Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
81.08% |
30 / 37 |
|
80.00% |
4 / 5 |
CRAP | |
0.00% |
0 / 1 |
| EditedSectionFinder | |
81.08% |
30 / 37 |
|
80.00% |
4 / 5 |
18.96 | |
0.00% |
0 / 1 |
| findEditedSectionsBetweenTextContents | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| findEditedSectionsBetweenRevisions | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
20 | |||
| findEditedSectionsByDiffOps | |
100.00% |
19 / 19 |
|
100.00% |
1 / 1 |
7 | |||
| isSectionLine | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getSectionTitlesFromLines | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
4 | |||
| 1 | <?php |
| 2 | declare( strict_types = 1 ); |
| 3 | |
| 4 | namespace ContentTranslation\Service; |
| 5 | |
| 6 | use MediaWiki\Content\TextContent; |
| 7 | use MediaWiki\Revision\RevisionRecord; |
| 8 | use MediaWiki\Revision\SlotRecord; |
| 9 | use Wikimedia\Diff\DiffOp; |
| 10 | |
| 11 | class EditedSectionFinder { |
| 12 | // Top section titles follow this pattern: |
| 13 | // == Section Title == |
| 14 | private const SECTION_TITLE_PATTERN = "/^==([^=].*?)==$/"; |
| 15 | private const WIKITEXT_COMMENT_PATTERN = "/<!--(.*?)-->/"; |
| 16 | |
| 17 | /** |
| 18 | * Given two TextContent instances, one representing the new version of the |
| 19 | * text content and the second the old one, this method an array of diff |
| 20 | * operations (DiffOp objects), that represents the changes between these |
| 21 | * two text contents. |
| 22 | * |
| 23 | * @param TextContent $newContent |
| 24 | * @param TextContent $oldContent |
| 25 | * @return string[] |
| 26 | */ |
| 27 | public function findEditedSectionsBetweenTextContents( TextContent $newContent, TextContent $oldContent ): array { |
| 28 | $diff = $oldContent->diff( $newContent ); |
| 29 | $diffOps = $diff->getEdits(); |
| 30 | |
| 31 | return $this->findEditedSectionsByDiffOps( $diffOps ); |
| 32 | } |
| 33 | |
| 34 | /** |
| 35 | * Given a revision, this static method finds all sections that were edited |
| 36 | * in this revision, and returns an array filled with these section titles |
| 37 | * |
| 38 | * @param RevisionRecord $currentRevision |
| 39 | * @param RevisionRecord|null $previousRevision |
| 40 | * @return string[] |
| 41 | */ |
| 42 | public function findEditedSectionsBetweenRevisions( |
| 43 | RevisionRecord $currentRevision, |
| 44 | ?RevisionRecord $previousRevision |
| 45 | ): array { |
| 46 | $newContent = $currentRevision->getContent( SlotRecord::MAIN, RevisionRecord::RAW ); |
| 47 | // If no previous revision, set old content to empty TextContent object |
| 48 | if ( $previousRevision instanceof RevisionRecord ) { |
| 49 | $oldContent = $previousRevision->getContent( SlotRecord::MAIN, RevisionRecord::RAW ); |
| 50 | } else { |
| 51 | // TextContentHandler::makeEmptyContent method doesn't initialize the content model |
| 52 | // properly, leading to CI errors. To fix this issue, instantiate oldContent manually |
| 53 | $oldContent = new TextContent( '', $newContent->getModel() ); |
| 54 | } |
| 55 | |
| 56 | // We can only calculate the difference for TextContent instances. |
| 57 | // If any of the content variables is not an instance of TextContent |
| 58 | // class, then return an empty array. In fact, both variables are |
| 59 | // expected to be instances of TextContent class when the method is |
| 60 | // actually called, however that may not be the case for CI tests. |
| 61 | if ( !$newContent instanceof TextContent || !$oldContent instanceof TextContent ) { |
| 62 | return []; |
| 63 | } |
| 64 | |
| 65 | return $this->findEditedSectionsBetweenTextContents( $newContent, $oldContent ); |
| 66 | } |
| 67 | |
| 68 | /** |
| 69 | * Given an array of diff operations, find all sections that were changed during the |
| 70 | * latest page edit |
| 71 | * |
| 72 | * @param DiffOp[] $diffOps |
| 73 | * @return string[] |
| 74 | */ |
| 75 | private function findEditedSectionsByDiffOps( array $diffOps ): array { |
