Physikerwelt has uploaded a new change for review.
https://gerrit.wikimedia.org/r/200100
Change subject: Live index update
......................................................................
Live index update
* use onPageContentSaveComplete hook
* still quite prototypical
Change-Id: Iff6e07e2276f32c11f4b828e55ba9e27bd33aa5d
---
M MathSearch.hooks.php
M MathSearch.php
A includes/MwsDumpWriter.php
M includes/engines/MathEngineBaseX.php
M includes/engines/MathEngineRest.php
M maintenance/CreateMWSHarvest.php
M maintenance/UpdateMath.php
A tests/MwsDumpWriterTest.php
8 files changed, 464 insertions(+), 79 deletions(-)
git pull ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/MathSearch
refs/changes/00/200100/1
diff --git a/MathSearch.hooks.php b/MathSearch.hooks.php
index de78e6d..83ae417 100644
--- a/MathSearch.hooks.php
+++ b/MathSearch.hooks.php
@@ -1,11 +1,11 @@
<?php
+
/**
* MediaWiki MathSearch extension
*
* (c) 2012 various MediaWiki contributors
* GPLv2 license; info in main package.
*/
-
class MathSearchHooks {
static $nextID = 0;
@@ -35,29 +35,31 @@
if ( is_null( $updater ) ) {
throw new MWException( "Mathsearch extension requires
Mediawiki 1.18 or above" );
}
- $type = $updater->getDB()->getType();
- if ( $type == "mysql" ) {
- $dir = __DIR__ . '/db/' ;
+ $type = $updater->getDB()->getType();
+ if ( $type == "mysql" ) {
+ $dir = __DIR__ . '/db/';
$updater->addExtensionTable( 'mathindex', $dir .
'mathindex.sql' );
- $updater->addExtensionTable( 'mathobservation', $dir .
'mathobservation.sql' );
+ $updater->addExtensionTable( 'mathobservation', $dir .
'mathobservation.sql' );
$updater->addExtensionTable( 'mathvarstat', $dir .
'mathvarstat.sql' );
$updater->addExtensionTable( 'mathrevisionstat', $dir .
'mathrevisionstat.sql' );
$updater->addExtensionTable( 'mathsemantics', $dir .
'mathsemantics.sql' );
$updater->addExtensionTable( 'mathperformance', $dir .
'mathperformance.sql' );
$updater->addExtensionTable( 'mathidentifier', $dir .
'mathidentifier.sql' );
- if ( $wgMathWmcServer ){
+ if ( $wgMathWmcServer ) {
$wmcDir = $dir . 'wmc/persistent/';
- $updater->addExtensionTable( 'math_wmc_ref',
$wmcDir . "math_wmc_ref.sql");
- $updater->addExtensionTable( 'math_wmc_runs',
$wmcDir . "math_wmc_runs.sql");
- $updater->addExtensionTable(
'math_wmc_results', $wmcDir . "math_wmc_results.sql");
- $updater->addExtensionTable(
'math_wmc_assessed_formula', $wmcDir . "math_wmc_assessed_formula.sql");
- $updater->addExtensionTable(
'math_wmc_assessed_revision', $wmcDir . "math_wmc_assessed_revision.sql");
+ $updater->addExtensionTable( 'math_wmc_ref',
$wmcDir . "math_wmc_ref.sql" );
+ $updater->addExtensionTable( 'math_wmc_runs',
$wmcDir . "math_wmc_runs.sql" );
+ $updater->addExtensionTable(
'math_wmc_results', $wmcDir . "math_wmc_results.sql" );
+ $updater->addExtensionTable(
'math_wmc_assessed_formula',
+ $wmcDir .
"math_wmc_assessed_formula.sql" );
+ $updater->addExtensionTable(
'math_wmc_assessed_revision',
+ $wmcDir .
"math_wmc_assessed_revision.sql" );
}
- } elseif ( $type == 'sqlite' ){
+ } elseif ( $type == 'sqlite' ) {
// Don't scare Jenkins with an exception.
} else {
- throw new Exception( "Math extension does not currently
support $type database." );
- }
+ throw new Exception( "Math extension does not currently
support $type database." );
+ }
return true;
}
@@ -65,8 +67,8 @@
* Checks if the db2 php client is installed
* @return boolean
*/
- public static function isDB2Supported(){
- if ( function_exists('db2_connect') ){
+ public static function isDB2Supported() {
+ if ( function_exists( 'db2_connect' ) ) {
return true;
} else {
return false;
@@ -81,7 +83,7 @@
* @param string $inputHash hash of tex string (used as database entry)
* @param string $tex the user input hash
*/
- private static function updateIndex($revId, $eid, $inputHash, $tex){
+ private static function updateIndex( $revId, $eid, $inputHash, $tex ) {
try {
$dbr = wfGetDB( DB_SLAVE );
$exists = $dbr->selectRow( 'mathindex',
@@ -89,15 +91,17 @@
array(
'mathindex_revision_id' => $revId,
'mathindex_anchor' => $eid,
- 'mathindex_inputhash' => $inputHash)
- ) ;
+ 'mathindex_inputhash' => $inputHash
+ )
+ );
if ( $exists ) {
- wfDebugLog( 'MathSearch', 'Index $' . $tex . '$
already in database.');
- wfDebugLog( 'MathSearch', "$revId-$eid with
hash ". bin2hex($inputHash) );
+ wfDebugLog( 'MathSearch', 'Index $' . $tex . '$
already in database.' );
+ wfDebugLog( 'MathSearch', "$revId-$eid with
hash " . bin2hex( $inputHash ) );
} else {
self::writeMathIndex( $revId, $eid, $inputHash,
$tex );
}
- } catch ( Exception $e ) {
+ }
+ catch ( Exception $e ) {
wfDebugLog( "MathSearch", 'Problem writing to math
index!'
