Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
47.95% |
35 / 73 |
|
20.00% |
4 / 20 |
CRAP | |
0.00% |
0 / 1 |
| Result | |
47.95% |
35 / 73 |
|
20.00% |
4 / 20 |
317.08 | |
0.00% |
0 / 1 |
| __construct | |
80.65% |
25 / 31 |
|
0.00% |
0 / 1 |
15.42 | |||
| pickTextSnippet | |
31.58% |
6 / 19 |
|
0.00% |
0 / 1 |
34.95 | |||
| getTitleSnippet | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getRedirectTitle | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| clearRedirectTitle | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
20 | |||
| getRedirectSnippet | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getTextSnippet | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getSectionSnippet | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getSectionTitle | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getCategorySnippet | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getWordCount | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getByteSize | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getTimestamp | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| isFileMatch | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getInterwikiPrefix | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getInterwikiNamespaceText | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getDocId | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getScore | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getExplanation | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getTitleHelper | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace CirrusSearch\Search; |
| 4 | |
| 5 | use CirrusSearch\Search\Fetch\HighlightingTrait; |
| 6 | use MediaWiki\Title\Title; |
| 7 | use MediaWiki\Utils\MWTimestamp; |
| 8 | |
| 9 | /** |
| 10 | * An individual search result from Elasticsearch. |
| 11 | * |
| 12 | * This program is free software; you can redistribute it and/or modify |
| 13 | * it under the terms of the GNU General Public License as published by |
| 14 | * the Free Software Foundation; either version 2 of the License, or |
| 15 | * (at your option) any later version. |
| 16 | * |
| 17 | * This program is distributed in the hope that it will be useful, |
| 18 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 19 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 20 | * GNU General Public License for more details. |
| 21 | * |
| 22 | * You should have received a copy of the GNU General Public License along |
| 23 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 24 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 25 | * http://www.gnu.org/copyleft/gpl.html |
| 26 | */ |
| 27 | class Result extends CirrusSearchResult { |
| 28 | use HighlightingTrait; |
| 29 | |
| 30 | /** @var string */ |
| 31 | private $titleSnippet = ''; |
| 32 | /** @var Title|null */ |
| 33 | private $redirectTitle = null; |
| 34 | /** @var string */ |
| 35 | private $redirectSnippet = ''; |
| 36 | /** @var Title|null */ |
| 37 | private $sectionTitle = null; |
| 38 | /** @var string */ |
| 39 | private $sectionSnippet = ''; |
| 40 | /** @var string */ |
| 41 | private $categorySnippet = ''; |
| 42 | /** @var string */ |
| 43 | private $textSnippet; |
| 44 | /** @var bool */ |
| 45 | private $isFileMatch = false; |
| 46 | /** @var string */ |
| 47 | private $namespaceText; |
| 48 | /** @var int */ |
| 49 | private $wordCount; |
| 50 | /** @var int */ |
| 51 | private $byteSize; |
| 52 | /** @var string */ |
| 53 | private $timestamp; |
| 54 | /** @var string */ |
| 55 | private $docId; |
| 56 | /** @var float */ |
| 57 | private $score; |
| 58 | /** @var array */ |
| 59 | private $explanation; |
| 60 | /** @var TitleHelper */ |
| 61 | private $titleHelper; |
| 62 | |
| 63 | /** |
| 64 | * Build the result. |
| 65 | * |
| 66 | * @param mixed $results Unused |