| 76 | $diffOps = array_values( $diffOps ); |
| 77 | $editedSections = []; |
| 78 | foreach ( $diffOps as $i => $diffOp ) { |
| 79 | // If first diff operation is an "add" operation, it should refer |
| 80 | // to an edit in the lead section of the page. Since we only care |
| 81 | // about edits in non-lead sections, we skip first operation. |
| 82 | // If current diff operation is a copy, then we just skip, as copies |
| 83 | // that are followed by other copies do not contain edited sections |
| 84 | if ( $i === 0 || $diffOp->getType() === "copy" ) { |
| 85 | continue; |
| 86 | } |
| 87 | |
| 88 | /** |
| 89 | * This variable contains all lines on the right ("new") side of the diff, or |
| 90 | * false when it's a delete operation. |
| 91 | * In case of delete operations, we do no consider sections that were deleted as |
| 92 | * edited sections. |
| 93 | * |
| 94 | * @type string[]|false |
| 95 | */ |
| 96 | $textAfterOp = $diffOp->getClosing(); |
| 97 | $currentEditedSections = []; |
| 98 | // This variable indicates whether the first line of the current diff operation |
| 99 | // is a section title. When true, this variable indicates that a new section |
| 100 | // was added without modifying the previous one. |
| 101 | // If current diff operation is a "delete" operation, this variable is always |
| 102 | // false. |
| 103 | $isFirstLineSectionTitle = false; |
| 104 | if ( $textAfterOp ) { |
| 105 | $isFirstLineSectionTitle = $this->isSectionLine( $diffOp->getClosing( 0 ) ); |
| 106 | $currentEditedSections = $this->getSectionTitlesFromLines( $textAfterOp ); |
| 107 | } |
| 108 | |
| 109 | // If first line of the current diff operation is a section title, then |
| 110 | // previous section was not edited. If not, the title of the currently |
| 111 | // edited section, is contained inside the previous diff operation |
| 112 | // (which is a copy). |
| 113 | if ( !$isFirstLineSectionTitle ) { |
| 114 | // This operation is the first operation or a copy operation |
| 115 | $previousOp = $diffOps[$i - 1]; |
| 116 | // The title of the current edited section is the last section title |
| 117 | // present inside the previous diff operation |
| 118 | $revertedPreviousLines = array_reverse( $previousOp->orig ); |
| 119 | $previousSectionTitles = $this->getSectionTitlesFromLines( $revertedPreviousLines ); |
| 120 | if ( $previousSectionTitles ) { |
| 121 | $editedSections[] = current( $previousSectionTitles ); |
| 122 | } |
| 123 | } |
| 124 | $editedSections = array_merge( $editedSections, $currentEditedSections ); |
| 125 | } |
| 126 | return $editedSections; |
| 127 | } |
| 128 | |
| 129 | /** |
| 130 | * Given a string, this method returns a boolean indicating whether this |
| 131 | * string is a line containing a section title. |
| 132 | * |
| 133 | * @param string $line |
| 134 | * @return bool |
| 135 | */ |
| 136 | private static function isSectionLine( string $line ): bool { |
| 137 | return (bool)preg_match( self::SECTION_TITLE_PATTERN, $line ); |
| 138 | } |
| 139 | |
| 140 | /** |
| 141 | * Given an array of strings, representing text lines, this static method |
| 142 | * filters only the lines that contain a section title, and finally returns |
| 143 | * an array with the actual section titles among these lines (if any). |
| 144 | * |
| 145 | * @param string[] $lines |
| 146 | * @return string[] |
| 147 | */ |
| 148 | private function getSectionTitlesFromLines( array $lines ): array { |
| 149 | $editedSections = []; |
| 150 | foreach ( $lines as $line ) { |
| 151 | // Find the section title without the equal signs (==) |
| 152 | preg_match( self::SECTION_TITLE_PATTERN, trim( $line ), $matches ); |
| 153 | if ( $matches && isset( $matches[1] ) ) { |
| 154 | // Remove comments |
| 155 | $sectionTitle = preg_replace( self::WIKITEXT_COMMENT_PATTERN, '', $matches[1 ] ); |
| 156 | $editedSections[] = trim( $sectionTitle ); |
| 157 | } |
| 158 | } |
| 159 | return $editedSections; |
| 160 | } |
| 161 | } |