Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
Total | |
0.00% |
0 / 123 |
|
0.00% |
0 / 15 |
CRAP | |
0.00% |
0 / 1 |
SearchSqlite | |
0.00% |
0 / 123 |
|
0.00% |
0 / 15 |
1260 | |
0.00% |
0 / 1 |
fulltextSearchSupported | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
2 | |||
parseQuery | |
0.00% |
0 / 44 |
|
0.00% |
0 / 1 |
132 | |||
regexTerm | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
12 | |||
legalSearchChars | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
doSearchTextInDB | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
doSearchTitleInDB | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
searchInternal | |
0.00% |
0 / 13 |
|
0.00% |
0 / 1 |
12 | |||
queryNamespaces | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
12 | |||
limitResult | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
getQuery | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
2 | |||
getIndexField | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
6 | |||
queryMain | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
2 | |||
getCountQuery | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
2 | |||
update | |
0.00% |
0 / 11 |
|
0.00% |
0 / 1 |
6 | |||
updateTitle | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
6 |
1 | <?php |
2 | /** |
3 | * SQLite search backend, based upon SearchMysql |
4 | * |
5 | * This program is free software; you can redistribute it and/or modify |
6 | * it under the terms of the GNU General Public License as published by |
7 | * the Free Software Foundation; either version 2 of the License, or |
8 | * (at your option) any later version. |
9 | * |
10 | * This program is distributed in the hope that it will be useful, |
11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
13 | * GNU General Public License for more details. |
14 | * |
15 | * You should have received a copy of the GNU General Public License along |
16 | * with this program; if not, write to the Free Software Foundation, Inc., |
17 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
18 | * http://www.gnu.org/copyleft/gpl.html |
19 | * |
20 | * @file |
21 | * @ingroup Search |
22 | */ |
23 | |
24 | use MediaWiki\MediaWikiServices; |
25 | use Wikimedia\AtEase\AtEase; |
26 | use Wikimedia\Rdbms\IDatabase; |
27 | |
28 | /** |
29 | * Search engine hook for SQLite |
30 | * @ingroup Search |
31 | */ |
32 | class SearchSqlite extends SearchDatabase { |
33 | /** |
34 | * Whether fulltext search is supported by current schema |
35 | * @return bool |
36 | */ |
37 | private function fulltextSearchSupported() { |
38 | $dbr = $this->dbProvider->getReplicaDatabase(); |
39 | $sql = (string)$dbr->newSelectQueryBuilder() |
40 | ->select( 'sql' ) |
41 | ->from( $dbr->addIdentifierQuotes( 'sqlite_master' ) ) |
42 | ->where( [ 'tbl_name' => $dbr->tableName( 'searchindex', 'raw' ) ] ) |
43 | ->caller( __METHOD__ )->fetchField(); |
44 | |
45 | return ( stristr( $sql, 'fts' ) !== false ); |
46 | } |
47 | |
48 | /** |
49 | * Parse the user's query and transform it into an SQL fragment which will |
50 | * become part of a WHERE clause |
51 | * |
52 | * @param string $filteredText |
53 | * @param bool $fulltext |
54 | * @return string |
55 | */ |
56 | private function parseQuery( $filteredText, $fulltext ) { |
57 | $lc = $this->legalSearchChars( self::CHARS_NO_SYNTAX ); // Minus syntax chars (" and *) |
58 | $searchon = ''; |
59 | $this->searchTerms = []; |
60 | |
61 | $m = []; |
62 | if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/', |
63 | $filteredText, $m, PREG_SET_ORDER ) ) { |
64 | foreach ( $m as $bits ) { |
65 | AtEase::suppressWarnings(); |
66 | [ /* all */, $modifier, $term, $nonQuoted, $wildcard ] = $bits; |
67 | AtEase::restoreWarnings(); |
68 | |
69 | if ( $nonQuoted != '' ) { |
70 | $term = $nonQuoted; |
71 | $quote = ''; |
72 | } else { |
73 | $term = str_replace( '"', '', $term ); |
74 | $quote = '"'; |
75 | } |
76 | |
77 | if ( $searchon !== '' ) { |
78 | $searchon .= ' '; |
79 | } |
80 | |
81 | // Some languages such as Serbian store the input form in the search index, |
82 | // so we may need to search for matches in multiple writing system variants. |
83 | |
84 | $converter = MediaWikiServices::getInstance()->getLanguageConverterFactory() |
85 | ->getLanguageConverter(); |
86 | $convertedVariants = $converter->autoConvertToAllVariants( $term ); |
87 | if ( is_array( $convertedVariants ) ) { |
88 | $variants = array_unique( array_values( $convertedVariants ) ); |