. ' You might want the rebuild the index by
running:'
. '"php
extensions/MathSearch/ReRenderMath.php". The error is'
@@ -117,15 +121,15 @@
* @return bool true if an ID has been assigned manually,
* false if the automatic fallback math{$id} was used.
*/
- public static function setMathId( &$id, MathRenderer $renderer, $revId)
{
- if ( $renderer->getID() ){
+ public static function setMathId( &$id, MathRenderer $renderer, $revId
) {
+ if ( $renderer->getID() ) {
$id = $renderer->getID();
return true;
} else {
- if ( is_null( $id ) ){
- $id = self::$nextID++;
+ if ( is_null( $id ) ) {
+ $id = self::$nextID ++;
$id = self::generateMathAnchorString( $revId,
$id, '' );
- $renderer->setID($id);
+ $renderer->setID( $id );
}
return false;
}
@@ -143,10 +147,12 @@
if ( $revId > 0 ) { // Only store something if a pageid was set.
// Use manually assigned IDs whenever possible
// and fallback to automatic IDs otherwise.
- if ( self::setMathId( $eid , $renderer, $revId ) ===
false ){
- $Result = preg_replace(
'/(class="mwe-math-mathml-(inline|display))/', "id=\"$eid\" \\1", $Result );
+ if ( self::setMathId( $eid, $renderer, $revId ) ===
false ) {
+ $Result =
+ preg_replace(
'/(class="mwe-math-mathml-(inline|display))/', "id=\"$eid\" \\1",
+ $Result );
}
- self::updateIndex( $revId , $eid ,
$renderer->getInputHash() , $renderer->getTex() );
+ self::updateIndex( $revId, $eid,
$renderer->getInputHash(), $renderer->getTex() );
}
return true;
}
@@ -158,13 +164,15 @@
* @param string|null $Result reference to the rendering result
* @return bool
*/
- static function addIdentifierDescription( Parser $parser, MathRenderer
$renderer, &$Result = null ) {
+ static function addIdentifierDescription( Parser $parser, MathRenderer
$renderer,
+ &$Result = null ) {
$revId = $parser->getRevisionId();
- self::setMathId( $eid , $renderer, $revId );
- $mo = MathObject::cloneFromRenderer($renderer);
- $mo->setRevisionID($revId);
- $mo->setID($eid);
- $Result = preg_replace_callback("#<(mi|mo)(
([^>].*?))?>(.*?)</\\1>#u", array( $mo , 'addIdentifierTitle' ), $Result);
+ self::setMathId( $eid, $renderer, $revId );
+ $mo = MathObject::cloneFromRenderer( $renderer );
+ $mo->setRevisionID( $revId );
+ $mo->setID( $eid );
+ $Result = preg_replace_callback( "#<(mi|mo)(
([^>].*?))?>(.*?)</\\1>#u",
+ array( $mo, 'addIdentifierTitle' ), $Result );
return true;
}
@@ -175,10 +183,14 @@
* @param string|null $Result reference to the rendering result
* @return bool
*/
- static function addLinkToFormulaInfoPage( Parser $parser, MathRenderer
$renderer, &$Result = null ) {
+ static function addLinkToFormulaInfoPage( Parser $parser, MathRenderer
$renderer,
+ &$Result = null ) {
$revId = $parser->getRevisionId();
- self::setMathId( $eid , $renderer, $revId );
- $url = SpecialPage::getTitleFor( 'FormulaInfo' )->getLocalUrl(
array( 'pid' => $revId, 'eid' => $eid ) );
+ self::setMathId( $eid, $renderer, $revId );
+ $url = SpecialPage::getTitleFor( 'FormulaInfo' )->getLocalUrl(
array(
+ 'pid' => $revId,
+ 'eid' => $eid
+ ) );
$Result = "<span><a href=\"$url\" id=\"$eid\"
style=\"color:inherit;\">$Result</a></span>";
return true;
}
@@ -195,9 +207,10 @@
* @param null $Result
* @return bool
*/
- static function onMathFormulaRenderedNoLink( Parser $parser,
MathRenderer $renderer, &$Result = null ) {
+ static function onMathFormulaRenderedNoLink( Parser $parser,
MathRenderer $renderer,
+ &$Result = null ) {
$revId = $parser->getRevisionId();
- self::setMathId($eid, $renderer, $revId);
+ self::setMathId( $eid, $renderer, $revId );
if ( $revId > 0 ) { // Only store something if a pageid was set.
