/**
* This class contains a list of pages that the client has requested.
- * Initially, when the client passes in titles=, pageids=, or revisions= parameter,
- * an instance of the ApiPageSet class will normalize titles,
- * determine if the pages/revisions exist, and prefetch any additional data page data requested.
- *
- * When generator is used, the result of the generator will become the input for the
- * second instance of this class, and all subsequent actions will go use the second instance
- * for all their work.
- *
- * @addtogroup API
+ * Initially, when the client passes in titles=, pageids=, or revisions=
+ * parameter, an instance of the ApiPageSet class will normalize titles,
+ * determine if the pages/revisions exist, and prefetch any additional page
+ * data requested.
+ *
+ * When a generator is used, the result of the generator will become the input
+ * for the second instance of this class, and all subsequent actions will use
+ * the second instance for all their work.
+ *
+ * @ingroup API
*/
class ApiPageSet extends ApiQueryBase {
- private $mAllPages; // [ns][dbkey] => page_id or 0 when missing
- private $mTitles, $mGoodTitles, $mMissingTitles, $mMissingPageIDs, $mRedirectTitles;
+ private $mAllPages; // [ns][dbkey] => page_id or negative when missing
+ private $mTitles, $mGoodTitles, $mMissingTitles, $mInvalidTitles;
+ private $mMissingPageIDs, $mRedirectTitles;
private $mNormalizedTitles, $mInterwikiTitles;
private $mResolveRedirects, $mPendingRedirectIDs;
private $mGoodRevIDs, $mMissingRevIDs;
private $mRequestedPageFields;
+ /**
+ * Constructor
+ * @param $query ApiQuery
+ * @param $resolveRedirects bool Whether redirects should be resolved
+ */
public function __construct($query, $resolveRedirects = false) {
- parent :: __construct($query, __CLASS__);
+ parent :: __construct($query, 'query');
$this->mAllPages = array ();
$this->mTitles = array();
$this->mGoodTitles = array ();
$this->mMissingTitles = array ();
+ $this->mInvalidTitles = array ();
$this->mMissingPageIDs = array ();
$this->mRedirectTitles = array ();
$this->mNormalizedTitles = array ();
$this->mResolveRedirects = $resolveRedirects;
if($resolveRedirects)
$this->mPendingRedirectIDs = array();
-
+
$this->mFakePageId = -1;
}
+ /**
+ * Check whether this PageSet is resolving redirects
+ * @return bool
+ */
public function isResolvingRedirects() {
return $this->mResolveRedirects;
}
+ /**
+ * Request an additional field from the page table. Must be called
+ * before execute()
+ * @param $fieldName string Field name
+ */
public function requestField($fieldName) {
$this->mRequestedPageFields[$fieldName] = null;
}
+ /**
+ * Get the value of a custom field previously requested through
+ * requestField()
+ * @param $fieldName string Field name
+ * @return mixed Field value
+ */
public function getCustomField($fieldName) {
return $this->mRequestedPageFields[$fieldName];
}
/**
- * Get fields that modules have requested from the page table
+ * Get the fields that have to be queried from the page table:
+ * the ones requested through requestField() and a few basic ones
+ * we always need
+ * @return array of field names
*/
public function getPageTableFields() {
// Ensure we get minimum required fields
+ // DON'T change this order
$pageFlds = array (
- 'page_id' => null,
'page_namespace' => null,
- 'page_title' => null
+ 'page_title' => null,
+ 'page_id' => null,
);
- // only store non-default fields
- $this->mRequestedPageFields = array_diff_key($this->mRequestedPageFields, $pageFlds);
-
if ($this->mResolveRedirects)
$pageFlds['page_is_redirect'] = null;
+ // only store non-default fields
+ $this->mRequestedPageFields = array_diff_key($this->mRequestedPageFields, $pageFlds);
+
$pageFlds = array_merge($pageFlds, $this->mRequestedPageFields);
return array_keys($pageFlds);
}
/**
- * Returns an array [ns][dbkey] => page_id for all requested titles
- * page_id is a unique negative number in case title was not found
+ * Returns an array [ns][dbkey] => page_id for all requested titles.
