Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 184 |
|
0.00% |
0 / 9 |
CRAP | |
0.00% |
0 / 1 |
| UpdateMath | |
0.00% |
0 / 179 |
|
0.00% |
0 / 9 |
1260 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 22 |
|
0.00% |
0 / 1 |
2 | |||
| time | |
0.00% |
0 / 14 |
|
0.00% |
0 / 1 |
12 | |||
| populateSearchIndex | |
0.00% |
0 / 36 |
|
0.00% |
0 / 1 |
42 | |||
| doUpdate | |
0.00% |
0 / 63 |
|
0.00% |
0 / 1 |
240 | |||
| getParserOptions | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
6 | |||
| getParser | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
6 | |||
| execute | |
0.00% |
0 / 15 |
|
0.00% |
0 / 1 |
2 | |||
| getMathMLForExport | |
0.00% |
0 / 12 |
|
0.00% |
0 / 1 |
12 | |||
| exportMMLtoFile | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
6 | |||
| 1 | #!/usr/bin/env php |
| 2 | <?php |
| 3 | /** |
| 4 | * This program is free software; you can redistribute it and/or modify |
| 5 | * it under the terms of the GNU General Public License as published by |
| 6 | * the Free Software Foundation; either version 2 of the License, or |
| 7 | * (at your option) any later version. |
| 8 | * |
| 9 | * This program is distributed in the hope that it will be useful, |
| 10 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 11 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 12 | * GNU General Public License for more details. |
| 13 | * |
| 14 | * You should have received a copy of the GNU General Public License along |
| 15 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 16 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 17 | * http://www.gnu.org/copyleft/gpl.html |
| 18 | * |
| 19 | * @ingroup Maintenance |
| 20 | */ |
| 21 | |
| 22 | use MediaWiki\Extension\Math\MathRenderer; |
| 23 | use MediaWiki\MediaWikiServices; |
| 24 | |
| 25 | require_once __DIR__ . '/../../../maintenance/Maintenance.php'; |
| 26 | |
| 27 | class UpdateMath extends Maintenance { |
| 28 | |
| 29 | /** @var bool */ |
| 30 | private $purge = false; |
| 31 | /** @var bool */ |
| 32 | private $verbose; |
| 33 | /** @var \Wikimedia\Rdbms\IDatabase */ |
| 34 | private $dbw; |
| 35 | /** @var \Wikimedia\Rdbms\IDatabase */ |
| 36 | private $db; |
| 37 | /** @var MathRenderer */ |
| 38 | private $current; |
| 39 | /** @var float */ |
| 40 | private $time = 0.0; // microtime( true ); |
| 41 | /** @var float[] */ |
| 42 | private $performance = []; |
| 43 | /** @var string */ |
| 44 | private $renderingMode = 'latexml'; |
| 45 | /** @var int */ |
| 46 | private $chunkSize = 1000; |
| 47 | private $parser; |
| 48 | private $parserOptions; |
| 49 | |
| 50 | public function __construct() { |
| 51 | parent::__construct(); |
| 52 | $this->addDescription( 'Updates the index of Mathematical formulae.' ); |
| 53 | $this->addOption( 'purge', |
| 54 | "If set all formulae are rendered again without using caches. (Very time consuming!)", |
| 55 | false, false, "f" ); |
| 56 | $this->addArg( 'min', "If set processing is started at the page with rank(pageID)>min", |
| 57 | false ); |
| 58 | $this->addArg( 'max', "If set processing is stopped at the page with rank(pageID)<=max", |
| 59 | false ); |
| 60 | $this->addOption( 'verbose', "If set output for successful rendering will produced", false, |
| 61 | false, 'v' ); |
| 62 | $this->addOption( 'SVG', "If set SVG images will be produced", false, false ); |
| 63 | $this->addOption( 'hooks', "If set hooks will be skipped, but index will be updated.", |
| 64 | false, false ); |
| 65 | $this->addOption( 'texvccheck', "If set texvccheck will be skipped", false, false ); |
| 66 | $this->addOption( 'mode', 'Rendering mode to be used (mathml, latexml)', false, true, |
| 67 | 'm' ); |