89 | } else { |
90 | $variants = [ $term ]; |
91 | } |
92 | |
93 | // The low-level search index does some processing on input to work |
94 | // around problems with minimum lengths and encoding in MySQL's |
95 | // fulltext engine. |
96 | // For Chinese this also inserts spaces between adjacent Han characters. |
97 | $strippedVariants = array_map( |
98 | [ MediaWikiServices::getInstance()->getContentLanguage(), |
99 | 'normalizeForSearch' ], |
100 | $variants ); |
101 | |
102 | // Some languages such as Chinese force all variants to a canonical |
103 | // form when stripping to the low-level search index, so to be sure |
104 | // let's check our variants list for unique items after stripping. |
105 | $strippedVariants = array_unique( $strippedVariants ); |
106 | |
107 | $searchon .= $modifier; |
108 | if ( count( $strippedVariants ) > 1 ) { |
109 | $searchon .= '('; |
110 | } |
111 | foreach ( $strippedVariants as $stripped ) { |
112 | if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) { |
113 | // Hack for Chinese: we need to toss in quotes for |
114 | // multiple-character phrases since normalizeForSearch() |
115 | // added spaces between them to make word breaks. |
116 | $stripped = '"' . trim( $stripped ) . '"'; |
117 | } |
118 | $searchon .= "$quote$stripped$quote$wildcard "; |
119 | } |
120 | if ( count( $strippedVariants ) > 1 ) { |
121 | $searchon .= ')'; |
122 | } |
123 | |
124 | // Match individual terms or quoted phrase in result highlighting... |
125 | // Note that variants will be introduced in a later stage for highlighting! |
126 | $regexp = $this->regexTerm( $term, $wildcard ); |
127 | $this->searchTerms[] = $regexp; |
128 | } |
129 | |
130 | } else { |
131 | wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'" ); |
132 | } |
133 | |
134 | $dbr = $this->dbProvider->getReplicaDatabase(); |
135 | $searchon = $dbr->addQuotes( $searchon ); |
136 | $field = $this->getIndexField( $fulltext ); |
137 | |
138 | return " $field MATCH $searchon "; |
139 | } |
140 | |
141 | private function regexTerm( $string, $wildcard ) { |
142 | $regex = preg_quote( $string, '/' ); |
143 | if ( MediaWikiServices::getInstance()->getContentLanguage()->hasWordBreaks() ) { |
144 | if ( $wildcard ) { |
145 | // Don't cut off the final bit! |
146 | $regex = "\b$regex"; |
147 | } else { |
148 | $regex = "\b$regex\b"; |
149 | } |
150 | } else { |
151 | // For Chinese, words may legitimately abut other words in the text literal. |
152 | // Don't add \b boundary checks... note this could cause false positives |
153 | // for Latin chars. |
154 | } |
155 | return $regex; |
156 | } |
157 | |
158 | public function legalSearchChars( $type = self::CHARS_ALL ) { |
159 | $searchChars = parent::legalSearchChars( $type ); |
160 | if ( $type === self::CHARS_ALL ) { |
161 | // " for phrase, * for wildcard |
162 | $searchChars = "\"*" . $searchChars; |
163 | } |
164 | return $searchChars; |
165 | } |
166 | |
167 | /** |
168 | * Perform a full text search query and return a result set. |
169 | * |
170 | * @param string $term Raw search term |
171 | * @return SqlSearchResultSet|null |
172 | */ |
173 | protected function doSearchTextInDB( $term ) { |
174 | return $this->searchInternal( $term, true ); |
175 | } |
176 | |
177 | /** |
178 | * Perform a title-only search query and return a result set. |
179 | * |
180 | * @param string $term Raw search term |
181 | * @return SqlSearchResultSet|null |
182 | */ |
183 | protected function doSearchTitleInDB( $term ) { |
184 | return $this->searchInternal( $term, false ); |
185 | } |
186 | |
187 | protected function searchInternal( $term, $fulltext ) { |
188 | if ( !$this->fulltextSearchSupported() ) { |
189 | return null; |
190 | } |
191 | |
192 | $filteredTerm = |
193 | $this->filter( MediaWikiServices::getInstance()->getContentLanguage()->lc( $term ) ); |
194 | $dbr = $this->dbProvider->getReplicaDatabase(); |
195 | // The real type is still IDatabase, but IReplicaDatabase is used for safety. |
196 | '@phan-var IDatabase $dbr'; |
197 | // phpcs:ignore MediaWiki.Usage.DbrQueryUsage.DbrQueryFound |
198 | $resultSet = $dbr->query( $this->getQuery( $filteredTerm, $fulltext ), __METHOD__ ); |
199 | |
200 | $total = null; |
201 | // phpcs:ignore MediaWiki.Usage.DbrQueryUsage.DbrQueryFound |
202 | $totalResult = $dbr->query( $this->getCountQuery( $filteredTerm, $fulltext ), __METHOD__ ); |
203 | $row = $totalResult->fetchObject(); |
204 | if ( $row ) { |
