunno you.
71			default:
72			$return = false;
73			break;
74			}
75
76			// Maybe parent got support
77			if (!$return)
78			$return = parent::supportsMethod($methodName, $query_params);
79
80			return $return;
81			}
82
83			/**
84			* fulltext_search::_getMinWordLength()
85			*
86			* What is the minimum word length full text supports?
87			*
88			* @return int The minimum word length
89			*/
90			protected function _getMinWordLength()
91			{
92			global $smcFunc, $db_type;
93
94			if ($db_type == 'postgresql')
95			return 0;
96			// Try to determine the minimum number of letters for a fulltext search.
97			$request = $smcFunc['db_search_query']('max_fulltext_length', '
98			SHOW VARIABLES
99			LIKE {string:fulltext_minimum_word_length}',
100			array(
101			'fulltext_minimum_word_length' => 'ft_min_word_len',
102			)
103			);
104			if ($request !== false && $smcFunc['db_num_rows']($request) == 1)
105			{
106			list (, $min_word_length) = $smcFunc['db_fetch_row']($request);
107			$smcFunc['db_free_result']($request);
108			}
109			// 4 is the MySQL default...
110			else
111			$min_word_length = 4;
112
113			return $min_word_length;
114			}
115
116			/**
117			* {@inheritDoc}
118			*/
119			public function searchSort($a, $b)
120			{
121			global $excludedWords, $smcFunc;
122
123			$x = $smcFunc['strlen']($a) - (in_array($a, $excludedWords) ? 1000 : 0);
124			$y = $smcFunc['strlen']($b) - (in_array($b, $excludedWords) ? 1000 : 0);
125
126			return $x < $y ? 1 : ($x > $y ? -1 : 0);
127			}
128
129			/**
130			* {@inheritDoc}
131			*/
132			public function prepareIndexes($word, array &$wordsSearch, array &$wordsExclude, $isExcluded)
133			{
134			global $modSettings, $smcFunc;
135
136			$subwords = text2words($word, null, false);
137
138			if (empty($modSettings['search_force_index']))
139			{
140			// A boolean capable search engine and not forced to only use an index, we may use a non indexed search
141			// this is harder on the server so we are restrictive here
142			if (count($subwords) > 1 && preg_match('~[.:@$]~', $word))
143			{
144			// using special characters that a full index would ignore and the remaining words are short which would also be ignored
145			if (($smcFunc['strlen'](current($subwords)) < $this->min_word_length) && ($smcFunc['strlen'](next($subwords)) < $this->min_word_length))
146			{
147			$wordsSearch['words'][] = trim($word, "/*- ");
148			$wordsSearch['complex_words'][] = count($subwords) === 1 ? $word : '"' . $word . '"';
149			}
150			}
151			elseif ($smcFunc['strlen'](trim($word, "/*- ")) < $this->min_word_length)
152			{
153			// short words have feelings too
154			$wordsSearch['words'][] = trim($word, "/*- ");
155			$wordsSearch['complex_words'][] = count($subwords) === 1 ? $word : '"' . $word . '"';
156			}
157			}
158
159			$fulltextWord = count($subwords) === 1 ? $word : '"' . $word . '"';
160			$wordsSearch['indexed_words'][] = $fulltextWord;
161			if ($isExcluded)
162			$wordsExclude[] = $fulltextWord;
163			}
164
165			/**
166			* {@inheritDoc}
167			*/
168			public function indexedWordQuery(array $words, array $search_data)
169			{
170			global $modSettings, $smcFunc;
171
172			// Specify the function to search with. Regex is for word boundaries.
173			$is_search_regex = !empty($modSettings['search_match_words']) && !$search_data['no_regexp'];
174			$query_match_type = $is_search_regex ? 'RLIKE' : 'LIKE';
175			$word_boundary_wrapper = function(string $str) use ($smcFunc): string
176			{
177			return sprintf($smcFunc['db_supports_pcre'] ? '\\b%s\\b' : '[[:<:]]%s[[:>:]]', $str);
178			};
179			$escape_sql_regex = function(string $str): string
180			{
181			return addcslashes(preg_replace('/[\[\]$.+*?&^\|{}()]/', '[$0]', $str), '\\\'');
182			};
183
184			$query_select = array(
185			'id_msg' => 'm.id_msg',
186			);
187			$query_where = array();
188			$query_params = $search_data['params'];
189
190			if ($smcFunc['db_title'] === POSTGRE_TITLE)
191			$modSettings['search_simple_fulltext'] = true;
192
193			if ($query_params['id_search'])
194			$query_select['id_search'] = '{int:id_search}';
195
196			$count = 0;
197			if (empty($modSettings['search_simple_fulltext']))
198			foreach ($words['words'] as $regularWord)
199			{
200			if (in_array($regularWord, $query_params['excluded_words']))
201			$query_where[] = 'm.body NOT ' . $query_match_type . ' {string:complex_body_' . $count . '}';
202			else
203			$query_where[] = 'm.body ' . $query_match_type . ' {string:complex_body_' . $count . '}';
204
205			if ($is_search_regex)
206			$query_params['complex_body_' . $count++] = $word_boundary_wrapper($escape_sql_regex($regularWord));
207			else
