MediaWiki  1.29.1
SearchMySQL.php
Go to the documentation of this file.
1 <?php
31 class SearchMySQL extends SearchDatabase {
32  protected $strictMatching = true;
33 
34  private static $mMinSearchLength;
35 
45  function parseQuery( $filteredText, $fulltext ) {
47 
48  $lc = $this->legalSearchChars( self::CHARS_NO_SYNTAX ); // Minus syntax chars (" and *)
49  $searchon = '';
50  $this->searchTerms = [];
51 
52  # @todo FIXME: This doesn't handle parenthetical expressions.
53  $m = [];
54  if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/',
55  $filteredText, $m, PREG_SET_ORDER ) ) {
56  foreach ( $m as $bits ) {
57  MediaWiki\suppressWarnings();
58  list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits;
59  MediaWiki\restoreWarnings();
60 
61  if ( $nonQuoted != '' ) {
62  $term = $nonQuoted;
63  $quote = '';
64  } else {
65  $term = str_replace( '"', '', $term );
66  $quote = '"';
67  }
68 
69  if ( $searchon !== '' ) {
70  $searchon .= ' ';
71  }
72  if ( $this->strictMatching && ( $modifier == '' ) ) {
73  // If we leave this out, boolean op defaults to OR which is rarely helpful.
74  $modifier = '+';
75  }
76 
77  // Some languages such as Serbian store the input form in the search index,
78  // so we may need to search for matches in multiple writing system variants.
79  $convertedVariants = $wgContLang->autoConvertToAllVariants( $term );
80  if ( is_array( $convertedVariants ) ) {
81  $variants = array_unique( array_values( $convertedVariants ) );
82  } else {
83  $variants = [ $term ];
84  }
85 
86  // The low-level search index does some processing on input to work
87  // around problems with minimum lengths and encoding in MySQL's
88  // fulltext engine.
89  // For Chinese this also inserts spaces between adjacent Han characters.
90  $strippedVariants = array_map(
91  [ $wgContLang, 'normalizeForSearch' ],
92  $variants );
93 
94  // Some languages such as Chinese force all variants to a canonical
95  // form when stripping to the low-level search index, so to be sure
96  // let's check our variants list for unique items after stripping.
97  $strippedVariants = array_unique( $strippedVariants );
98 
99  $searchon .= $modifier;
100  if ( count( $strippedVariants ) > 1 ) {
101  $searchon .= '(';
102  }
103  foreach ( $strippedVariants as $stripped ) {
104  $stripped = $this->normalizeText( $stripped );
105  if ( $nonQuoted && strpos( $stripped, ' ' ) !== false ) {
106  // Hack for Chinese: we need to toss in quotes for
107  // multiple-character phrases since normalizeForSearch()
108  // added spaces between them to make word breaks.
109  $stripped = '"' . trim( $stripped ) . '"';
110  }
111  $searchon .= "$quote$stripped$quote$wildcard ";
112  }
113  if ( count( $strippedVariants ) > 1 ) {
114  $searchon .= ')';
115  }
116 
117  // Match individual terms or quoted phrase in result highlighting...
118  // Note that variants will be introduced in a later stage for highlighting!
119  $regexp = $this->regexTerm( $term, $wildcard );
120  $this->searchTerms[] = $regexp;
121  }
122  wfDebug( __METHOD__ . ": Would search with '$searchon'\n" );
123  wfDebug( __METHOD__ . ': Match with /' . implode( '|', $this->searchTerms ) . "/\n" );
124  } else {
125  wfDebug( __METHOD__ . ": Can't understand search query '{$filteredText}'\n" );
126  }
127 
128  $searchon = $this->db->addQuotes( $searchon );
129  $field = $this->getIndexField( $fulltext );
130  return " MATCH($field) AGAINST($searchon IN BOOLEAN MODE) ";
131  }
132 
133  function regexTerm( $string, $wildcard ) {
135 
136  $regex = preg_quote( $string, '/' );
137  if ( $wgContLang->hasWordBreaks() ) {
138  if ( $wildcard ) {
139  // Don't cut off the final bit!
140  $regex = "\b$regex";
141  } else {
142  $regex = "\b$regex\b";
143  }
144  } else {
145  // For Chinese, words may legitimately abut other words in the text literal.