self::updateIndex( $revId, $eid,
$renderer->getInputHash(), $renderer->getTex() );
}
@@ -221,15 +234,16 @@
return true;
}
- static function generateMathAnchorString($revId, $anchorID, $prefix =
"#"){
+ static function generateMathAnchorString( $revId, $anchorID, $prefix =
"#" ) {
$result = "{$prefix}math.$revId.$anchorID";
- Hooks::run( "MathSearchGenerateAnchorString" , array( $revId,
$anchorID, $prefix, &$result ) );
+ Hooks::run( "MathSearchGenerateAnchorString",
+ array( $revId, $anchorID, $prefix, &$result ) );
return $result;
}
/**
- * @param int $oldID
- * @param int $eid
+ * @param int $oldID
+ * @param int $eid
* @param string $inputHash
* @param string $tex
*/
@@ -238,10 +252,10 @@
$dbw = wfGetDB( DB_MASTER );
$dbw->onTransactionIdle( function () use ( $oldID, $eid,
$inputHash, $dbw ) {
$dbw->replace( 'mathindex', array(
'mathindex_revision_id', 'mathindex_anchor' ), array(
- 'mathindex_revision_id' => $oldID,
- 'mathindex_anchor' => $eid,
- 'mathindex_inputhash' => $inputHash
- ) );
+ 'mathindex_revision_id' => $oldID,
+ 'mathindex_anchor' => $eid,
+ 'mathindex_inputhash' => $inputHash
+ ) );
} );
}
@@ -253,7 +267,7 @@
*/
static function onParserFirstCallInit( $parser ) {
$parser->setHook( 'mquery', array( 'MathSearchHooks',
'mQueryTagHook' ) );
- wfDebugLog('MathSearch','mquery tag registered');
+ wfDebugLog( 'MathSearch', 'mquery tag registered' );
return true;
}
@@ -270,18 +284,87 @@
if ( trim( $content ) === '' ) { // bug 8372
return '';
}
- wfDebugLog('MathSearch','Render mquery tag.');
+ wfDebugLog( 'MathSearch', 'Render mquery tag.' );
//TODO: Report %\n problem to LaTeXML upstream
$content = preg_replace( '/%\n/', '', $content );
$renderer = new MathLaTeXML( $content );
- $mQuerySettings = $wgMathDefaultLaTeXMLSetting;
+ $mQuerySettings = $wgMathDefaultLaTeXMLSetting;
$mQuerySettings['preload'][] = 'mws.sty';
- $renderer->setLaTeXMLSettings($mQuerySettings);
- $renderer->render( );
+ $renderer->setLaTeXMLSettings( $mQuerySettings );
+ $renderer->render();
$renderedMath = $renderer->getHtmlOutput();
$renderer->writeCache();
return array( $renderedMath, "markerType" => 'nowiki' );
}
+ static function onArticleDeleteComplete() {
+
+ }
+
+
+
+ /**
+ * Occurs after the save page request has been processed.
+ * @see
https://www.mediawiki.org/wiki/Manual:Hooks/PageContentSaveComplete
+ *
+ * @param WikiPage $article
+ * @param User $user
+ * @param Content $content
+ * @param string $summary
+ * @param boolean $isMinor
+ * @param boolean $isWatch
+ * @param $section Deprecated
+ * @param integer $flags
+ * @param Revision|null $revision
+ * @param Status $status
+ * @param integer $baseRevId
+ *
+ * @return boolean
+ */
+ public static function onPageContentSaveComplete( $article, $user,
$content, $summary, $isMinor,
+ $isWatch, $section, $flags, $revision, $status, $baseRevId ) {
+ //TODO: Update to JOB
+ if ( $revision == null ) {
+ wfDebugLog( 'MathSearch', "Empty update for
{$article->getTitle()->getFullText()}." );
+ return true;
+ }
+ $mathTags =
+ MathObject::extractMathTagsFromWikiText(
ContentHandler::getContentText( $content ) );
+ $revId = $revision->getId();
+ $harvest = "";
+ if ( $mathTags ) {
+ $dw = new MwsDumpWriter();
+ self::resetId();
+ foreach ( $mathTags as $tag ) {
+ $id = null;
+ $tagContent = $tag[1];
+ $attributes = $tag[2];
+ // $fullElement = $tag[3];
+ $renderer = MathRenderer::getRenderer(
$tagContent, $attributes, MW_MATH_LATEXML );
+ $renderer->render();
+ self::setMathId( $id, $renderer, $revId );
+ $dw->addMwsExpression( $renderer->getMathml(),
$revId, $id );
+ }
+ $harvest = $dw->getOutput();
+ }
+ //TODO: Figure out whats wrong with
$revision->getPrevious()->getId();
+ $prevRevId =
+ Revision::newFromId(
$article->getTitle()->getLatestRevID() )->getPrevious()->getId();
+ if ( $prevRevId != null ) {
+ $baseXUpdater = new MathEngineBaseX();
+ $res = $baseXUpdater->update( $harvest, array(
$prevRevId ) );
+ }
+ if ( $res ) {
+ wfDebugLog( 'MathSearch', "Update for $revId (was
$prevRevId) successful." );
+ } else {
+ wfDebugLog( 'MathSearch', "Update for $revId (was
$prevRevId , $baseRevId) failed." );
+ }
+
+ return true;
+ }
+
+ static function resetId() {
+ self::$nextID = 0;
+ }
}
diff --git a/MathSearch.php b/MathSearch.php
index 5a82ede..5fbbafc 100644
--- a/MathSearch.php
+++ b/MathSearch.php
@@ -48,6 +48,7 @@
$wgAutoloadClasses['ImportCsv'] = __DIR__ . '/includes/ImportCsv.php';
$wgAutoloadClasses['MathSearchUtils'] = __DIR__ .