208			$query_params['complex_body_' . $count++] = '%' . $smcFunc['db_escape_wildcard_string']($regularWord) . '%';
209			}
210
211			if ($query_params['user_query'])
212			$query_where[] = '{raw:user_query}';
213			if ($query_params['board_query'])
214			$query_where[] = 'm.id_board {raw:board_query}';
215
216			if ($query_params['topic'])
217			$query_where[] = 'm.id_topic = {int:topic}';
218			if ($query_params['min_msg_id'])
219			$query_where[] = 'm.id_msg >= {int:min_msg_id}';
220			if ($query_params['max_msg_id'])
221			$query_where[] = 'm.id_msg <= {int:max_msg_id}';
222
223			$count = 0;
224			if (!empty($query_params['excluded_phrases']) && empty($modSettings['search_force_index']))
225			foreach ($query_params['excluded_phrases'] as $phrase)
226			{
227			$query_where[] = 'subject NOT ' . $query_match_type . ' {string:exclude_subject_phrase_' . $count . '}';
228
229			if ($is_search_regex)
230			$query_params['exclude_subject_phrase_' . $count++] = $word_boundary_wrapper($escape_sql_regex($phrase));
231			else
232			$query_params['exclude_subject_phrase_' . $count++] = '%' . $smcFunc['db_escape_wildcard_string']($phrase) . '%';
233			}
234			$count = 0;
235			if (!empty($query_params['excluded_subject_words']) && empty($modSettings['search_force_index']))
236			foreach ($query_params['excluded_subject_words'] as $excludedWord)
237			{
238			$query_where[] = 'subject NOT ' . $query_match_type . ' {string:exclude_subject_words_' . $count . '}';
239
240			if ($is_search_regex)
241			$query_params['exclude_subject_words_' . $count++] = $word_boundary_wrapper($escape_sql_regex($excludedWord));
242			else
243			$query_params['exclude_subject_words_' . $count++] = '%' . $smcFunc['db_escape_wildcard_string']($excludedWord) . '%';
244			}
245
246			if (!empty($modSettings['search_simple_fulltext']))
247			{
248			if ($smcFunc['db_title'] === POSTGRE_TITLE)
249			{
250			$language_ftx = $smcFunc['db_search_language']();
251
252			$query_where[] = 'to_tsvector({string:language_ftx},body) @@ plainto_tsquery({string:language_ftx},{string:body_match})';
253			$query_params['language_ftx'] = $language_ftx;
254			}
255			else
256			$query_where[] = 'MATCH (body) AGAINST ({string:body_match})';
257			$query_params['body_match'] = implode(' ', array_diff($words['indexed_words'], $query_params['excluded_index_words']));
258			}
259			else
260			{
261			$query_params['boolean_match'] = '';
262
263			// remove any indexed words that are used in the complex body search terms
264			$words['indexed_words'] = array_diff($words['indexed_words'], $words['complex_words']);
265
266			if ($smcFunc['db_title'] === POSTGRE_TITLE)
267			{
268			$row = 0;
269			foreach ($words['indexed_words'] as $fulltextWord)
270			{
271			$query_params['boolean_match'] .= ($row <> 0 ? '&' : '');
272			$query_params['boolean_match'] .= (in_array($fulltextWord, $query_params['excluded_index_words']) ? '!' : '') . $fulltextWord . ' ';
273			$row++;
274			}
275			}
276			else
277			foreach ($words['indexed_words'] as $fulltextWord)
278			$query_params['boolean_match'] .= (in_array($fulltextWord, $query_params['excluded_index_words']) ? '-' : '+') . $fulltextWord . ' ';
279
280			$query_params['boolean_match'] = substr($query_params['boolean_match'], 0, -1);
281
282			// if we have bool terms to search, add them in
283			if ($query_params['boolean_match'])
284			{
285			if ($smcFunc['db_title'] === POSTGRE_TITLE)
286			{
287			$language_ftx = $smcFunc['db_search_language']();
288
289			$query_where[] = 'to_tsvector({string:language_ftx},body) @@ plainto_tsquery({string:language_ftx},{string:boolean_match})';
290			$query_params['language_ftx'] = $language_ftx;
291			}
292			else
293			$query_where[] = 'MATCH (body) AGAINST ({string:boolean_match} IN BOOLEAN MODE)';
294			}
295			}
296
297			$ignoreRequest = $smcFunc['db_search_query']('insert_into_log_messages_fulltext', ($smcFunc['db_support_ignore'] ? ('
298			INSERT IGNORE INTO {db_prefix}' . $search_data['insert_into'] . '
299			(' . implode(', ', array_keys($query_select)) . ')') : '') . '
300			SELECT ' . implode(', ', $query_select) . '
301			FROM {db_prefix}messages AS m
302			WHERE ' . implode('
303			AND ', $query_where) . (empty($search_data['max_results']) ? '' : '
304			LIMIT ' . ($search_data['max_results'] - $search_data['indexed_results'])),
305			$query_params
306			);
307
308			return $ignoreRequest;
309			}
310			}
311
312			?>

albertlast / SMF2.1

fulltext_search::indexedWordQuery() F last analyzed 2025-07-20 06:27 UTC

Complexity

Size

Duplication

Importance

How to fix Long Method Complexity

Long Method

Duplication Side-by-Side

Filter issues like

fulltext_search::indexedWordQuery() F
last analyzed 2025-07-20 06:27 UTC