Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
89.60% |
112 / 125 |
|
50.00% |
5 / 10 |
CRAP | |
0.00% |
0 / 1 |
| ApiHooksHandler | |
89.60% |
112 / 125 |
|
50.00% |
5 / 10 |
74.36 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| onAPIGetAllowedParams | |
100.00% |
13 / 13 |
|
100.00% |
1 / 1 |
10 | |||
| onApiQueryBaseBeforeQuery | |
85.29% |
29 / 34 |
|
0.00% |
0 / 1 |
13.54 | |||
| onApiQueryBaseAfterQuery | |
80.00% |
20 / 25 |
|
0.00% |
0 / 1 |
15.57 | |||
| loadScoresForRevisions | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
3 | |||
| onApiQueryBaseProcessRow | |
91.67% |
11 / 12 |
|
0.00% |
0 / 1 |
7.03 | |||
| addScoresForAPI | |
94.44% |
17 / 18 |
|
0.00% |
0 / 1 |
10.02 | |||
| onApiQueryWatchlistPrepareWatchedItemQueryServiceOptions | |
85.71% |
6 / 7 |
|
0.00% |
0 / 1 |
7.14 | |||
| onApiQueryWatchlistExtractOutputData | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
3 | |||
| onWatchedItemQueryServiceExtensions | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Copyright (C) 2016 Brad Jorsch <bjorsch@wikimedia.org> |
| 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 3 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 |
| 16 | * along with this program. If not, see <http://www.gnu.org/licenses/>. |
| 17 | */ |
| 18 | |
| 19 | namespace ORES\Hooks\Api; |
| 20 | |
| 21 | use MediaWiki\Api\ApiBase; |
| 22 | use MediaWiki\Api\ApiQueryAllRevisions; |
| 23 | use MediaWiki\Api\ApiQueryBase; |
| 24 | use MediaWiki\Api\ApiQueryGeneratorBase; |
| 25 | use MediaWiki\Api\ApiQueryRecentChanges; |
| 26 | use MediaWiki\Api\ApiQueryRevisions; |
| 27 | use MediaWiki\Api\ApiQueryUserContribs; |
| 28 | use MediaWiki\Api\ApiQueryWatchlist; |
| 29 | use MediaWiki\Api\ApiResult; |
| 30 | use MediaWiki\Api\Hook\APIGetAllowedParamsHook; |
| 31 | use MediaWiki\Api\Hook\ApiQueryBaseAfterQueryHook; |
| 32 | use MediaWiki\Api\Hook\ApiQueryBaseBeforeQueryHook; |
| 33 | use MediaWiki\Api\Hook\ApiQueryBaseProcessRowHook; |
| 34 | use MediaWiki\Api\Hook\ApiQueryWatchlistExtractOutputDataHook; |
| 35 | use MediaWiki\Api\Hook\ApiQueryWatchlistPrepareWatchedItemQueryServiceOptionsHook; |
| 36 | use MediaWiki\Hook\WatchedItemQueryServiceExtensionsHook; |
| 37 | use MediaWiki\Watchlist\WatchedItem; |
| 38 | use MediaWiki\Watchlist\WatchedItemQueryService; |
| 39 | use ORES\Hooks\Helpers; |
| 40 | use ORES\Services\ORESServices; |
| 41 | use Wikimedia\ParamValidator\ParamValidator; |
| 42 | use Wikimedia\Rdbms\IConnectionProvider; |
| 43 | use Wikimedia\Rdbms\IResultWrapper; |
| 44 | |
| 45 | class ApiHooksHandler implements |
| 46 | APIGetAllowedParamsHook, |
| 47 | ApiQueryBaseBeforeQueryHook, |
| 48 | ApiQueryBaseAfterQueryHook, |
| 49 | ApiQueryBaseProcessRowHook, |
| 50 | ApiQueryWatchlistExtractOutputDataHook, |
| 51 | ApiQueryWatchlistPrepareWatchedItemQueryServiceOptionsHook, |
| 52 | WatchedItemQueryServiceExtensionsHook |
| 53 | { |
| 54 | |
| 55 | private IConnectionProvider $dbProvider; |
| 56 | |
| 57 | public function __construct( IConnectionProvider $dbProvider ) { |
| 58 | $this->dbProvider = $dbProvider; |
| 59 | } |
| 60 | |
| 61 | /** |
| 62 | * Inject parameters into certain API modules |
| 63 | * |
| 64 | * - Adds an 'oresscores' prop to ApiQueryRevisions, ApiQueryAllRevisions, |
| 65 | * ApiQueryRecentChanges, ApiQueryWatchlist, and ApiQueryUserContribs |