'/includes/MathSearchUtils.php';
$wgAutoloadClasses['MathSearchTerm'] = __DIR__ .
'/includes/MathSearchTerm.php';
+$wgAutoloadClasses['MwsDumpWriter'] = __DIR__ . '/includes/MwsDumpWriter.php';
$wgMessagesDirs['MathSeach'] = __DIR__ . '/i18n';
$wgExtensionMessagesFiles['MathSearch'] = __DIR__ . '/MathSearch.i18n.php';
@@ -73,6 +74,8 @@
$wgHooks['MathFormulaPostRender']['addLink'] =
'MathSearchHooks::addLinkToFormulaInfoPage';
$wgHooks['UnitTestsList'][] = 'MathSearchHooks::onRegisterUnitTests';
$wgHooks['ParserFirstCallInit'][] = 'MathSearchHooks::onParserFirstCallInit';
+$wgHooks['ArticleDeleteComplete'][] =
'MathSearchHooks::onArticleDeleteComplete';
+$wgHooks['PageContentSaveComplete'][] =
'MathSearchHooks::onPageContentSaveComplete';
$wgMathSearchBaseXBackendUrl = 'http://localhost:10043/';
diff --git a/includes/MwsDumpWriter.php b/includes/MwsDumpWriter.php
new file mode 100644
index 0000000..5595e15
--- /dev/null
+++ b/includes/MwsDumpWriter.php
@@ -0,0 +1,77 @@
+<?php
+
+class MwsDumpWriter {
+ private $mwsns = 'mws:';
+ private $XMLHead;
+ private $XMLFooter;
+ private $outBuffer = '';
+
+ public function __construct() {
+ $this->InitializeHeader();
+ }
+
+ public function InitializeHeader( ) {
+ $ns = $this->mwsns;
+ $this->XMLHead = "<?xml version=\"1.0\"?>\n<" . $ns .
+ 'harvest xmlns:mws="http://search.mathweb.org/ns";
xmlns:m="http://www.w3.org/1998/Math/MathML";>';
+ $this->XMLFooter = '</' . $ns . 'harvest>';
+ }
+
+ /**
+ * @param stdClass $row
+ *
+ * @return string
+ */
+ public function generateIndexString( $row ) {
+ $xml = simplexml_load_string( utf8_decode( $row->math_mathml )
);
+ if ( !$xml ) {
+ echo "ERROR while converting:\n " . var_export(
$row->math_mathml, true ) . "\n";
+ foreach ( libxml_get_errors() as $error ) {
+ echo "\t", $error->message;
+ }
+ libxml_clear_errors();
+ return '';
+ }
+ return $this->getMwsExpression( utf8_decode( $row->math_mathml
),
+ $row->mathindex_revision_id, $row->mathindex_anchor );
+
+ }
+
+ public function getHead() {
+ return $this->XMLHead;
+ }
+
+ public function getFooter() {
+ return $this->XMLFooter;
+ }
+
+ public function getMwsExpression( $mathML, $revId, $eId ) {
+ $out = "\n<" . $this->mwsns . "expr
url=\"${revId}#${eId}\">\n\t";
+ $out .= $mathML;
+ $out .= "\n</" . $this->mwsns . "expr>\n";
+ return $out;
+ }
+ public function addMwsExpression( $mathML, $revId, $eId ) {
+ $this->outBuffer .= $this->getMwsExpression( $mathML, $revId,
$eId );
+ return true;
+ }
+
+
+ /**
+ * @return string
+ */
+ public function getMwsns() {
+ return $this->mwsns;
+ }
+
+ /**
+ * @param string $mwsns
+ */
+ public function setMwsns( $mwsns ) {
+ $this->mwsns = $mwsns;
+ }
+
+ public function getOutput(){
+ return $this->getHead() . $this->outBuffer . $this->getFooter();
+ }
+}
\ No newline at end of file
diff --git a/includes/engines/MathEngineBaseX.php
b/includes/engines/MathEngineBaseX.php
index 47fc0d6..3c1be83 100644
--- a/includes/engines/MathEngineBaseX.php
+++ b/includes/engines/MathEngineBaseX.php
@@ -51,9 +51,13 @@
function processMathResults( $xmlRoot ) {
foreach ( $xmlRoot->children( )->children() as $page ) {
$attrs = $page->attributes();
- $uri = explode( ".", $attrs["id"] );
- $revisionID = $uri[1];
- $AnchorID = $uri[2];
+ $uri = explode( "#", $attrs["id"] );
+ if ( sizeof($uri) != 2 ) {
+ wfDebugLog('MathSearch','Can not parse'.