146  // Don't add \b boundary checks... note this could cause false positives
147  // for latin chars.
148  }
149  return $regex;
150  }
151 
152  public static function legalSearchChars( $type = self::CHARS_ALL ) {
153  $searchChars = parent::legalSearchChars( $type );
154  if ( $type === self::CHARS_ALL ) {
155  // " for phrase, * for wildcard
156  $searchChars = "\"*" . $searchChars;
157  }
158  return $searchChars;
159  }
160 
167  function searchText( $term ) {
168  return $this->searchInternal( $term, true );
169  }
170 
177  function searchTitle( $term ) {
178  return $this->searchInternal( $term, false );
179  }
180 
181  protected function searchInternal( $term, $fulltext ) {
182  // This seems out of place, why is this called with empty term?
183  if ( trim( $term ) === '' ) {
184  return null;
185  }
186 
187  $filteredTerm = $this->filter( $term );
188  $query = $this->getQuery( $filteredTerm, $fulltext );
189  $resultSet = $this->db->select(
190  $query['tables'], $query['fields'], $query['conds'],
191  __METHOD__, $query['options'], $query['joins']
192  );
193 
194  $total = null;
195  $query = $this->getCountQuery( $filteredTerm, $fulltext );
196  $totalResult = $this->db->select(
197  $query['tables'], $query['fields'], $query['conds'],
198  __METHOD__, $query['options'], $query['joins']
199  );
200 
201  $row = $totalResult->fetchObject();
202  if ( $row ) {
203  $total = intval( $row->c );
204  }
205  $totalResult->free();
206 
207  return new SqlSearchResultSet( $resultSet, $this->searchTerms, $total );
208  }
209 
210  public function supports( $feature ) {
211  switch ( $feature ) {
212  case 'title-suffix-filter':
213  return true;
214  default:
215  return parent::supports( $feature );
216  }
217  }
218 
224  protected function queryFeatures( &$query ) {
225  foreach ( $this->features as $feature => $value ) {
226  if ( $feature === 'title-suffix-filter' && $value ) {
227  $query['conds'][] = 'page_title' . $this->db->buildLike( $this->db->anyString(), $value );
228  }
229  }
230  }
231 
237  function queryNamespaces( &$query ) {
238  if ( is_array( $this->namespaces ) ) {
239  if ( count( $this->namespaces ) === 0 ) {
240  $this->namespaces[] = '0';
241  }
242  $query['conds']['page_namespace'] = $this->namespaces;
243  }
244  }
245 
251  protected function limitResult( &$query ) {
252  $query['options']['LIMIT'] = $this->limit;
253  $query['options']['OFFSET'] = $this->offset;
254  }
255 
264  function getQuery( $filteredTerm, $fulltext ) {
265  $query = [
266  'tables' => [],
267  'fields' => [],
268  'conds' => [],
269  'options' => [],
270  'joins' => [],
271  ];
272 
273  $this->queryMain( $query, $filteredTerm, $fulltext );
274  $this->queryFeatures( $query );
275  $this->queryNamespaces( $query );
276  $this->limitResult( $query );
277 
278  return $query;
279  }
280 
286  function getIndexField( $fulltext ) {
287  return $fulltext ? 'si_text' : 'si_title';
288  }
289 
298  function queryMain( &$query, $filteredTerm, $fulltext ) {
299  $match = $this->parseQuery( $filteredTerm, $fulltext );
300  $query['tables'][] = 'page';
301  $query['tables'][] = 'searchindex';
302  $query['fields'][] = 'page_id';
303  $query['fields'][] = 'page_namespace';
304  $query['fields'][] = 'page_title';
305  $query['conds'][] = 'page_id=si_page';
306  $query['conds'][] = $match;
307  }
308 
315  function getCountQuery( $filteredTerm, $fulltext ) {
316  $match = $this->parseQuery( $filteredTerm, $fulltext );