| 68 | $this->addOption( 'exportmml', 'export LaTeX and generated MathML to the specified file', false, true, |
| 69 | 'e' ); |
| 70 | $this->addOption( 'chunk-size', |
| 71 | 'Determines how many pages are updated in one database transaction.', false, true ); |
| 72 | $this->requireExtension( 'MathSearch' ); |
| 73 | } |
| 74 | |
| 75 | /** |
| 76 | * Measures time in ms. |
| 77 | * In order to have a formula centric evaluation, we can not just the build in profiler |
| 78 | * @param string $category |
| 79 | * |
| 80 | * @return int |
| 81 | */ |
| 82 | private function time( $category = 'default' ) { |
| 83 | global $wgMathDebug; |
| 84 | $delta = ( microtime( true ) - $this->time ) * 1000; |
| 85 | if ( isset( $this->performance[$category] ) ) { |
| 86 | $this->performance[$category] += $delta; |
| 87 | } else { |
| 88 | $this->performance[$category] = $delta; |
| 89 | } |
| 90 | if ( $wgMathDebug ) { |
| 91 | $this->db->insert( 'mathperformance', [ |
| 92 | 'math_inputhash' => $this->current->getInputHash(), |
| 93 | 'mathperformance_name' => substr( $category, 0, 10 ), |
| 94 | 'mathperformance_time' => $delta, |
| 95 | 'mathperformance_mode' => MathObject::MODE_2_USER_OPTION[ $this->renderingMode ] |
| 96 | ] ); |
| 97 | } |
| 98 | $this->time = microtime( true ); |
| 99 | |
| 100 | return (int)$delta; |
| 101 | } |
| 102 | |
| 103 | /** |
| 104 | * Populates the search index with content from all pages |
| 105 | * |
| 106 | * @param int $n |
| 107 | * @param int $cMax |
| 108 | */ |
| 109 | protected function populateSearchIndex( $n = 0, $cMax = -1 ) { |
| 110 | $s = $this->db->selectRow( 'revision', 'MAX(rev_id) AS count', '' ); |
| 111 | $count = $s->count; |
| 112 | if ( $cMax > 0 && $count > $cMax ) { |
| 113 | $count = $cMax; |
| 114 | } |
| 115 | $this->output( |
| 116 | "Rebuilding index fields for pages with revision < {$count} with option {$this->purge}...\n" |
| 117 | ); |
| 118 | $fCount = 0; |
| 119 | // return; |
| 120 | while ( $n < $count ) { |
| 121 | if ( $n ) { |
| 122 | $this->output( $n . " of $count \n" ); |
| 123 | } |
| 124 | $end = min( $n + $this->chunkSize - 1, $count ); |
| 125 | |
| 126 | # For filtering page by namespace add condition 'page_namespace = 4' |
| 127 | $res = $this->db->select( [ 'page', 'slots', 'content', 'text', 'revision' ], |
| 128 | [ 'page_id', 'page_namespace', 'page_title', 'page_latest', |
| 129 | 'content_address', 'old_text', 'old_flags', 'rev_id' ], |
| 130 | [ "rev_id BETWEEN $n AND $end" ], |
| 131 | __METHOD__, |
| 132 | [], |
| 133 | [ |
| 134 | 'slots' => [ 'INNER JOIN', [ 'slot_origin = page_latest' ] ], |
| 135 | 'content' => [ 'INNER JOIN', [ 'content_id = slot_content_id' ] ], |
| 136 | 'text' => [ 'INNER JOIN', [ 'old_id = substr(content_address,4)' ] ], |
| 137 | 'revision' => [ 'INNER JOIN', [ 'page_latest = rev_id' ] ] ] |
| 138 | ); |
| 139 | |
| 140 | $this->dbw->begin( __METHOD__ ); |
| 141 | $revisionStore = MediaWikiServices::getInstance()->getRevisionStore(); |
| 142 | // echo "before" +$this->dbw->selectField('mathindex', 'count(*)')."\n"; |
| 143 | foreach ( $res as $s ) { |
| 144 | $this->output( "\nr{$s->rev_id} namespace: {$s->page_namespace} page title: {$s->page_title}" ); |
| 145 | $fCount += $this->doUpdate( $s->page_id, $s->old_text, $s->page_title, $s->rev_id ); |
| 146 | } |
| 147 | // echo "before" +$this->dbw->selectField('mathindex', 'count(*)')."\n"; |
| 148 | $start = microtime( true ); |
| 149 | $this->dbw->commit( __METHOD__ ); |
| 150 | echo " committed in " . ( microtime( true ) - $start ) . "s\n\n"; |
| 151 | var_dump( $this->performance ); |
| 152 | // echo "after" +$this->dbw->selectField('mathindex', 'count(*)')."\n"; |