$attrs['id']);
+ continue;
+ }
+ $revisionID = $uri[0];
+ $AnchorID = $uri[1];
$this->relevanceMap[] = $revisionID;
$substarr = array();
//TODO: Add hit support.
@@ -69,4 +73,19 @@
function getPostData( $numProcess ){
return json_encode( array( "type" => $this->type, "query" =>
$this->query->getCQuery()) );
}
+
+ function update( $harvest = "", array $delte=array() ){
+ global $wgMathSearchBaseXBackendUrl;
+ $json_payload = json_encode( array( "harvest" => $harvest,
"delete" => $delte) );
+ $res = self::doPost( $wgMathSearchBaseXBackendUrl.
'api/update', $json_payload);
+ if($res){
+ $resJson = json_decode($res);
+ if ($resJson->success==true){
+ return true;
+ } else {
+ wfDebugLog("MathSearch", "harvest update
failed" . var_export($resJson,true));
+ }
+ }
+ return false;
+ }
}
\ No newline at end of file
diff --git a/includes/engines/MathEngineRest.php
b/includes/engines/MathEngineRest.php
index 1b4d3b1..4b8a2b5 100644
--- a/includes/engines/MathEngineRest.php
+++ b/includes/engines/MathEngineRest.php
@@ -19,6 +19,29 @@
/** @type string */
protected $backendUrl = "http://localhost:9090";;
+ protected static function doPost( $url, $postData ) {
+ $res = Http::post( $url, array( "postData" => $postData,
"timeout" => 60 ) );
+ if ( $res === false ) {
+ if ( function_exists( 'curl_init' ) ) {
+ $handle = curl_init();
+ $options = array(
+ CURLOPT_URL => $url,
+ CURLOPT_CUSTOMREQUEST => 'POST', // GET
POST PUT PATCH DELETE HEAD OPTIONS
+ );
+ // TODO: Figure out how not to write the error
in a message and not in top of the output page
+ curl_setopt_array( $handle, $options );
+ $details = curl_exec( $handle );
+ } else {
+ $details = "curl is not installed.";
+ }
+ wfDebugLog( "MathSearch", "Nothing retreived from $url.
Check if server is running. Error:" .
+ var_export( $details, true ) );
+ return false;
+ } else {
+ return $res;
+ }
+ }
+
/**
* @return string
*/
@@ -80,25 +103,12 @@
global $wgMathDebug;
$numProcess = 30000;
$postData = $this->getPostData( $numProcess );
- $res = Http::post( $this->backendUrl, array( "postData" =>
$postData, "timeout" => 60 ) );
- if ( $res === false ) {
- if ( function_exists( 'curl_init' ) ) {
- $handle = curl_init();
- $options = array(
- CURLOPT_URL => $this->backendUrl,
- CURLOPT_CUSTOMREQUEST => 'POST', // GET
POST PUT PATCH DELETE HEAD OPTIONS
- );
- // TODO: Figure out how not to write the error
in a message and not in top of the output page
- curl_setopt_array( $handle, $options );
- $details = curl_exec( $handle );
- } else {
- $details = "curl is not installed.";
- }
- wfDebugLog( "MathSearch", "Nothing retreived from
$this->backendUrl. Check if mwsd is running. Error:" .
- var_export( $details, true ) );
+ $res = self::doPost($this->backendUrl,$postData);
+ if ( $res === false ){
return false;
+ } else {
+ return $this->processResults( $res, $numProcess );
}
- return $this->processResults( $res, $numProcess );
}
/**
diff --git a/maintenance/CreateMWSHarvest.php b/maintenance/CreateMWSHarvest.php
index ff54580..ec30e7e 100644
--- a/maintenance/CreateMWSHarvest.php
+++ b/maintenance/CreateMWSHarvest.php
@@ -41,6 +41,13 @@
$this->addOption( 'mwsns', 'The namespace or mws normally
"mws:"', false );
}
+ public function InitializeHeader() {
+ self::$XMLHead =
+ "<?xml version=\"1.0\"?>\n<" . self::$mwsns .