317 
318  $query = [
319  'tables' => [ 'page', 'searchindex' ],
320  'fields' => [ 'COUNT(*) as c' ],
321  'conds' => [ 'page_id=si_page', $match ],
322  'options' => [],
323  'joins' => [],
324  ];
325 
326  $this->queryFeatures( $query );
327  $this->queryNamespaces( $query );
328 
329  return $query;
330  }
331 
340  function update( $id, $title, $text ) {
341  $dbw = wfGetDB( DB_MASTER );
342  $dbw->replace( 'searchindex',
343  [ 'si_page' ],
344  [
345  'si_page' => $id,
346  'si_title' => $this->normalizeText( $title ),
347  'si_text' => $this->normalizeText( $text )
348  ], __METHOD__ );
349  }
350 
358  function updateTitle( $id, $title ) {
359  $dbw = wfGetDB( DB_MASTER );
360 
361  $dbw->update( 'searchindex',
362  [ 'si_title' => $this->normalizeText( $title ) ],
363  [ 'si_page' => $id ],
364  __METHOD__,
365  [ $dbw->lowPriorityOption() ] );
366  }
367 
375  function delete( $id, $title ) {
376  $dbw = wfGetDB( DB_MASTER );
377 
378  $dbw->delete( 'searchindex', [ 'si_page' => $id ], __METHOD__ );
379  }
380 
387  function normalizeText( $string ) {
389 
390  $out = parent::normalizeText( $string );
391 
392  // MySQL fulltext index doesn't grok utf-8, so we
393  // need to fold cases and convert to hex
394  $out = preg_replace_callback(
395  "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
396  [ $this, 'stripForSearchCallback' ],
397  $wgContLang->lc( $out ) );
398 
399  // And to add insult to injury, the default indexing
400  // ignores short words... Pad them so we can pass them
401  // through without reconfiguring the server...
402  $minLength = $this->minSearchLength();
403  if ( $minLength > 1 ) {
404  $n = $minLength - 1;
405  $out = preg_replace(
406  "/\b(\w{1,$n})\b/",
407  "$1u800",
408  $out );
409  }
410 
411  // Periods within things like hostnames and IP addresses
412  // are also important -- we want a search for "example.com"
413  // or "192.168.1.1" to work sanely.
414  // MySQL's search seems to ignore them, so you'd match on
415  // "example.wikipedia.com" and "192.168.83.1" as well.
416  $out = preg_replace(
417  "/(\w)\.(\w|\*)/u",
418  "$1u82e$2",
419  $out );
420 
421  return $out;
422  }
423 
431  protected function stripForSearchCallback( $matches ) {
432  return 'u8' . bin2hex( $matches[1] );
433  }
434 
441  protected function minSearchLength() {
442  if ( is_null( self::$mMinSearchLength ) ) {
443  $sql = "SHOW GLOBAL VARIABLES LIKE 'ft\\_min\\_word\\_len'";
444 
445  $dbr = wfGetDB( DB_REPLICA );
446  $result = $dbr->query( $sql, __METHOD__ );
447  $row = $result->fetchObject();
448  $result->free();
449 
450  if ( $row && $row->Variable_name == 'ft_min_word_len' ) {
451  self::$mMinSearchLength = intval( $row->Value );
452  } else {
453  self::$mMinSearchLength = 0;
454  }
455  }
457  }
458 }
SearchEngine\$limit
int $limit
Definition: SearchEngine.php:42
SearchMySQL\limitResult
limitResult(&$query)
Add limit options.
Definition: SearchMySQL.php:251
SearchMySQL\queryNamespaces
queryNamespaces(&$query)
Add namespace conditions.
Definition: SearchMySQL.php:237
captcha-old.count
count
Definition: captcha-old.py:225
SearchMySQL
Search engine hook for MySQL 4+.
Definition: SearchMySQL.php:31
SearchMySQL\searchTitle
searchTitle( $term)
Perform a title-only search query and return a result set.
Definition: SearchMySQL.php:177
SearchMySQL\minSearchLength
minSearchLength()
Check MySQL server's ft_min_word_len setting so we know if we need to pad short words....