| 153 | $n += $this->chunkSize; |
| 154 | } |
| 155 | $this->output( "Updated {$fCount} formulae!\n" ); |
| 156 | } |
| 157 | |
| 158 | /** |
| 159 | * @param int $pid |
| 160 | * @param string $pText |
| 161 | * @param string $pTitle |
| 162 | * @param int $revId |
| 163 | * |
| 164 | * @return number |
| 165 | */ |
| 166 | private function doUpdate( $pid, $pText, $pTitle = "", $revId = 0 ) { |
| 167 | $allFormula = []; |
| 168 | |
| 169 | $notused = ''; |
| 170 | // MathSearchHooks::setNextID($eId); |
| 171 | $math = MathObject::extractMathTagsFromWikiText( $pText ); |
| 172 | $matches = count( $math ); |
| 173 | if ( $matches ) { |
| 174 | echo ( "\t processing $matches math fields for {$pTitle} page\n" ); |
| 175 | foreach ( $math as $formula ) { |
| 176 | $this->time = microtime( true ); |
| 177 | /** @var MathRenderer $renderer */ |
| 178 | $renderer = MediaWikiServices::getInstance()->get( 'Math.RendererFactory' ) |
| 179 | ->getRenderer( $formula[1], $formula[2], $this->renderingMode ); |
| 180 | $this->current = $renderer; |
| 181 | $this->time( "loadClass" ); |
| 182 | if ( $this->getOption( "texvccheck", false ) ) { |
| 183 | $checked = true; |
| 184 | } else { |
| 185 | $checked = $renderer->checkTeX(); |
| 186 | $this->time( "checkTex" ); |
| 187 | } |
| 188 | if ( $checked ) { |
| 189 | if ( !$renderer->isInDatabase() || $this->purge ) { |
| 190 | $renderer->render( $this->purge ); |
| 191 | if ( $renderer->getMathml() ) { |
| 192 | $this->time( "render" ); |
| 193 | } else { |
| 194 | $this->time( "Failing" ); |
| 195 | } |
| 196 | if ( $this->getOption( "SVG", false ) ) { |
| 197 | $svg = $renderer->getSvg(); |
| 198 | if ( $svg ) { |
| 199 | $this->time( "SVG-Rendering" ); |
| 200 | } else { |
| 201 | $this->time( "SVG-Fail" ); |
| 202 | } |
| 203 | } |
| 204 | } else { |
| 205 | $this->time( 'checkInDB' ); |
| 206 | } |
| 207 | } else { |
| 208 | $this->time( "checkTex-Fail" ); |
| 209 | echo "\nF:\t\t" . $renderer->getInputHash() . " texvccheck error:" . |
| 210 | $renderer->getLastError(); |
| 211 | continue; |
| 212 | } |
| 213 | $renderer->writeCache(); |
| 214 | $this->time( "write Cache" ); |
| 215 | if ( !$this->getOption( "hooks", false ) ) { |
| 216 | $hookContainer = MediaWikiServices::getInstance()->getHookContainer(); |
| 217 | $hookContainer->run( |
| 218 | 'MathFormulaPostRender', |
| 219 | [ |
| 220 | $this->getParser( $revId ), |
| 221 | &$renderer, |
| 222 | &$notused |
| 223 | ] |
| 224 | ); |
| 225 | $this->time( "hooks" ); |
| 226 | } else { |
| 227 | $eId = null; |
| 228 | MathSearchHooks::setMathId( $eId, $renderer, $revId ); |
| 229 | MathSearchHooks::writeMathIndex( $revId, $eId, $renderer->getInputHash(), '' ); |
| 230 | $this->time( "index" ); |
| 231 | } |
| 232 | if ( $renderer->getLastError() ) { |
| 233 | echo "\n\t\t" . $renderer->getLastError(); |
| 234 | echo "\nF:\t\t" . $renderer->getInputHash() . " equation " . ( $eId ) . |
| 235 | "-failed beginning with\n\t\t'" . substr( $formula, 0, 100 ) |
| 236 | . "'\n\t\tmathml:" . substr( $renderer->getMathml(), 0, 10 ) . "\n "; |
| 237 | } else { |
| 238 | if ( $this->verbose ) { |
| 239 | echo "\nS:\t\t" . $renderer->getInputHash(); |
| 240 | } |
| 241 | } |
| 242 | if ( $this->getOption( "exportmml", false ) ) { |
| 243 | $allFormula = $this->getMathMLForExport( $formula[1], $renderer, $allFormula ); |
| 244 | } |
| 245 | } |
| 246 | $mmlPath = $this->getOption( "exportmml", false ); |
| 247 | if ( $mmlPath ) { |
| 248 | $this->exportMMLtoFile( $mmlPath, $allFormula, $pTitle ); |
| 249 | } |
| 250 | |
| 251 | return $matches; |
| 252 | |
| 253 | } |
| 254 | return 0; |
| 255 | } |
| 256 | |
| 257 | private function getParserOptions(): ParserOptions { |