| 66 | * - Adds 'oresreview' and '!oresreview' to the 'show' parameters of |
| 67 | * ApiQueryRecentChanges, ApiQueryWatchlist, and ApiQueryUserContribs. |
| 68 | * |
| 69 | * The actual implementations of these new parameters are handled by the |
| 70 | * various hook functions below and by \ORES\WatchedItemQueryServiceExtension. |
| 71 | * |
| 72 | * @param ApiBase $module |
| 73 | * @param array &$params |
| 74 | * @param int $flags zero or OR-ed flags like ApiBase::GET_VALUES_FOR_HELP |
| 75 | */ |
| 76 | public function onAPIGetAllowedParams( $module, &$params, $flags ) { |
| 77 | if ( $module instanceof ApiQueryRevisions || |
| 78 | $module instanceof ApiQueryAllRevisions || |
| 79 | $module instanceof ApiQueryRecentChanges || |
| 80 | $module instanceof ApiQueryWatchlist || |
| 81 | $module instanceof ApiQueryUserContribs |
| 82 | ) { |
| 83 | $params['prop'][ParamValidator::PARAM_TYPE][] = 'oresscores'; |
| 84 | } |
| 85 | |
| 86 | if ( Helpers::isModelEnabled( 'damaging' ) && ( |
| 87 | $module instanceof ApiQueryRecentChanges || |
| 88 | $module instanceof ApiQueryWatchlist || |
| 89 | $module instanceof ApiQueryUserContribs |
| 90 | ) ) { |
| 91 | $params['show'][ParamValidator::PARAM_TYPE][] = 'oresreview'; |
| 92 | $params['show'][ParamValidator::PARAM_TYPE][] = '!oresreview'; |
| 93 | $params['show'][ApiBase::PARAM_HELP_MSG_APPEND][] = 'ores-api-show-note'; |
| 94 | } |
| 95 | } |
| 96 | |
| 97 | /** |
| 98 | * Modify the API query before it's made. |
| 99 | * |
| 100 | * This mainly adds the joins and conditions necessary to implement the |
| 101 | * 'oresreview' and '!oresreview' values added to the 'show' parameters of |
| 102 | * ApiQueryRecentChanges and ApiQueryUserContribs. |
| 103 | * |
| 104 | * It also ensures that the query from ApiQueryRecentChanges includes the |
| 105 | * fields necessary to process rcprop=oresscores. |
| 106 | * |
| 107 | * @warning Any joins added *must* join on a unique key of the target table |
| 108 | * unless you really know what you're doing. |
| 109 | * @param ApiQueryBase $module |
| 110 | * @param array &$tables tables to be queried |
| 111 | * @param array &$fields columns to select |
| 112 | * @param array &$conds WHERE conditionals for query |
| 113 | * @param array &$options options for the database request |
| 114 | * @param array &$joinConds join conditions for the tables |
| 115 | * @param array &$hookData Inter-hook communication |
| 116 | */ |
| 117 | public function onApiQueryBaseBeforeQuery( |
| 118 | $module, &$tables, &$fields, &$conds, &$options, &$joinConds, &$hookData |
| 119 | ) { |
| 120 | $params = $module->extractRequestParams(); |
| 121 | |
| 122 | if ( $module instanceof ApiQueryRecentChanges ) { |
| 123 | $field = 'rc_this_oldid'; |
| 124 | |
| 125 | // Make sure the needed fields are included in the query, if necessary |
| 126 | if ( !$module->isInGeneratorMode() && in_array( 'oresscores', $params['prop'], true ) ) { |
| 127 | if ( !in_array( 'rc_this_oldid', $fields, true ) ) { |
| 128 | $fields[] = 'rc_this_oldid'; |
| 129 | } |
| 130 | if ( !in_array( 'rc_type', $fields, true ) ) { |
| 131 | $fields[] = 'rc_type'; |
| 132 | } |
| 133 | } |
| 134 | } elseif ( $module instanceof ApiQueryUserContribs ) { |
| 135 | $field = 'rev_id'; |
| 136 | } else { |
| 137 | return; |
| 138 | } |
| 139 | |
| 140 | $show = Helpers::isModelEnabled( 'damaging' ) |