+ * page_id is a unique negative number in case title was not found.
+ * Invalid titles will also have negative page IDs and will be in namespace 0
+ * @return array
*/
public function getAllTitlesByNamespace() {
return $this->mAllPages;
/**
* Returns the number of unique pages (not revisions) in the set.
+ * @return int
*/
public function getTitleCount() {
return count($this->mTitles);
/**
* Returns the number of found unique pages (not revisions) in the set.
+ * @return int
*/
public function getGoodTitleCount() {
return count($this->mGoodTitles);
return $this->mMissingTitles;
}
+ /**
+ * Titles that were deemed invalid by Title::newFromText()
+ * The array's index will be unique and negative for each item
+ * @return array of strings (not Title objects)
+ */
+ public function getInvalidTitles() {
+ return $this->mInvalidTitles;
+ }
+
/**
* Page IDs that were not found in the database
* @return array of page IDs
}
/**
- * Get a list of redirects when doing redirect resolution
+ * Get a list of redirect resolutions - maps a title to its redirect
+ * target.
* @return array prefixed_title (string) => prefixed_title (string)
*/
public function getRedirectTitles() {
}
/**
- * Get a list of title normalizations - maps the title given
- * with its normalized version.
- * @return array raw_prefixed_title (string) => prefixed_title (string)
+ * Get a list of title normalizations - maps a title to its normalized
+ * version.
+ * @return array raw_prefixed_title (string) => prefixed_title (string)
*/
public function getNormalizedTitles() {
return $this->mNormalizedTitles;
}
/**
- * Get a list of interwiki titles - maps the title given
- * with to the interwiki prefix.
- * @return array raw_prefixed_title (string) => interwiki_prefix (string)
+ * Get a list of interwiki titles - maps a title to its interwiki
+ * prefix.
+ * @return array raw_prefixed_title (string) => interwiki_prefix (string)
*/
public function getInterwikiTitles() {
return $this->mInterwikiTitles;
}
/**
- * Get the list of revision IDs (requested with revids= parameter)
+ * Get the list of revision IDs (requested with the revids= parameter)
* @return array revID (int) => pageID (int)
*/
public function getRevisionIDs() {
}
/**
- * Returns the number of revisions (requested with revids= parameter)
+ * Returns the number of revisions (requested with revids= parameter)\
+ * @return int
*/
public function getRevisionCount() {
return count($this->getRevisionIDs());
}
/**
- * Populate from the request parameters
+ * Populate the PageSet from the request parameters.
*/
public function execute() {
$this->profileIn();
- $titles = $pageids = $revids = null;
- extract($this->extractRequestParams());
+ $params = $this->extractRequestParams();
// Only one of the titles/pageids/revids is allowed at the same time
$dataSource = null;
- if (isset ($titles))
+ if (isset ($params['titles']))
$dataSource = 'titles';
- if (isset ($pageids)) {
+ if (isset ($params['pageids'])) {
if (isset ($dataSource))
$this->dieUsage("Cannot use 'pageids' at the same time as '$dataSource'", 'multisource');
$dataSource = 'pageids';
}
- if (isset ($revids)) {
+ if (isset ($params['revids'])) {
if (isset ($dataSource))
$this->dieUsage("Cannot use 'revids' at the same time as '$dataSource'", 'multisource');
$dataSource = 'revids';
switch ($dataSource) {
case 'titles' :
- $this->initFromTitles($titles);
+ $this->initFromTitles($params['titles']);
break;
case 'pageids' :
- $this->initFromPageIds($pageids);
+ $this->initFromPageIds($params['pageids']);
break;
case 'revids' :
if($this->mResolveRedirects)
- $this->dieUsage('revids may not be used with redirect resolution', 'params');
- $this->initFromRevIDs($revids);
+ $this->setWarning('Redirect resolution cannot be used together with the revids= parameter. '.