+ 'harvest xmlns:mws="http://search.mathweb.org/ns";
xmlns:m="http://www.w3.org/1998/Math/MathML";>';
+ self::$XMLFooter = '</' . self::$mwsns . 'harvest>';
+ }
+
/**
* @param stdClass $row
*
@@ -85,10 +92,7 @@
*/
public function execute() {
self::$mwsns = $this->getOption( 'mwsns', '' );
- self::$XMLHead =
- "<?xml version=\"1.0\"?>\n<" . self::$mwsns .
- 'harvest xmlns:mws="http://search.mathweb.org/ns";
xmlns:m="http://www.w3.org/1998/Math/MathML";>';
- self::$XMLFooter = '</' . self::$mwsns . 'harvest>';
+ $this->InitializeHeader();
parent::execute();
}
}
diff --git a/maintenance/UpdateMath.php b/maintenance/UpdateMath.php
index 43d04d9..2f1dada 100644
--- a/maintenance/UpdateMath.php
+++ b/maintenance/UpdateMath.php
@@ -25,7 +25,7 @@
* Class UpdateMath
*/
class UpdateMath extends Maintenance {
- const RTI_CHUNK_SIZE = 100;
+ const RTI_CHUNK_SIZE = 10000;
public $purge = false;
/** @var boolean */
private $verbose;
diff --git a/tests/MwsDumpWriterTest.php b/tests/MwsDumpWriterTest.php
new file mode 100644
index 0000000..e36052b
--- /dev/null
+++ b/tests/MwsDumpWriterTest.php
@@ -0,0 +1,189 @@
+<?php
+/**
+ * Test the MathSearchUtils script.
+ *
+ * @group MathSearch
+ */
+class MwsDumpWriterTest extends MediaWikiTestCase {
+ //TODO: update tests strategy resources etc.
+ private $testWikiText = <<<'WikiText'
+<math>
+E = m \power c 1
+</math>
+
+<math>
+E = m \power c 2
+</math>
+
+<math id=incorrect>
+E = m \power c 3
+</math>
+
+<math>
+E = m \power c 4
+</math>
+WikiText;
+
+ private $expectedOutput = <<<'XML'
+<?xml version="1.0"?>
+<mws:harvest xmlns:mws="http://search.mathweb.org/ns";
xmlns:m="http://www.w3.org/1998/Math/MathML";>
+<mws:expr url="28351#math.28351.0">
+ <math id="p1.1.m1.1" class="ltx_Math" alttext="{\displaystyle
E=m{c^{1}}}" display="inline">
+ <semantics id="p1.1.m1.1a">
+ <mrow id="p1.1.m1.1.5" xref="p1.1.m1.1.5.cmml">
+ <mi id="p1.1.m1.1.1" xref="p1.1.m1.1.1.cmml">E</mi>
+ <mo id="p1.1.m1.1.2" xref="p1.1.m1.1.2.cmml">=</mo>
+ <mrow id="p1.1.m1.1.5.1" xref="p1.1.m1.1.5.1.cmml">
+ <mi id="p1.1.m1.1.3" xref="p1.1.m1.1.3.cmml">m</mi>
+ <mo id="p1.1.m1.1.5.1.1" xref="p1.1.m1.1.5.1.1.cmml"></mo>
+ <msup id="p1.1.m1.1.4.4" xref="p1.1.m1.1.4.1.cmml">
+ <mi id="p1.1.m1.1.4.2" xref="p1.1.m1.1.4.2.cmml">c</mi>
+ <mn id="p1.1.m1.1.4.3.1" xref="p1.1.m1.1.4.3.1.cmml">1</mn>
+ </msup>
+ </mrow>
+ </mrow>
+ <annotation-xml encoding="MathML-Content" id="p1.1.m1.1b">
+ <apply id="p1.1.m1.1.5.cmml" xref="p1.1.m1.1.5">
+ <eq id="p1.1.m1.1.2.cmml" xref="p1.1.m1.1.2"></eq>
+ <ci id="p1.1.m1.1.1.cmml" xref="p1.1.m1.1.1">E</ci>
+ <apply id="p1.1.m1.1.5.1.cmml" xref="p1.1.m1.1.5.1">
+ <times id="p1.1.m1.1.5.1.1.cmml" xref="p1.1.m1.1.5.1.1"></times>
+ <ci id="p1.1.m1.1.3.cmml" xref="p1.1.m1.1.3">m</ci>
+ <apply id="p1.1.m1.1.4.1.cmml" xref="p1.1.m1.1.4.4">
+ <power id="p1.1.m1.1.4.1.1.cmml"></power>
+ <ci id="p1.1.m1.1.4.2.cmml" xref="p1.1.m1.1.4.2">c</ci>
+ <cn type="integer" id="p1.1.m1.1.4.3.1.cmml"