| 141 | ? array_flip( $params['show'] ?? [] ) |
| 142 | : []; |
| 143 | if ( isset( $show['oresreview'] ) || isset( $show['!oresreview'] ) ) { |
| 144 | if ( isset( $show['oresreview'] ) && isset( $show['!oresreview'] ) ) { |
| 145 | $module->dieWithError( 'apierror-show' ); |
| 146 | } |
| 147 | |
| 148 | $threshold = |
| 149 | Helpers::getThreshold( 'damaging', $module->getUser(), $module->getTitle() ); |
| 150 | $dbr = $this->dbProvider->getReplicaDatabase(); |
| 151 | |
| 152 | $tables[] = 'ores_classification'; |
| 153 | |
| 154 | if ( isset( $show['oresreview'] ) ) { |
| 155 | $join = 'INNER JOIN'; |
| 156 | |
| 157 | // Filter out non-damaging and unscored edits. |
| 158 | $conds[] = $dbr->expr( 'oresc_probability', '>', $threshold ); |
| 159 | |
| 160 | // Performance hack: add STRAIGHT_JOIN (T146111) |
| 161 | $options[] = 'STRAIGHT_JOIN'; |
| 162 | } else { |
| 163 | $join = 'LEFT JOIN'; |
| 164 | |
| 165 | // Filter out damaging edits. |
| 166 | $conds[] = $dbr->expr( 'oresc_probability', '<=', $threshold ) |
| 167 | ->or( 'oresc_probability', '=', null ); |
| 168 | } |
| 169 | |
| 170 | $modelId = ORESServices::getModelLookup()->getModelId( 'damaging' ); |
| 171 | $joinConds['ores_classification'] = [ $join, [ |
| 172 | "oresc_rev=$field", |
| 173 | 'oresc_model' => $modelId, |
| 174 | 'oresc_class' => 1 |
| 175 | ] ]; |
| 176 | } |
| 177 | } |
| 178 | |
| 179 | /** |
| 180 | * Perform work after the API query is made |
| 181 | * |
| 182 | * This fetches the data necessary to handle the 'oresscores' prop added to |
| 183 | * ApiQueryRevisions, ApiQueryAllRevisions, ApiQueryRecentChanges, and |
| 184 | * ApiQueryUserContribs, to avoid having to make up to 5000 fetches to do |
| 185 | * it individually per row. |
| 186 | * |
| 187 | * The list of revids is extracted from $res and scores are fetched using |
| 188 | * self::loadScoresForRevisions(). The following keys are written into |
| 189 | * $hookData, if our ApiQueryBaseProcessRow hook function needs to do |
| 190 | * anything at all: |
| 191 | * - oresField: (string) Field in the result rows holding the revid |
| 192 | * - oresCheckRCType: (bool) Whether to skip rows where rc_type is not |
| 193 | * RC_EDIT or RC_NEW. |
| 194 | * - oresScores: (array) Array of arrays of row objects holding the scores |
| 195 | * for each revision we were able to fetch. |
| 196 | * |
| 197 | * @param ApiQueryBase $module |
| 198 | * @param IResultWrapper|bool $res |
| 199 | * @param array &$hookData Inter-hook communication |
| 200 | */ |
| 201 | public function onApiQueryBaseAfterQuery( $module, $res, &$hookData ) { |
| 202 | if ( !$res ) { |
| 203 | return; |
| 204 | } |
| 205 | |
| 206 | // If the module is being used as a generator, don't bother. Generators |
| 207 | // don't return props. |
| 208 | if ( $module instanceof ApiQueryGeneratorBase && $module->isInGeneratorMode() ) { |
| 209 | return; |
| 210 | } |
| 211 | |
| 212 | if ( $module instanceof ApiQueryRevisions || |
| 213 | $module instanceof ApiQueryAllRevisions || |
| 214 | $module instanceof ApiQueryUserContribs |
| 215 | ) { |
| 216 | $field = 'rev_id'; |
| 217 | $checkRCType = false; |
| 218 | } elseif ( $module instanceof ApiQueryRecentChanges ) { |
| 219 | $field = 'rc_this_oldid'; |
| 220 | $checkRCType = true; |
| 221 | } else { |
| 222 | return; |
| 223 | } |
| 224 | |
| 225 | $params = $module->extractRequestParams(); |