+ 'Any redirects the revids= point to have not been resolved.');
+ $this->mResolveRedirects = false;
+ $this->initFromRevIDs($params['revids']);
break;
default :
// Do nothing - some queries do not need any of the data sources.
}
/**
- * Initialize PageSet from a list of Titles
+ * Populate this PageSet from a list of Titles
+ * @param $titles array of Title objects
*/
public function populateFromTitles($titles) {
$this->profileIn();
}
/**
- * Initialize PageSet from a list of Page IDs
+ * Populate this PageSet from a list of page IDs
+ * @param $pageIDs array of page IDs
*/
public function populateFromPageIDs($pageIDs) {
$this->profileIn();
}
/**
- * Initialize PageSet from a rowset returned from the database
+ * Populate this PageSet from a rowset returned from the database
+ * @param $db Database object
+ * @param $queryResult Query result object
*/
public function populateFromQueryResult($db, $queryResult) {
$this->profileIn();
}
/**
- * Initialize PageSet from a list of Revision IDs
+ * Populate this PageSet from a list of revision IDs
+ * @param $revIDs array of revision IDs
*/
public function populateFromRevisionIDs($revIDs) {
$this->profileIn();
- $revIDs = array_map('intval', $revIDs); // paranoia
$this->initFromRevIDs($revIDs);
$this->profileOut();
}
/**
* Extract all requested fields from the row received from the database
+ * @param $row Result row
*/
public function processDbRow($row) {
-
+
// Store Title object in various data structures
$title = Title :: makeTitle($row->page_namespace, $row->page_title);
-
- $pageId = intval($row->page_id);
+
+ $pageId = intval($row->page_id);
$this->mAllPages[$row->page_namespace][$row->page_title] = $pageId;
$this->mTitles[] = $title;
foreach ($this->mRequestedPageFields as $fieldName => & $fieldValues)
$fieldValues[$pageId] = $row-> $fieldName;
}
-
+
+ /**
+ * Resolve redirects, if applicable
+ */
public function finishPageSetGeneration() {
$this->profileIn();
$this->resolvePendingRedirects();
$this->profileOut();
}
-
+
/**
* This method populates internal variables with page information
* based on the given array of title strings.
- *
+ *
* Steps:
* #1 For each title, get data from `page` table
* #2 If page was not found in the DB, store it as missing
- *
+ *
* Additionally, when resolving redirects:
* #3 If no more redirects left, stop.
- * #4 For each redirect, get its links from `pagelinks` table.
+ * #4 For each redirect, get its target from the `redirect` table.
* #5 Substitute the original LinkBatch object with the new list
- * #6 Repeat from step #1
+ * #6 Repeat from step #1
+ *
+ * @param $titles array of Title objects or strings
*/
private function initFromTitles($titles) {
$linkBatch = $this->processTitlesArray($titles);
if($linkBatch->isEmpty())
return;
-
+
$db = $this->getDB();
$set = $linkBatch->constructSet('page', $db);
// Get pageIDs data from the `page` table
$this->profileDBIn();
- $res = $db->select('page', $this->getPageTableFields(), $set, __METHOD__);
+ $res = $db->select('page', $this->getPageTableFields(), $set,
+ __METHOD__);
$this->profileDBOut();
// Hack: get the ns:titles stored in array(ns => array(titles)) format
$this->resolvePendingRedirects();
}
+ /**
+ * Does the same as initFromTitles(), but is based on page IDs instead
+ * @param $pageids array of page IDs
+ */
private function initFromPageIds($pageids) {
- if(empty($pageids))
+ if(!count($pageids))
return;
$pageids = array_map('intval', $pageids); // paranoia
$set = array (
'page_id' => $pageids
);
-
$db = $this->getDB();
// Get pageIDs data from the `page` table
$this->profileDBIn();
- $res = $db->select('page', $this->getPageTableFields(), $set, __METHOD__);
+ $res = $db->select('page', $this->getPageTableFields(), $set,
+ __METHOD__);
$this->profileDBOut();
-
- $this->initFromQueryResult($db, $res, array_flip($pageids), false); // process PageIDs
+
+ $remaining = array_flip($pageids);
+ $this->initFromQueryResult($db, $res, $remaining, false); // process PageIDs
// Resolve any found redirects
$this->resolvePendingRedirects();
}
-
+
/**
* Iterate through the result of the query on 'page' table,
* and for each row create and store title object and save any extra fields requested.