Definition: SearchMySQL.php:441
$result
The index of the header message $result[1]=The index of the body text message $result[2 through n]=Parameters passed to body text message. Please note the header message cannot receive/use parameters. 'ImportHandleLogItemXMLTag':When parsing a XML tag in a log item. Return false to stop further processing of the tag $reader:XMLReader object $logInfo:Array of information 'ImportHandlePageXMLTag':When parsing a XML tag in a page. Return false to stop further processing of the tag $reader:XMLReader object & $pageInfo:Array of information 'ImportHandleRevisionXMLTag':When parsing a XML tag in a page revision. Return false to stop further processing of the tag $reader:XMLReader object $pageInfo:Array of page information $revisionInfo:Array of revision information 'ImportHandleToplevelXMLTag':When parsing a top level XML tag. Return false to stop further processing of the tag $reader:XMLReader object 'ImportHandleUploadXMLTag':When parsing a XML tag in a file upload. Return false to stop further processing of the tag $reader:XMLReader object $revisionInfo:Array of information 'ImportLogInterwikiLink':Hook to change the interwiki link used in log entries and edit summaries for transwiki imports. & $fullInterwikiPrefix:Interwiki prefix, may contain colons. & $pageTitle:String that contains page title. 'ImportSources':Called when reading from the $wgImportSources configuration variable. Can be used to lazy-load the import sources list. & $importSources:The value of $wgImportSources. Modify as necessary. See the comment in DefaultSettings.php for the detail of how to structure this array. 'InfoAction':When building information to display on the action=info page. $context:IContextSource object & $pageInfo:Array of information 'InitializeArticleMaybeRedirect':MediaWiki check to see if title is a redirect. & $title:Title object for the current page & $request:WebRequest & $ignoreRedirect:boolean to skip redirect check & $target:Title/string of redirect target & $article:Article object 'InternalParseBeforeLinks':during Parser 's internalParse method before links but after nowiki/noinclude/includeonly/onlyinclude and other processings. & $parser:Parser object & $text:string containing partially parsed text & $stripState:Parser 's internal StripState object 'InternalParseBeforeSanitize':during Parser 's internalParse method just before the parser removes unwanted/dangerous HTML tags and after nowiki/noinclude/includeonly/onlyinclude and other processings. Ideal for syntax-extensions after template/parser function execution which respect nowiki and HTML-comments. & $parser:Parser object & $text:string containing partially parsed text & $stripState:Parser 's internal StripState object 'InterwikiLoadPrefix':When resolving if a given prefix is an interwiki or not. Return true without providing an interwiki to continue interwiki search. $prefix:interwiki prefix we are looking for. & $iwData:output array describing the interwiki with keys iw_url, iw_local, iw_trans and optionally iw_api and iw_wikiid. 'InvalidateEmailComplete':Called after a user 's email has been invalidated successfully. $user:user(object) whose email is being invalidated 'IRCLineURL':When constructing the URL to use in an IRC notification. Callee may modify $url and $query, URL will be constructed as $url . $query & $url:URL to index.php & $query:Query string $rc:RecentChange object that triggered url generation 'IsFileCacheable':Override the result of Article::isFileCacheable()(if true) & $article:article(object) being checked 'IsTrustedProxy':Override the result of IP::isTrustedProxy() & $ip:IP being check & $result:Change this value to override the result of IP::isTrustedProxy() 'IsUploadAllowedFromUrl':Override the result of UploadFromUrl::isAllowedUrl() $url:URL used to upload from & $allowed:Boolean indicating if uploading is allowed for given URL 'isValidEmailAddr':Override the result of Sanitizer::validateEmail(), for instance to return false if the domain name doesn 't match your organization. $addr:The e-mail address entered by the user & $result:Set this and return false to override the internal checks 'isValidPassword':Override the result of User::isValidPassword() $password:The password entered by the user & $result:Set this and return false to override the internal checks $user:User the password is being validated for 'Language::getMessagesFileName':$code:The language code or the language we 're looking for a messages file for & $file:The messages file path, you can override this to change the location. 