| 226 | if ( in_array( 'oresscores', $params['prop'], true ) ) { |
| 227 | // Extract revision IDs from the result set |
| 228 | $revids = []; |
| 229 | foreach ( $res as $row ) { |
| 230 | if ( !$checkRCType || (int)$row->rc_type === RC_EDIT || (int)$row->rc_type === RC_NEW ) { |
| 231 | $revids[] = $row->$field; |
| 232 | } |
| 233 | } |
| 234 | $res->rewind(); |
| 235 | |
| 236 | if ( $revids ) { |
| 237 | $hookData['oresField'] = $field; |
| 238 | $hookData['oresCheckRCType'] = $checkRCType; |
| 239 | $scores = self::loadScoresForRevisions( $revids ); |
| 240 | $hookData['oresScores'] = $scores; |
| 241 | } |
| 242 | } |
| 243 | } |
| 244 | |
| 245 | /** |
| 246 | * Load ORES score data for a list of revisions |
| 247 | * |
| 248 | * Scores already cached are fetched from the database. |
| 249 | * |
| 250 | * @param int[] $revids Revision IDs |
| 251 | * @return array |
| 252 | */ |
| 253 | public static function loadScoresForRevisions( array $revids ) { |
| 254 | $scores = []; |
| 255 | $models = []; |
| 256 | foreach ( ORESServices::getModelLookup()->getModels() as $modelName => $modelDatum ) { |
| 257 | $models[$modelDatum['id']] = $modelName; |
| 258 | } |
| 259 | |
| 260 | // Load cached score data |
| 261 | $dbResult = ORESServices::getScoreLookup()->getScores( $revids, array_values( $models ) ); |
| 262 | foreach ( $dbResult as $row ) { |
| 263 | $scores[$row->oresc_rev][] = $row; |
| 264 | } |
| 265 | |
| 266 | return $scores; |
| 267 | } |
| 268 | |
| 269 | /** |
| 270 | * Modify each data row before it's returned. |
| 271 | * |
| 272 | * This uses the data added to $hookData by |
| 273 | * self::onApiQueryBaseAfterQuery() to actually inject the scores into the |
| 274 | * result data structure. See the documentation of that method for the |
| 275 | * details of that data. |
| 276 | * |
| 277 | * @param ApiQueryBase $module |
| 278 | * @param \stdClass $row |
| 279 | * @param array &$data |
| 280 | * @param array &$hookData Inter-hook communication |
| 281 | */ |
| 282 | public function onApiQueryBaseProcessRow( |
| 283 | $module, |
| 284 | $row, |
| 285 | &$data, |
| 286 | &$hookData |
| 287 | ) { |
| 288 | if ( isset( $hookData['oresField'] ) && |
| 289 | ( !$hookData['oresCheckRCType'] || |
| 290 | (int)$row->rc_type === RC_NEW || (int)$row->rc_type === RC_EDIT |
| 291 | ) |
| 292 | ) { |
| 293 | $data['oresscores'] = []; |
| 294 | $revid = $row->{$hookData['oresField']}; |
| 295 | |
| 296 | $modelData = ORESServices::getModelLookup()->getModels(); |
| 297 | $models = []; |
| 298 | foreach ( $modelData as $modelName => $modelDatum ) { |
| 299 | $models[$modelDatum['id']] = $modelName; |
| 300 | } |
| 301 | |
| 302 | if ( !isset( $hookData['oresScores'][$revid] ) ) { |
| 303 | return; |
| 304 | } |
| 305 | |
| 306 | self::addScoresForAPI( $data, $hookData['oresScores'][$revid], $models ); |
| 307 | } |
| 308 | } |
| 309 | |
| 310 | /** |
| 311 | * Helper to actuall add scores to an API result array |
| 312 | * |
| 313 | * @param array &$data Output array |
| 314 | * @param \stdClass[] $scores Array of score data |
| 315 | * @param array $models |
| 316 | */ |
| 317 | private static function addScoresForAPI( array &$data, array $scores, array $models ) { |
| 318 | global $wgOresModelClasses; |
| 319 | static $classMap = null; |
| 320 | |
| 321 | if ( $classMap === null ) { |