* @param $db Database
* @param $res DB Query result
- * @param $remaining Array of either pageID or ns/title elements (optional).
+ * @param $remaining array of either pageID or ns/title elements (optional).
* If given, any missing items will go to $mMissingPageIDs and $mMissingTitles
* @param $processTitles bool Must be provided together with $remaining.
* If true, treat $remaining as an array of [ns][title]
* If false, treat it as an array of [pageIDs]
- * @return Array of redirect IDs (only when resolving redirects)
*/
private function initFromQueryResult($db, $res, &$remaining = null, $processTitles = null) {
if (!is_null($remaining) && is_null($processTitles))
ApiBase :: dieDebug(__METHOD__, 'Missing $processTitles parameter when $remaining is provided');
-
+
while ($row = $db->fetchObject($res)) {
$pageId = intval($row->page_id);
else
unset ($remaining[$pageId]);
}
-
+
// Store any extra fields requested by modules
$this->processDbRow($row);
}
$db->freeResult($res);
-
+
if(isset($remaining)) {
// Any items left in the $remaining list are added as missing
if($processTitles) {
- // The remaining titles in $remaining are non-existant pages
+ // The remaining titles in $remaining are non-existent pages
foreach ($remaining as $ns => $dbkeys) {
foreach ( $dbkeys as $dbkey => $unused ) {
$title = Title :: makeTitle($ns, $dbkey);
else
{
// The remaining pageids do not exist
- if(empty($this->mMissingPageIDs))
+ if(!$this->mMissingPageIDs)
$this->mMissingPageIDs = array_keys($remaining);
else
$this->mMissingPageIDs = array_merge($this->mMissingPageIDs, array_keys($remaining));
}
}
+ /**
+ * Does the same as initFromTitles(), but is based on revision IDs
+ * instead
+ * @param $revids array of revision IDs
+ */
private function initFromRevIDs($revids) {
- if(empty($revids))
+ if(!count($revids))
return;
-
+
+ $revids = array_map('intval', $revids); // paranoia
$db = $this->getDB();
$pageids = array();
$remaining = array_flip($revids);
-
- $tables = array('revision');
- $fields = array('rev_id','rev_page');
- $where = array('rev_deleted' => 0, 'rev_id' => $revids);
-
+
+ $tables = array('revision', 'page');
+ $fields = array('rev_id', 'rev_page');
+ $where = array('rev_id' => $revids, 'rev_page = page_id');
+
// Get pageIDs data from the `page` table
$this->profileDBIn();
- $res = $db->select( $tables, $fields, $where, __METHOD__ );
- while ( $row = $db->fetchObject( $res ) ) {
+ $res = $db->select($tables, $fields, $where, __METHOD__);
+ while ($row = $db->fetchObject($res)) {
$revid = intval($row->rev_id);
$pageid = intval($row->rev_page);
$this->mGoodRevIDs[$revid] = $pageid;
$pageids[$pageid] = '';
unset($remaining[$revid]);
}
- $db->freeResult( $res );
+ $db->freeResult($res);
$this->profileDBOut();
$this->mMissingRevIDs = array_keys($remaining);
// Populate all the page information
- if($this->mResolveRedirects)
- ApiBase :: dieDebug(__METHOD__, 'revids may not be used with redirect resolution');
$this->initFromPageIds(array_keys($pageids));
}
+ /**
+ * Resolve any redirects in the result if redirect resolution was
+ * requested. This function is called repeatedly until all redirects
+ * have been resolved.