'LanguageGetMagic':DEPRECATED! Use $magicWords in a file listed in $wgExtensionMessagesFiles instead. Use this to define synonyms of magic words depending of the language & $magicExtensions:associative array of magic words synonyms $lang:language code(string) 'LanguageGetNamespaces':Provide custom ordering for namespaces or remove namespaces. Do not use this hook to add namespaces. Use CanonicalNamespaces for that. & $namespaces:Array of namespaces indexed by their numbers 'LanguageGetSpecialPageAliases':DEPRECATED! Use $specialPageAliases in a file listed in $wgExtensionMessagesFiles instead. Use to define aliases of special pages names depending of the language & $specialPageAliases:associative array of magic words synonyms $lang:language code(string) 'LanguageGetTranslatedLanguageNames':Provide translated language names. & $names:array of language code=> language name $code:language of the preferred translations 'LanguageLinks':Manipulate a page 's language links. This is called in various places to allow extensions to define the effective language links for a page. $title:The page 's Title. & $links:Array with elements of the form "language:title" in the order that they will be output. & $linkFlags:Associative array mapping prefixed links to arrays of flags. Currently unused, but planned to provide support for marking individual language links in the UI, e.g. for featured articles. 'LanguageSelector':Hook to change the language selector available on a page. $out:The output page. $cssClassName:CSS class name of the language selector. 'LinkBegin':DEPRECATED! Use HtmlPageLinkRendererBegin instead. Used when generating internal and interwiki links in Linker::link(), before processing starts. Return false to skip default processing and return $ret. See documentation for Linker::link() for details on the expected meanings of parameters. $skin:the Skin object $target:the Title that the link is pointing to & $html:the contents that the< a > tag should have(raw HTML) $result
Definition: hooks.txt:1954
SearchMySQL\getIndexField
getIndexField( $fulltext)
Picks which field to index on, depending on what type of query.
Definition: SearchMySQL.php:286
SearchEngine\$namespaces
int[] null $namespaces
Definition: SearchEngine.php:39
SearchMySQL\normalizeText
normalizeText( $string)
Converts some characters for MySQL's indexing to grok it correctly, and pads short words to overcome ...
Definition: SearchMySQL.php:387
SearchMySQL\queryMain
queryMain(&$query, $filteredTerm, $fulltext)
Get the base part of the search query.
Definition: SearchMySQL.php:298
$term
external whereas SearchGetNearMatch runs after $term
Definition: hooks.txt:2759
$type
do that in ParserLimitReportFormat instead use this to modify the parameters of the image and a DIV can begin in one section and end in another Make sure your code can handle that case gracefully See the EditSectionClearerLink extension for an example zero but section is usually empty its values are the globals values before the output is cached my talk my contributions etc etc otherwise the built in rate limiting checks are if enabled allows for interception of redirect as a string mapping parameter names to values & $type
Definition: hooks.txt:2536
SearchMySQL\queryFeatures
queryFeatures(&$query)
Add special conditions.
Definition: SearchMySQL.php:224
php
injection txt This is an overview of how MediaWiki makes use of dependency injection The design described here grew from the discussion of RFC T384 The term dependency this means that anything an object needs to operate should be injected from the the object itself should only know narrow no concrete implementation of the logic it relies on The requirement to inject everything typically results in an architecture that based on two main types of and essentially stateless service objects that use other service objects to operate on the value objects As of the beginning MediaWiki is only starting to use the DI approach Much of the code still relies on global state or direct resulting in a highly cyclical dependency which acts as the top level factory for services in MediaWiki which can be used to gain access to default instances of various services MediaWikiServices however also allows new services to be defined and default services to be redefined Services are defined or redefined by providing a callback the instantiator that will return a new instance of the service When it will create an instance of MediaWikiServices and populate it with the services defined in the files listed by thereby bootstrapping the DI framework Per $wgServiceWiringFiles lists includes ServiceWiring php
Definition: injection.txt:35
SearchDatabase
Base search engine base class for database-backed searches.
Definition: SearchDatabase.php:31
$query
null for the wiki Added should default to null in handler for backwards compatibility add a value to it if you want to add a cookie that have to vary cache options can modify $query
Definition: hooks.txt:1572
namespaces
to move a page</td >< td > &*You are moving the page across namespaces
Definition: All_system_messages.txt:2670
SearchMySQL\stripForSearchCallback
stripForSearchCallback( $matches)
Armor a case-folded UTF-8 string to get through MySQL's fulltext search without being mucked up by fu...
Definition: SearchMySQL.php:431
SearchMySQL\updateTitle
updateTitle( $id, $title)
Update a search index record's title only.