xref="p1.1.m1.1.4.3.1">1</cn>
+ </apply>
+ </apply>
+ </apply>
+ </annotation-xml>
+ <annotation encoding="application/x-tex" id="p1.1.m1.1c">{\displaystyle
E=m{c^{1}}}</annotation>
+ </semantics>
+</math>
+</mws:expr>
+
+<mws:expr url="28351#math.28351.1">
+ <math id="p1.1.m1.1" class="ltx_Math" alttext="{\displaystyle
E=m{c^{2}}}" display="inline">
+ <semantics id="p1.1.m1.1a">
+ <mrow id="p1.1.m1.1.5" xref="p1.1.m1.1.5.cmml">
+ <mi id="p1.1.m1.1.1" xref="p1.1.m1.1.1.cmml">E</mi>
+ <mo id="p1.1.m1.1.2" xref="p1.1.m1.1.2.cmml">=</mo>
+ <mrow id="p1.1.m1.1.5.1" xref="p1.1.m1.1.5.1.cmml">
+ <mi id="p1.1.m1.1.3" xref="p1.1.m1.1.3.cmml">m</mi>
+ <mo id="p1.1.m1.1.5.1.1" xref="p1.1.m1.1.5.1.1.cmml"></mo>
+ <msup id="p1.1.m1.1.4.4" xref="p1.1.m1.1.4.1.cmml">
+ <mi id="p1.1.m1.1.4.2" xref="p1.1.m1.1.4.2.cmml">c</mi>
+ <mn id="p1.1.m1.1.4.3.1" xref="p1.1.m1.1.4.3.1.cmml">2</mn>
+ </msup>
+ </mrow>
+ </mrow>
+ <annotation-xml encoding="MathML-Content" id="p1.1.m1.1b">
+ <apply id="p1.1.m1.1.5.cmml" xref="p1.1.m1.1.5">
+ <eq id="p1.1.m1.1.2.cmml" xref="p1.1.m1.1.2"></eq>
+ <ci id="p1.1.m1.1.1.cmml" xref="p1.1.m1.1.1">E</ci>
+ <apply id="p1.1.m1.1.5.1.cmml" xref="p1.1.m1.1.5.1">
+ <times id="p1.1.m1.1.5.1.1.cmml" xref="p1.1.m1.1.5.1.1"></times>
+ <ci id="p1.1.m1.1.3.cmml" xref="p1.1.m1.1.3">m</ci>
+ <apply id="p1.1.m1.1.4.1.cmml" xref="p1.1.m1.1.4.4">
+ <power id="p1.1.m1.1.4.1.1.cmml"></power>
+ <ci id="p1.1.m1.1.4.2.cmml" xref="p1.1.m1.1.4.2">c</ci>
+ <cn type="integer" id="p1.1.m1.1.4.3.1.cmml"
xref="p1.1.m1.1.4.3.1">2</cn>
+ </apply>
+ </apply>
+ </apply>
+ </annotation-xml>
+ <annotation encoding="application/x-tex" id="p1.1.m1.1c">{\displaystyle
E=m{c^{2}}}</annotation>
+ </semantics>
+</math>
+</mws:expr>
+
+<mws:expr url="28351#incorrect">
+ <math id="p1.1.m1.1" class="ltx_Math" alttext="{\displaystyle
E=m{c^{3}}}" display="inline">
+ <semantics id="p1.1.m1.1a">
+ <mrow id="p1.1.m1.1.5" xref="p1.1.m1.1.5.cmml">
+ <mi id="p1.1.m1.1.1" xref="p1.1.m1.1.1.cmml">E</mi>
+ <mo id="p1.1.m1.1.2" xref="p1.1.m1.1.2.cmml">=</mo>
+ <mrow id="p1.1.m1.1.5.1" xref="p1.1.m1.1.5.1.cmml">
+ <mi id="p1.1.m1.1.3" xref="p1.1.m1.1.3.cmml">m</mi>
+ <mo id="p1.1.m1.1.5.1.1" xref="p1.1.m1.1.5.1.1.cmml"></mo>
+ <msup id="p1.1.m1.1.4.4" xref="p1.1.m1.1.4.1.cmml">
+ <mi id="p1.1.m1.1.4.2" xref="p1.1.m1.1.4.2.cmml">c</mi>
+ <mn id="p1.1.m1.1.4.3.1" xref="p1.1.m1.1.4.3.1.cmml">3</mn>
+ </msup>
+ </mrow>
+ </mrow>
+ <annotation-xml encoding="MathML-Content" id="p1.1.m1.1b">
+ <apply id="p1.1.m1.1.5.cmml" xref="p1.1.m1.1.5">
+ <eq id="p1.1.m1.1.2.cmml" xref="p1.1.m1.1.2"></eq>
+ <ci id="p1.1.m1.1.1.cmml" xref="p1.1.m1.1.1">E</ci>
+ <apply id="p1.1.m1.1.5.1.cmml" xref="p1.1.m1.1.5.1">
+ <times id="p1.1.m1.1.5.1.1.cmml" xref="p1.1.m1.1.5.1.1"></times>
+ <ci id="p1.1.m1.1.3.cmml" xref="p1.1.m1.1.3">m</ci>
+ <apply id="p1.1.m1.1.4.1.cmml" xref="p1.1.m1.1.4.4">
+ <power id="p1.1.m1.1.4.1.1.cmml"></power>