+ */
private function resolvePendingRedirects() {
if($this->mResolveRedirects) {
$db = $this->getDB();
$pageFlds = $this->getPageTableFields();
-
+
// Repeat until all redirects have been resolved
// The infinite loop is prevented by keeping all known pages in $this->mAllPages
- while (!empty ($this->mPendingRedirectIDs)) {
-
+ while ($this->mPendingRedirectIDs) {
+
// Resolve redirects by querying the pagelinks table, and repeat the process
// Create a new linkBatch object for the next pass
$linkBatch = $this->getRedirectTargets();
-
+
if ($linkBatch->isEmpty())
break;
-
+
$set = $linkBatch->constructSet('page', $db);
- if(false === $set)
+ if($set === false)
break;
-
+
// Get pageIDs data from the `page` table
$this->profileDBIn();
$res = $db->select('page', $pageFlds, $set, __METHOD__);
$this->profileDBOut();
-
+
// Hack: get the ns:titles stored in array(ns => array(titles)) format
$this->initFromQueryResult($db, $res, $linkBatch->data, true);
}
}
}
+ /**
+ * Get the targets of the pending redirects from the database
+ *
+ * Also creates entries in the redirect table for redirects that don't
+ * have one.
+ * @return LinkBatch
+ */
private function getRedirectTargets() {
-
- $linkBatch = new LinkBatch();
+ $lb = new LinkBatch();
$db = $this->getDB();
- // find redirect targets for all redirect pages
$this->profileDBIn();
- $res = $db->select('pagelinks', array (
- 'pl_from',
- 'pl_namespace',
- 'pl_title'
- ), array (
- 'pl_from' => array_keys($this->mPendingRedirectIDs
- )), __METHOD__);
+ $res = $db->select('redirect', array(
+ 'rd_from',
+ 'rd_namespace',
+ 'rd_title'
+ ), array('rd_from' => array_keys($this->mPendingRedirectIDs)),
+ __METHOD__
+ );
$this->profileDBOut();
- while ($row = $db->fetchObject($res)) {
-
- $plfrom = intval($row->pl_from);
-
- // Bug 7304 workaround
- // ( http://bugzilla.wikipedia.org/show_bug.cgi?id=7304 )
- // A redirect page may have more than one link.
- // This code will only use the first link returned.
- if (isset ($this->mPendingRedirectIDs[$plfrom])) { // remove line when bug 7304 is fixed
-
- $titleStrFrom = $this->mPendingRedirectIDs[$plfrom]->getPrefixedText();
- $titleStrTo = Title :: makeTitle($row->pl_namespace, $row->pl_title)->getPrefixedText();
- unset ($this->mPendingRedirectIDs[$plfrom]); // remove line when bug 7304 is fixed
-
- // Avoid an infinite loop by checking if we have already processed this target
- if (!isset ($this->mAllPages[$row->pl_namespace][$row->pl_title])) {
- $linkBatch->add($row->pl_namespace, $row->pl_title);
- }
- } else {
- // This redirect page has more than one link.