| 67 | * @param \Elastica\Result $result containing the given search result |
| 68 | * @param TitleHelper|null $titleHelper |
| 69 | */ |
| 70 | public function __construct( $results, $result, ?TitleHelper $titleHelper = null ) { |
| 71 | $this->titleHelper = $titleHelper ?: new TitleHelper(); |
| 72 | parent::__construct( $this->titleHelper->makeTitle( $result ) ); |
| 73 | $this->namespaceText = $result->namespace_text; |
| 74 | $this->docId = $result->getId(); |
| 75 | |
| 76 | $fields = $result->getFields(); |
| 77 | // Not all results requested a word count. Just pretend we have none if so |
| 78 | $this->wordCount = isset( $fields['text.word_count'] ) ? $fields['text.word_count'][ 0 ] : 0; |
| 79 | $this->byteSize = $result->text_bytes; |
| 80 | $this->timestamp = new MWTimestamp( $result->timestamp ); |
| 81 | $highlights = $result->getHighlights(); |
| 82 | // Evil hax to not special case .plain fields for intitle regex |
| 83 | foreach ( [ 'title', 'redirect.title' ] as $field ) { |
| 84 | if ( isset( $highlights["$field.plain"] ) && !isset( $highlights[$field] ) ) { |
| 85 | $highlights[$field] = $highlights["$field.plain"]; |
| 86 | unset( $highlights["$field.plain"] ); |
| 87 | } |
| 88 | } |
| 89 | if ( isset( $highlights[ 'title' ] ) ) { |
| 90 | $nstext = $this->getTitle()->getNamespace() === 0 ? '' : |
| 91 | $this->titleHelper->getNamespaceText( $this->getTitle() ) . ':'; |
| 92 | $this->titleSnippet = $nstext . $this->escapeHighlightedText( $highlights[ 'title' ][ 0 ] ); |
| 93 | } elseif ( $this->getTitle()->isExternal() ) { |
| 94 | // Interwiki searches are weird. They won't have title highlights by design, but |
| 95 | // if we don't return a title snippet we'll get weird display results. |
| 96 | $this->titleSnippet = $this->getTitle()->getText(); |
| 97 | } |
| 98 | |
| 99 | if ( !isset( $highlights[ 'title' ] ) && isset( $highlights[ 'redirect.title' ] ) ) { |
| 100 | // Make sure to find the redirect title before escaping because escaping breaks it.... |
| 101 | $this->redirectTitle = $this->findRedirectTitle( $result, $highlights[ 'redirect.title' ][ 0 ] ); |
| 102 | if ( $this->redirectTitle !== null ) { |
| 103 | $this->redirectSnippet = $this->escapeHighlightedText( $highlights[ 'redirect.title' ][ 0 ] ); |
| 104 | } |
| 105 | } |
| 106 | |
| 107 | $this->textSnippet = $this->escapeHighlightedText( $this->pickTextSnippet( $highlights ) ); |
| 108 | |
| 109 | if ( isset( $highlights[ 'heading' ] ) ) { |
| 110 | $this->sectionSnippet = $this->escapeHighlightedText( $highlights[ 'heading' ][ 0 ] ); |
| 111 | $this->sectionTitle = $this->findSectionTitle( $highlights[ 'heading' ][ 0 ], $this->getTitle() ); |
| 112 | } |
| 113 | |
| 114 | if ( isset( $highlights[ 'category' ] ) ) { |
| 115 | $this->categorySnippet = $this->escapeHighlightedText( $highlights[ 'category' ][ 0 ] ); |
| 116 | } |
| 117 | $this->score = $result->getScore(); |
| 118 | $this->explanation = $result->getExplanation(); |
| 119 | } |
| 120 | |
| 121 | /** |
| 122 | * @param string[] $highlights |
| 123 | * @return string |
| 124 | */ |
| 125 | private function pickTextSnippet( $highlights ) { |
| 126 | // This can get skipped if there the page was sent to Elasticsearch without text. |
| 127 | // This could be a bug or it could be that the page simply doesn't have any text. |
| 128 | $mainSnippet = ''; |
| 129 | // Prefer source_text.plain it's likely a regex |
| 130 | // TODO: use the priority system from the FetchPhaseConfigBuilder |
| 131 | if ( isset( $highlights[ 'source_text.plain' ] ) ) { |
| 132 | $sourceSnippet = $highlights[ 'source_text.plain' ][ 0 ]; |
| 133 | if ( $this->containsMatches( $sourceSnippet ) ) { |
| 134 | return $sourceSnippet; |
| 135 | } |
| 136 | } |
| 137 | if ( isset( $highlights[ 'text' ] ) ) { |
| 138 | $mainSnippet = $highlights[ 'text' ][ 0 ]; |
| 139 | if ( $this->containsMatches( $mainSnippet ) ) { |