Definition: SearchMySQL.php:358
$title
namespace and then decline to actually register it file or subcat img or subcat $title
Definition: hooks.txt:934
SearchEngine\$offset
int $offset
Definition: SearchEngine.php:45
SearchMySQL\regexTerm
regexTerm( $string, $wildcard)
Definition: SearchMySQL.php:133
wfGetDB
wfGetDB( $db, $groups=[], $wiki=false)
Get a Database object.
Definition: GlobalFunctions.php:3060
SearchMySQL\getQuery
getQuery( $filteredTerm, $fulltext)
Construct the SQL query to do the search.
Definition: SearchMySQL.php:264
$matches
$matches
Definition: NoLocalSettings.php:24
SearchMySQL\getCountQuery
getCountQuery( $filteredTerm, $fulltext)
Definition: SearchMySQL.php:315
global
when a variable name is used in a it is silently declared as a new masking the global
Definition: design.txt:93
DB_REPLICA
const DB_REPLICA
Definition: defines.php:25
DB_MASTER
const DB_MASTER
Definition: defines.php:26
wfDebug
wfDebug( $text, $dest='all', array $context=[])
Sends a line to the debug log if enabled or, optionally, to a comment in output.
Definition: GlobalFunctions.php:999
list
deferred txt A few of the database updates required by various functions here can be deferred until after the result page is displayed to the user For updating the view updating the linked to tables after a etc PHP does not yet have any way to tell the server to actually return and disconnect while still running these but it might have such a feature in the future We handle these by creating a deferred update object and putting those objects on a global list
Definition: deferred.txt:11
SqlSearchResultSet
This class is used for different SQL-based search engines shipped with MediaWiki.
Definition: SqlSearchResultSet.php:9
$value
$value
Definition: styleTest.css.php:45
SearchMySQL\supports
supports( $feature)
Definition: SearchMySQL.php:210
SearchMySQL\parseQuery
parseQuery( $filteredText, $fulltext)
Parse the user's query and transform it into an SQL fragment which will become part of a WHERE clause...
Definition: SearchMySQL.php:45
SearchMySQL\update
update( $id, $title, $text)
Create or update the search index record for the given page.
Definition: SearchMySQL.php:340
SearchMySQL\$strictMatching
$strictMatching
Definition: SearchMySQL.php:32
$dbr
if(! $regexes) $dbr
Definition: cleanup.php:94
SearchMySQL\$mMinSearchLength
static $mMinSearchLength
Definition: SearchMySQL.php:34
as
This document is intended to provide useful advice for parties seeking to redistribute MediaWiki to end users It s targeted particularly at maintainers for Linux since it s been observed that distribution packages of MediaWiki often break We ve consistently had to recommend that users seeking support use official tarballs instead of their distribution s and this often solves whatever problem the user is having It would be nice if this could such as
Definition: distributors.txt:9
SearchMySQL\legalSearchChars
static legalSearchChars( $type=self::CHARS_ALL)
Get chars legal for search NOTE: usage as static is deprecated and preserved only as BC measure.
Definition: SearchMySQL.php:152
SearchMySQL\searchText
searchText( $term)
Perform a full text search query and return a result set.
Definition: SearchMySQL.php:167
SearchDatabase\filter
filter( $text)
Return a 'cleaned up' search string.
Definition: SearchDatabase.php:55
SearchMySQL\searchInternal
searchInternal( $term, $fulltext)
Definition: SearchMySQL.php:181
$wgContLang
this class mediates it Skin Encapsulates a look and feel for the wiki All of the functions that render HTML and make choices about how to render it are here and are called from various other places when and is meant to be subclassed with other skins that may override some of its functions The User object contains a reference to a and so rather than having a global skin object we just rely on the global User and get the skin with $wgUser and also has some character encoding functions and other locale stuff The current user interface language is instantiated as and the content language as $wgContLang
Definition: design.txt:56
$out
this hook is for auditing only or null if authentication failed before getting that far or null if we can t even determine that probably a stub it is not rendered in wiki pages or galleries in category pages allow injecting custom HTML after the section Any uses of the hook need to handle escaping see BaseTemplate::getToolbox and BaseTemplate::makeListItem for details on the format of individual items inside of this array or by returning and letting standard HTTP rendering take place modifiable or by returning false and taking over the output $out
Definition: hooks.txt:783