- // This is very slow, but safer until bug 7304 is resolved
- $title = Title :: newFromID($plfrom);
- $titleStrFrom = $title->getPrefixedText();
-
+ while($row = $db->fetchObject($res))
+ {
+ $rdfrom = intval($row->rd_from);
+ $from = $this->mPendingRedirectIDs[$rdfrom]->getPrefixedText();
+ $to = Title::makeTitle($row->rd_namespace, $row->rd_title)->getPrefixedText();
+ unset($this->mPendingRedirectIDs[$rdfrom]);
+ if(!isset($this->mAllPages[$row->rd_namespace][$row->rd_title]))
+ $lb->add($row->rd_namespace, $row->rd_title);
+ $this->mRedirectTitles[$from] = $to;
+ }
+ $db->freeResult($res);
+ if($this->mPendingRedirectIDs)
+ {
+ # We found pages that aren't in the redirect table
+ # Add them
+ foreach($this->mPendingRedirectIDs as $id => $title)
+ {
$article = new Article($title);
- $text = $article->getContent();
- $titleTo = Title :: newFromRedirect($text);
- $titleStrTo = $titleTo->getPrefixedText();
-
- if (is_null($titleStrTo))
- ApiBase :: dieDebug(__METHOD__, 'Bug7304 workaround: redir target from {$title->getPrefixedText()} not found');
-
- // Avoid an infinite loop by checking if we have already processed this target
- if (!isset ($this->mAllPages[$titleTo->getNamespace()][$titleTo->getDBkey()])) {
- $linkBatch->addObj($titleTo);
- }
+ $rt = $article->insertRedirect();
+ if(!$rt)
+ # What the hell. Let's just ignore this
+ continue;
+ $lb->addObj($rt);
+ $this->mRedirectTitles[$title->getPrefixedText()] = $rt->getPrefixedText();
+ unset($this->mPendingRedirectIDs[$id]);
}
-
- $this->mRedirectTitles[$titleStrFrom] = $titleStrTo;
}
- $db->freeResult($res);
-
- // All IDs must exist in the page table
- if (!empty($this->mPendingRedirectIDs[$plfrom]))
- ApiBase :: dieDebug(__METHOD__, 'Invalid redirect IDs were found');
-
- return $linkBatch;
+ return $lb;
}
/**
* Given an array of title strings, convert them into Title objects.
* Alternativelly, an array of Title objects may be given.
- * This method validates access rights for the title,
+ * This method validates access rights for the title,
* and appends normalization values to the output.
- *
- * @return LinkBatch of title objects.
+ *
+ * @param $titles array of Title objects or strings
+ * @return LinkBatch
*/
private function processTitlesArray($titles) {
$linkBatch = new LinkBatch();
foreach ($titles as $title) {
-
+
$titleObj = is_string($title) ? Title :: newFromText($title) : $title;
if (!$titleObj)
- $this->dieUsage("bad title", 'invalidtitle');
-
+ {
+ # Handle invalid titles gracefully
+ $this->mAllpages[0][$title] = $this->mFakePageId;
+ $this->mInvalidTitles[$this->mFakePageId] = $title;
+ $this->mFakePageId--;
+ continue; // There's nothing else we can do
+ }
$iw = $titleObj->getInterwiki();
- if (!empty($iw)) {
+ if (strval($iw) !== '') {
// This title is an interwiki link.
$this->mInterwikiTitles[$titleObj->getPrefixedText()] = $iw;
- } else {
+ } else {
// Validation
if ($titleObj->getNamespace() < 0)
- $this->dieUsage("No support for special pages has been implemented", 'unsupportednamespace');
-
- $linkBatch->addObj($titleObj);
+ $this->setWarning("No support for special pages has been implemented");
+ else
+ $linkBatch->addObj($titleObj);
}
-
- // Make sure we remember the original title that was given to us
- // This way the caller can correlate new titles with the originally requested,
- // i.e. namespace is localized or capitalization is different
+
+ // Make sure we remember the original title that was
+ // given to us. This way the caller can correlate new
+ // titles with the originally requested when e.g. the
+ // namespace is localized or the capitalization is
+ // different
if (is_string($title) && $title !== $titleObj->getPrefixedText()) {
$this->mNormalizedTitles[$title] = $titleObj->getPrefixedText();
}
return $linkBatch;
}
-
+
protected function getAllowedParams() {
return array (
'titles' => array (
}
public function getVersion() {
- return __CLASS__ . ': $Id: ApiPageSet.php 24935 2007-08-20 08:13:16Z nickj $';
+ return __CLASS__ . ': $Id: ApiPageSet.php 47424 2009-02-18 05:29:11Z werdna $';
}
}
-