Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
77.98% |
85 / 109 |
|
75.00% |
6 / 8 |
CRAP | |
0.00% |
0 / 1 |
| TranslatablePageParser | |
77.98% |
85 / 109 |
|
75.00% |
6 / 8 |
28.65 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| containsMarkup | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| cleanupTags | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
1 | |||
| parse | |
71.43% |
30 / 42 |
|
0.00% |
0 / 1 |
9.49 | |||
| parseSection | |
100.00% |
18 / 18 |
|
100.00% |
1 / 1 |
3 | |||
| parseUnit | |
55.56% |
15 / 27 |
|
0.00% |
0 / 1 |
9.16 | |||
| armourNowiki | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
2 | |||
| unarmourNowiki | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| 1 | <?php |
| 2 | declare( strict_types = 1 ); |
| 3 | |
| 4 | namespace MediaWiki\Extension\Translate\PageTranslation; |
| 5 | |
| 6 | use MediaWiki\Extension\Translate\Utilities\ParsingPlaceholderFactory; |
| 7 | |
| 8 | /** |
| 9 | * Generates ParserOutput from text or removes all tags from a text. |
| 10 | * |
| 11 | * @author Niklas Laxström |
| 12 | * @license GPL-2.0-or-later |
| 13 | * @since 2020.08 |
| 14 | */ |
| 15 | class TranslatablePageParser { |
| 16 | private $placeholderFactory; |
| 17 | |
| 18 | public function __construct( ParsingPlaceholderFactory $placeholderFactory ) { |
| 19 | $this->placeholderFactory = $placeholderFactory; |
| 20 | } |
| 21 | |
| 22 | public function containsMarkup( string $text ): bool { |
| 23 | $nowiki = []; |
| 24 | $text = $this->armourNowiki( $nowiki, $text ); |
| 25 | return preg_match( '~</?translate[ >]~', $text ) !== 0; |
| 26 | } |
| 27 | |
| 28 | /** |
| 29 | * Remove all opening and closing translate tags following the same whitespace rules as the |
| 30 | * regular parsing. This doesn't try to parse the page, so it can handle unbalanced tags. |
| 31 | */ |
| 32 | public function cleanupTags( string $text ): string { |
| 33 | $nowiki = []; |
| 34 | $text = $this->armourNowiki( $nowiki, $text ); |
| 35 | $text = preg_replace( '~<translate( nowrap)?>\n?~s', '', $text ); |
| 36 | $text = preg_replace( '~\n?</translate>~s', '', $text ); |
| 37 | // Markers: headers and the rest |
| 38 | $ic = preg_quote( TranslationUnit::UNIT_MARKER_INVALID_CHARS, '~' ); |
| 39 | $text = preg_replace( "~(^=.*=) <!--T:[^$ic]+-->$~um", '\1', $text ); |
| 40 | $text = preg_replace( "~<!--T:[^$ic]+-->[\n ]?~um", '', $text ); |
| 41 | // Remove variables |
| 42 | $unit = new TranslationUnit( $text ); |
| 43 | $text = $unit->getTextForTrans(); |
| 44 | |
| 45 | $text = $this->unarmourNowiki( $nowiki, $text ); |
| 46 | return $text; |
| 47 | } |
| 48 | |
| 49 | /** @throws ParsingFailure */ |
| 50 | public function parse( string $text ): ParserOutput { |
| 51 | $nowiki = []; |
| 52 | $text = $this->armourNowiki( $nowiki, $text ); |
| 53 | |
| 54 | $sections = []; |
| 55 | $tagPlaceHolders = []; |
| 56 | |
| 57 | while ( true ) { |
| 58 | $re = '~(<translate(?: nowrap)?>)(.*?)</translate>~s'; |
| 59 | $matches = []; |
| 60 | $ok = preg_match( $re, $text, $matches, PREG_OFFSET_CAPTURE ); |
| 61 | |
| 62 | if ( $ok === 0 || $ok === false ) { |
| 63 | break; // No match or failure |
| 64 | } |
| 65 | |
| 66 | $contentWithTags = $matches[0][0]; |
| 67 | $contentWithoutTags = $matches[2][0]; |
| 68 | // These are offsets to the content inside the tags in $text |
| 69 | $offsetStart = $matches[0][1]; |
| 70 | $offsetEnd = $offsetStart + strlen( $contentWithTags ); |
| 71 | |
| 72 | // Replace the whole match with a placeholder |
| 73 | $ph = $this->placeholderFactory->make(); |
| 74 | $text = substr( $text, 0, $offsetStart ) . $ph . substr( $text, $offsetEnd ); |
| 75 | |
| 76 | if ( preg_match( '~<translate( nowrap)?>~', $contentWithoutTags ) !== 0 ) { |
| 77 | throw new ParsingFailure( |
| 78 | 'Nested tags', |
| 79 | [ 'pt-parse-nested', $contentWithoutTags ] |
| 80 | ); |
| 81 | } |
| 82 | |
| 83 | $openTag = $matches[1][0]; |
| 84 | $canWrap = $openTag !== '<translate nowrap>'; |
| 85 | |
| 86 | // Parse the content inside the tags |
| 87 | $contentWithoutTags = $this->unarmourNowiki( $nowiki, $contentWithoutTags ); |
| 88 | $parse = $this->parseSection( $contentWithoutTags, $canWrap ); |
| 89 | |
| 90 | // Update list of sections and the template with the results |
| 91 | $sections += $parse['sections']; |
| 92 | $tagPlaceHolders[$ph] = new Section( $openTag, $parse['template'], '</translate>' ); |