| 322 | $classMap = array_map( 'array_flip', $wgOresModelClasses ); |
| 323 | } |
| 324 | |
| 325 | foreach ( $scores as $row ) { |
| 326 | if ( !isset( $row->oresm_name ) && isset( $row->oresc_model ) ) { |
| 327 | $row->oresm_name = $models[$row->oresc_model]; |
| 328 | } |
| 329 | |
| 330 | if ( !isset( $row->oresm_name ) || !isset( $classMap[$row->oresm_name][$row->oresc_class] ) ) { |
| 331 | // Missing configuration, ignore it |
| 332 | continue; |
| 333 | } |
| 334 | $data['oresscores'][$row->oresm_name][$classMap[$row->oresm_name][$row->oresc_class]] = |
| 335 | (float)$row->oresc_probability; |
| 336 | } |
| 337 | |
| 338 | foreach ( $data['oresscores'] as $model => &$outputScores ) { |
| 339 | // Recalculate the class-0 result, as it's not stored in the database |
| 340 | if ( isset( $classMap[$model][0] ) && !isset( $outputScores[$classMap[$model][0]] ) ) { |
| 341 | $outputScores[$classMap[$model][0]] = 1.0 - array_sum( $outputScores ); |
| 342 | } |
| 343 | |
| 344 | ApiResult::setArrayType( $outputScores, 'kvp', 'name' ); |
| 345 | ApiResult::setIndexedTagName( $outputScores, 'class' ); |
| 346 | } |
| 347 | unset( $outputScores ); |
| 348 | |
| 349 | ApiResult::setArrayType( $data['oresscores'], 'kvp', 'name' ); |
| 350 | ApiResult::setIndexedTagName( $data['oresscores'], 'model' ); |
| 351 | } |
| 352 | |
| 353 | /** |
| 354 | * Convert API parameters to WatchedItemQueryService options |
| 355 | * |
| 356 | * @param ApiQueryBase $module |
| 357 | * @param array $params |
| 358 | * @param array &$options |
| 359 | */ |
| 360 | public function onApiQueryWatchlistPrepareWatchedItemQueryServiceOptions( |
| 361 | $module, $params, &$options |
| 362 | ) { |
| 363 | if ( in_array( 'oresscores', $params['prop'], true ) ) { |
| 364 | $options['includeFields'][] = 'oresscores'; |
| 365 | } |
| 366 | |
| 367 | $show = array_flip( $params['show'] ?? [] ); |
| 368 | if ( isset( $show['oresreview'] ) || isset( $show['!oresreview'] ) ) { |
| 369 | if ( isset( $show['oresreview'] ) && isset( $show['!oresreview'] ) ) { |
| 370 | $module->dieWithError( 'apierror-show' ); |
| 371 | } |
| 372 | |
| 373 | $options['filters'][] = isset( $show['oresreview'] ) ? 'oresreview' : '!oresreview'; |
| 374 | } |
| 375 | } |
| 376 | |
| 377 | /** |
| 378 | * Add data to ApiQueryWatchlist output |
| 379 | * |
| 380 | * @param ApiQueryBase $module |
| 381 | * @param WatchedItem $watchedItem |
| 382 | * @param array $recentChangeInfo |
| 383 | * @param array &$output |
| 384 | */ |
| 385 | public function onApiQueryWatchlistExtractOutputData( |
| 386 | $module, $watchedItem, $recentChangeInfo, &$output |
| 387 | ) { |
| 388 | if ( isset( $recentChangeInfo['oresScores'] ) ) { |
| 389 | $modelData = ORESServices::getModelLookup()->getModels(); |
| 390 | |
| 391 | $models = []; |
| 392 | foreach ( $modelData as $modelName => $modelDatum ) { |
| 393 | $models[$modelDatum['id']] = $modelName; |
| 394 | } |
| 395 | self::addScoresForAPI( $output, $recentChangeInfo['oresScores'], $models ); |
| 396 | } |
| 397 | } |
| 398 | |
| 399 | /** |
| 400 | * @param array &$extensions |
| 401 | * @param WatchedItemQueryService $watchedItemQueryService |
| 402 | */ |
| 403 | public function onWatchedItemQueryServiceExtensions( &$extensions, $watchedItemQueryService ) { |
| 404 | $extensions[] = new WatchedItemQueryServiceExtension(); |
| 405 | } |
| 406 | } |