| 140 | return $mainSnippet; |
| 141 | } |
| 142 | } |
| 143 | if ( isset( $highlights[ 'auxiliary_text' ] ) ) { |
| 144 | $auxSnippet = $highlights[ 'auxiliary_text' ][ 0 ]; |
| 145 | if ( $this->containsMatches( $auxSnippet ) ) { |
| 146 | return $auxSnippet; |
| 147 | } |
| 148 | } |
| 149 | if ( isset( $highlights[ 'file_text' ] ) ) { |
| 150 | $fileSnippet = $highlights[ 'file_text' ][ 0 ]; |
| 151 | if ( $this->containsMatches( $fileSnippet ) ) { |
| 152 | $this->isFileMatch = true; |
| 153 | return $fileSnippet; |
| 154 | } |
| 155 | } |
| 156 | return $mainSnippet; |
| 157 | } |
| 158 | |
| 159 | /** |
| 160 | * @return string |
| 161 | */ |
| 162 | public function getTitleSnippet() { |
| 163 | return $this->titleSnippet; |
| 164 | } |
| 165 | |
| 166 | /** |
| 167 | * @return Title|null |
| 168 | */ |
| 169 | public function getRedirectTitle() { |
| 170 | return $this->redirectTitle; |
| 171 | } |
| 172 | |
| 173 | protected function clearRedirectTitle(): bool { |
| 174 | $this->redirectTitle = null; |
| 175 | $this->redirectSnippet = ''; |
| 176 | |
| 177 | return !$this->containsHighlight( $this->textSnippet ) |
| 178 | && $this->titleSnippet === '' |
| 179 | && $this->sectionSnippet === '' |
| 180 | && $this->categorySnippet === ''; |
| 181 | } |
| 182 | |
| 183 | /** |
| 184 | * @return string |
| 185 | */ |
| 186 | public function getRedirectSnippet() { |
| 187 | return $this->redirectSnippet; |
| 188 | } |
| 189 | |
| 190 | /** |
| 191 | * @param array $terms |
| 192 | * @return string|null |
| 193 | */ |
| 194 | public function getTextSnippet( $terms = [] ) { |
| 195 | return $this->textSnippet; |
| 196 | } |
| 197 | |
| 198 | /** |
| 199 | * @return string |
| 200 | */ |
| 201 | public function getSectionSnippet() { |
| 202 | return $this->sectionSnippet; |
| 203 | } |
| 204 | |
| 205 | /** |
| 206 | * @return Title|null |
| 207 | */ |
| 208 | public function getSectionTitle() { |
| 209 | return $this->sectionTitle; |
| 210 | } |
| 211 | |
| 212 | /** |
| 213 | * @return string |
| 214 | */ |
| 215 | public function getCategorySnippet() { |
| 216 | return $this->categorySnippet; |
| 217 | } |
| 218 | |
| 219 | /** |
| 220 | * @return int |
| 221 | */ |
| 222 | public function getWordCount() { |
| 223 | return $this->wordCount; |
| 224 | } |
| 225 | |
| 226 | /** |
| 227 | * @return int |
| 228 | */ |
| 229 | public function getByteSize() { |
| 230 | return $this->byteSize; |
| 231 | } |
| 232 | |
| 233 | /** |
| 234 | * @return string |
| 235 | */ |
| 236 | public function getTimestamp() { |
| 237 | return $this->timestamp->getTimestamp( TS_MW ); |
| 238 | } |
| 239 | |
| 240 | /** |
| 241 | * @return bool |
| 242 | */ |
| 243 | public function isFileMatch() { |
| 244 | return $this->isFileMatch; |
| 245 | } |
| 246 | |
| 247 | /** |
| 248 | * @return string |
| 249 | */ |
| 250 | public function getInterwikiPrefix() { |
| 251 | return $this->getTitle()->getInterwiki(); |
| 252 | } |
| 253 | |
| 254 | /** |
| 255 | * @return string |
| 256 | */ |
| 257 | public function getInterwikiNamespaceText() { |
| 258 | // Seems to be only useful for API |
| 259 | return $this->namespaceText; |
| 260 | } |
| 261 | |
| 262 | /** |
| 263 | * @return string |
| 264 | */ |
| 265 | public function getDocId() { |
| 266 | return $this->docId; |
| 267 | } |
| 268 | |
| 269 | /** |
| 270 | * @return float the score |
| 271 | */ |
| 272 | public function getScore() { |
| 273 | return $this->score; |
| 274 | } |
| 275 | |
| 276 | /** |
| 277 | * @return array lucene score explanation |
| 278 | */ |
| 279 | public function getExplanation() { |
| 280 | return $this->explanation; |
| 281 | } |
| 282 | |
| 283 | /** |
| 284 | * @return TitleHelper |
| 285 | */ |
| 286 | protected function getTitleHelper(): TitleHelper { |
| 287 | return $this->titleHelper; |
| 288 | } |
| 289 | } |