| 93 | } |
| 94 | |
| 95 | $prettyTemplate = $text; |
| 96 | foreach ( $tagPlaceHolders as $ph => $value ) { |
| 97 | $prettyTemplate = str_replace( $ph, '[...]', $prettyTemplate ); |
| 98 | } |
| 99 | |
| 100 | if ( preg_match( '~<translate( nowrap)?>~', $text ) !== 0 ) { |
| 101 | throw new ParsingFailure( |
| 102 | 'Unmatched opening tag', |
| 103 | [ 'pt-parse-open', $prettyTemplate ] |
| 104 | ); |
| 105 | } elseif ( str_contains( $text, '</translate>' ) ) { |
| 106 | throw new ParsingFailure( |
| 107 | "Unmatched closing tag", |
| 108 | [ 'pt-parse-close', $prettyTemplate ] |
| 109 | ); |
| 110 | } |
| 111 | |
| 112 | $text = $this->unarmourNowiki( $nowiki, $text ); |
| 113 | |
| 114 | return new ParserOutput( $text, $tagPlaceHolders, $sections ); |
| 115 | } |
| 116 | |
| 117 | /** |
| 118 | * Splits the content marked with \<translate> tags into translation units, which are |
| 119 | * separated with two or more newlines. Extra whitespace is captured in the template and |
| 120 | * is not included in the translation units. |
| 121 | * @internal |
| 122 | */ |
| 123 | public function parseSection( string $text, bool $canWrap ): array { |
| 124 | $flags = PREG_SPLIT_NO_EMPTY | PREG_SPLIT_DELIM_CAPTURE; |
| 125 | $parts = preg_split( '~(^\s*|\s*\n\n\s*|\s*$)~', $text, -1, $flags ); |
| 126 | |
| 127 | $inline = preg_match( '~\n~', $text ) === 0; |
| 128 | |
| 129 | $template = ''; |
| 130 | $sections = []; |
| 131 | |
| 132 | foreach ( $parts as $_ ) { |
| 133 | if ( trim( $_ ) === '' ) { |
| 134 | $template .= $_; |
| 135 | } else { |
| 136 | $ph = $this->placeholderFactory->make(); |
| 137 | $tpsection = $this->parseUnit( $_ ); |
| 138 | $tpsection->setIsInline( $inline ); |
| 139 | $tpsection->setCanWrap( $canWrap ); |
| 140 | $sections[$ph] = $tpsection; |
| 141 | $template .= $ph; |
| 142 | } |
| 143 | } |
| 144 | |
| 145 | return [ |
| 146 | 'template' => $template, |
| 147 | 'sections' => $sections, |
| 148 | ]; |
| 149 | } |
| 150 | |
| 151 | /** |
| 152 | * Checks if this unit already contains a section marker. If there |
| 153 | * is not, a new one will be created. Marker will have the value of |
| 154 | * -1, which will later be replaced with a real value. |
| 155 | * @internal |
| 156 | */ |
| 157 | public function parseUnit( string $content ): TranslationUnit { |
| 158 | $re = '~<!--T:(.*?)-->~'; |
| 159 | $matches = []; |
| 160 | $count = preg_match_all( $re, $content, $matches, PREG_SET_ORDER ); |
| 161 | |
| 162 | if ( $count > 1 ) { |
| 163 | throw new ParsingFailure( |
| 164 | 'Multiple translation unit markers', |
| 165 | [ 'pt-shake-multiple', $content ] |
| 166 | ); |
| 167 | } |
| 168 | |
| 169 | // If no id given in the source, default to a new section id |
| 170 | $id = TranslationUnit::NEW_UNIT_ID; |
| 171 | if ( $count === 1 ) { |
| 172 | foreach ( $matches as $match ) { |
| 173 | [ /*full*/, $id ] = $match; |
| 174 | |
| 175 | // Currently handle only these two standard places. |
| 176 | // Is this too strict? |
| 177 | $rer1 = '~^<!--T:(.*?)-->( |\n)~'; // Normal sections |
| 178 | $rer2 = '~\s*<!--T:(.*?)-->$~m'; // Sections with title |
| 179 | $content = preg_replace( $rer1, '', $content ); |
| 180 | $content = preg_replace( $rer2, '', $content ); |
| 181 | |
| 182 | if ( preg_match( $re, $content ) === 1 ) { |
| 183 | throw new ParsingFailure( |
| 184 | 'Translation unit marker is in unsupported position', |
| 185 | [ 'pt-shake-position', $content ] |
| 186 | ); |
| 187 | } elseif ( trim( $content ) === '' ) { |
| 188 | throw new ParsingFailure( |
| 189 | 'Translation unit has no content besides marker', |
| 190 | [ 'pt-shake-empty', $id ] |
| 191 | ); |
| 192 | } |
| 193 | } |
| 194 | } |
| 195 | |
| 196 | return new TranslationUnit( $content, $id ); |
| 197 | } |
| 198 | |
| 199 | /** @internal */ |
| 200 | public function armourNowiki( array &$holders, string $text ): string { |
| 201 | $re = '~(<nowiki>)(.*?)(</nowiki>)~s'; |
| 202 | |
| 203 | while ( preg_match( $re, $text, $matches ) ) { |
| 204 | $ph = $this->placeholderFactory->make(); |
| 205 | $text = str_replace( $matches[0], $ph, $text ); |
| 206 | $holders[$ph] = $matches[0]; |
| 207 | } |
| 208 | |
| 209 | return $text; |
| 210 | } |
| 211 | |
| 212 | /** @internal */ |
| 213 | public function unarmourNowiki( array $holders, string $text ): string { |
| 214 | return strtr( $text, $holders ); |
| 215 | } |
| 216 | } |