205 | $total = intval( $row->c ); |
206 | } |
207 | $totalResult->free(); |
208 | |
209 | return new SqlSearchResultSet( $resultSet, $this->searchTerms, $total ); |
210 | } |
211 | |
212 | /** |
213 | * Return a partial WHERE clause to limit the search to the given namespaces |
214 | * @return string |
215 | */ |
216 | private function queryNamespaces() { |
217 | if ( $this->namespaces === null ) { |
218 | return ''; # search all |
219 | } |
220 | if ( $this->namespaces === [] ) { |
221 | $namespaces = NS_MAIN; |
222 | } else { |
223 | $dbr = $this->dbProvider->getReplicaDatabase(); |
224 | $namespaces = $dbr->makeList( $this->namespaces ); |
225 | } |
226 | return 'AND page_namespace IN (' . $namespaces . ')'; |
227 | } |
228 | |
229 | /** |
230 | * Returns a query with limit for number of results set. |
231 | * @param string $sql |
232 | * @return string |
233 | */ |
234 | private function limitResult( $sql ) { |
235 | return $this->dbProvider->getReplicaDatabase()->limitResult( $sql, $this->limit, $this->offset ); |
236 | } |
237 | |
238 | /** |
239 | * Construct the full SQL query to do the search. |
240 | * The guts shoulds be constructed in queryMain() |
241 | * @param string $filteredTerm |
242 | * @param bool $fulltext |
243 | * @return string |
244 | */ |
245 | private function getQuery( $filteredTerm, $fulltext ) { |
246 | return $this->limitResult( |
247 | $this->queryMain( $filteredTerm, $fulltext ) . ' ' . |
248 | $this->queryNamespaces() |
249 | ); |
250 | } |
251 | |
252 | /** |
253 | * Picks which field to index on, depending on what type of query. |
254 | * @param bool $fulltext |
255 | * @return string |
256 | */ |
257 | private function getIndexField( $fulltext ) { |
258 | return $fulltext ? 'si_text' : 'si_title'; |
259 | } |
260 | |
261 | /** |
262 | * Get the base part of the search query. |
263 | * |
264 | * @param string $filteredTerm |
265 | * @param bool $fulltext |
266 | * @return string |
267 | */ |
268 | private function queryMain( $filteredTerm, $fulltext ) { |
269 | $match = $this->parseQuery( $filteredTerm, $fulltext ); |
270 | $dbr = $this->dbProvider->getReplicaDatabase(); |
271 | $page = $dbr->tableName( 'page' ); |
272 | $searchindex = $dbr->tableName( 'searchindex' ); |
273 | return "SELECT $searchindex.rowid, page_namespace, page_title " . |
274 | "FROM $searchindex CROSS JOIN $page " . |
275 | "WHERE page_id=$searchindex.rowid AND $match"; |
276 | } |
277 | |
278 | private function getCountQuery( $filteredTerm, $fulltext ) { |
279 | $match = $this->parseQuery( $filteredTerm, $fulltext ); |
280 | $dbr = $this->dbProvider->getReplicaDatabase(); |
281 | $page = $dbr->tableName( 'page' ); |
282 | $searchindex = $dbr->tableName( 'searchindex' ); |
283 | return "SELECT COUNT(*) AS c " . |
284 | "FROM $searchindex CROSS JOIN $page " . |
285 | "WHERE page_id=$searchindex.rowid AND $match " . |
286 | $this->queryNamespaces(); |
287 | } |
288 | |
289 | /** |
290 | * Create or update the search index record for the given page. |
291 | * Title and text should be pre-processed. |
292 | * |
293 | * @param int $id |
294 | * @param string $title |
295 | * @param string $text |
296 | */ |
297 | public function update( $id, $title, $text ) { |
298 | if ( !$this->fulltextSearchSupported() ) { |
299 | return; |
300 | } |
301 | // @todo find a method to do it in a single request, |
302 | // couldn't do it so far due to typelessness of FTS3 tables. |
303 | $dbw = $this->dbProvider->getPrimaryDatabase(); |
304 | $dbw->newDeleteQueryBuilder() |
305 | ->deleteFrom( 'searchindex' ) |
306 | ->where( [ 'rowid' => $id ] ) |
307 | ->caller( __METHOD__ )->execute(); |
308 | $dbw->newInsertQueryBuilder() |
309 | ->insertInto( 'searchindex' ) |
310 | ->row( [ 'rowid' => $id, 'si_title' => $title, 'si_text' => $text ] ) |
311 | ->caller( __METHOD__ )->execute(); |
312 | } |
313 | |
314 | /** |
315 | * Update a search index record's title only. |
316 | * Title should be pre-processed. |
317 | * |
318 | * @param int $id |
319 | * @param string $title |
320 | */ |
321 | public function updateTitle( $id, $title ) { |
322 | if ( !$this->fulltextSearchSupported() ) { |
323 | return; |
324 | } |
325 | |
326 | $dbw = $this->dbProvider->getPrimaryDatabase(); |
327 | $dbw->newUpdateQueryBuilder() |
328 | ->update( 'searchindex' ) |
329 | ->set( [ 'si_title' => $title ] ) |
330 | ->where( [ 'rowid' => $id ] ) |
331 | ->caller( __METHOD__ )->execute(); |
332 | } |
333 | } |