| 258 | if ( !$this->parserOptions ) { |
| 259 | $this->parserOptions = ParserOptions::newFromAnon(); |
| 260 | } |
| 261 | return $this->parserOptions; |
| 262 | } |
| 263 | |
| 264 | private function getParser( $revId ): Parser { |
| 265 | if ( !$this->parser ) { |
| 266 | $this->parser = MediaWikiServices::getInstance()->getParserFactory()->create(); |
| 267 | } |
| 268 | // hack to set private field mRevisionId id |
| 269 | $this->parser->preprocess( |
| 270 | '', |
| 271 | null, |
| 272 | $this->getParserOptions(), |
| 273 | $revId ); |
| 274 | return $this->parser; |
| 275 | } |
| 276 | |
| 277 | public function execute() { |
| 278 | global $wgMathValidModes; |
| 279 | $this->dbw = MediaWikiServices::getInstance() |
| 280 | ->getConnectionProvider() |
| 281 | ->getPrimaryDatabase(); |
| 282 | $this->purge = $this->getOption( "purge", false ); |
| 283 | $this->verbose = $this->getOption( "verbose", false ); |
| 284 | $this->renderingMode = $this->getOption( "mode", 'latexml' ); |
| 285 | $this->chunkSize = $this->getOption( 'chunk-size', $this->chunkSize ); |
| 286 | $this->db = MediaWikiServices::getInstance() |
| 287 | ->getConnectionProvider() |
| 288 | ->getPrimaryDatabase(); |
| 289 | $wgMathValidModes[] = $this->renderingMode; |
| 290 | $this->output( "Loaded.\n" ); |
| 291 | $this->time = microtime( true ); |
| 292 | $this->populateSearchIndex( $this->getArg( 0, 0 ), $this->getArg( 1, -1 ) ); |
| 293 | } |
| 294 | |
| 295 | /** |
| 296 | * Fetches a MathML entry for exporting formulas from renderer and forms an entry for json export. |
| 297 | * @param string $formula formula in tex to save |
| 298 | * @param MathRenderer $renderer mathrenderer object which contains mathml |
| 299 | * @param array $allFormula array which is filled with formula entries |
| 300 | * @return array modified allFormula array |
| 301 | */ |
| 302 | public function getMathMLForExport( string $formula, MathRenderer $renderer, array $allFormula ): array { |
| 303 | if ( $this->verbose ) { |
| 304 | echo "\n Fetching MML for formula: " . $formula . "\n"; |
| 305 | } |
| 306 | $mathML = $renderer->getMathml(); |
| 307 | if ( $this->verbose ) { |
| 308 | echo "\n Input-type is: " . $renderer->getInputType(); |
| 309 | echo "\n MathML is" . substr( $mathML, 0, 50 ); |
| 310 | } |
| 311 | $allFormula[] = [ |
| 312 | 'tex' => $formula, |
| 313 | 'type' => $renderer->getInputType(), |
| 314 | 'mml' => $mathML, |
| 315 | ]; |
| 316 | return $allFormula; |
| 317 | } |
| 318 | |
| 319 | /** |
| 320 | * Writes the MathML content in allFormula to a file named '<mmlPath>/mmlAllResults-<mode>-<pTitle>.json' |
| 321 | * @param string $mmlPath path for saving the mathml (without filename) |
| 322 | * @param array $allFormula all formula array with mathml for the current page |
| 323 | * @param string $pTitle title of page |
| 324 | * @return void |
| 325 | * @throws InvalidArgumentException when the filepath defined by cli-arg is not a correct folder |
| 326 | */ |
| 327 | public function exportMMLtoFile( string $mmlPath, array $allFormula, string $pTitle ): void { |
| 328 | if ( !is_dir( $mmlPath ) ) { |
| 329 | throw new InvalidArgumentException( "Filepath for exportmml at not valid at: " . $mmlPath ); |
| 330 | } |
| 331 | $jsonData = json_encode( $allFormula, JSON_PRETTY_PRINT ); |
| 332 | $fullPath = realpath( $mmlPath ) . DIRECTORY_SEPARATOR . 'mmlRes-' . $this->renderingMode . |
| 333 | "-" . $pTitle . ".json"; |
| 334 | file_put_contents( $fullPath, $jsonData ); |
| 335 | } |
| 336 | } |
| 337 | |
| 338 | $maintClass = UpdateMath::class; |
| 339 | /** @noinspection PhpIncludeInspection */ |
| 340 | require_once RUN_MAINTENANCE_IF_MAIN; |