+ <ci id="p1.1.m1.1.4.2.cmml" xref="p1.1.m1.1.4.2">c</ci>
+ <cn type="integer" id="p1.1.m1.1.4.3.1.cmml"
xref="p1.1.m1.1.4.3.1">3</cn>
+ </apply>
+ </apply>
+ </apply>
+ </annotation-xml>
+ <annotation encoding="application/x-tex" id="p1.1.m1.1c">{\displaystyle
E=m{c^{3}}}</annotation>
+ </semantics>
+</math>
+</mws:expr>
+
+<mws:expr url="28351#math.28351.2">
+ <math id="p1.1.m1.1" class="ltx_Math" alttext="{\displaystyle
E=m{c^{4}}}" display="inline">
+ <semantics id="p1.1.m1.1a">
+ <mrow id="p1.1.m1.1.5" xref="p1.1.m1.1.5.cmml">
+ <mi id="p1.1.m1.1.1" xref="p1.1.m1.1.1.cmml">E</mi>
+ <mo id="p1.1.m1.1.2" xref="p1.1.m1.1.2.cmml">=</mo>
+ <mrow id="p1.1.m1.1.5.1" xref="p1.1.m1.1.5.1.cmml">
+ <mi id="p1.1.m1.1.3" xref="p1.1.m1.1.3.cmml">m</mi>
+ <mo id="p1.1.m1.1.5.1.1" xref="p1.1.m1.1.5.1.1.cmml"></mo>
+ <msup id="p1.1.m1.1.4.4" xref="p1.1.m1.1.4.1.cmml">
+ <mi id="p1.1.m1.1.4.2" xref="p1.1.m1.1.4.2.cmml">c</mi>
+ <mn id="p1.1.m1.1.4.3.1" xref="p1.1.m1.1.4.3.1.cmml">4</mn>
+ </msup>
+ </mrow>
+ </mrow>
+ <annotation-xml encoding="MathML-Content" id="p1.1.m1.1b">
+ <apply id="p1.1.m1.1.5.cmml" xref="p1.1.m1.1.5">
+ <eq id="p1.1.m1.1.2.cmml" xref="p1.1.m1.1.2"></eq>
+ <ci id="p1.1.m1.1.1.cmml" xref="p1.1.m1.1.1">E</ci>
+ <apply id="p1.1.m1.1.5.1.cmml" xref="p1.1.m1.1.5.1">
+ <times id="p1.1.m1.1.5.1.1.cmml" xref="p1.1.m1.1.5.1.1"></times>
+ <ci id="p1.1.m1.1.3.cmml" xref="p1.1.m1.1.3">m</ci>
+ <apply id="p1.1.m1.1.4.1.cmml" xref="p1.1.m1.1.4.4">
+ <power id="p1.1.m1.1.4.1.1.cmml"></power>
+ <ci id="p1.1.m1.1.4.2.cmml" xref="p1.1.m1.1.4.2">c</ci>
+ <cn type="integer" id="p1.1.m1.1.4.3.1.cmml"
xref="p1.1.m1.1.4.3.1">4</cn>
+ </apply>
+ </apply>
+ </apply>
+ </annotation-xml>
+ <annotation encoding="application/x-tex" id="p1.1.m1.1c">{\displaystyle
E=m{c^{4}}}</annotation>
+ </semantics>
+</math>
+</mws:expr>
+</mws:harvest>
+XML;
+
+
+ public function testExtract() {
+ $mathTags = MathObject::extractMathTagsFromWikiText(
$this->testWikiText ) ;
+ $revId = 28351;
+ $dw = new MwsDumpWriter();
+ MathSearchHooks::resetId();
+ foreach ( $mathTags as $tag ) {
+ $id = null;
+ $content = $tag[1];
+ $attributes = $tag[2];
+ $renderer = MathRenderer::getRenderer( $content,
$attributes, MW_MATH_LATEXML );
+ $renderer->render();
+ MathSearchHooks::setMathId( $id, $renderer, $revId );
+ $dw->addMwsExpression( $renderer->getMathml(), $revId,
$id );
+ }
+ $this->assertEquals( $this->expectedOutput, $dw->getOutput() );
+ }
+}
\ No newline at end of file
--
To view, visit https://gerrit.wikimedia.org/r/200100
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: newchange
Gerrit-Change-Id: Iff6e07e2276f32c11f4b828e55ba9e27bd33aa5d
Gerrit-PatchSet: 1
Gerrit-Project: mediawiki/extensions/MathSearch
Gerrit-Branch: master
Gerrit-Owner: Physikerwelt <w...@physikerwelt.de>
_______________________________________________
MediaWiki-commits mailing list
MediaWiki-commits@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
