wikimedia/mediawiki-extensions-TextExtracts
1
0
Fork 0
mirror of https://gerrit.wikimedia.org/r/mediawiki/extensions/TextExtracts synced 2024-11-15 03:35:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge "Fix separation of text into sentences."

Browse source
This commit is contained in:
jenkins-bot 2016-04-01 20:45:08 +00:00 committed by Gerrit Code Review
parent f321fc541a 472d84c9de
commit 647c39734a
2 changed files with 7 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
includes/ExtractFormatter.php
View file

@ -80,7 +80,7 @@ class ExtractFormatter extends HtmlFormatter {
public static function getFirstSentences( $text, $requestedSentenceCount ) { public static function getFirstSentences( $text, $requestedSentenceCount ) {
// Based on code from OpenSearchXml by Brion Vibber // Based on code from OpenSearchXml by Brion Vibber
$endchars = array( $endchars = array(
'[^\p{Lu}]\.\s', '\!\s', '\?\s', // regular ASCII '[^\p{Lu}]\.[ \n]', '\![ \n]', '\?[ \n]', // regular ASCII
'。', // full-width ideographic full-stop '。', // full-width ideographic full-stop
'．', '！', '？', // double-width roman forms '．', '！', '？', // double-width roman forms
'｡', // half-width ideographic full stop '｡', // half-width ideographic full stop

6
tests/ExtractFormatterTest.php
View file

@ -115,6 +115,12 @@ class ExtractFormatterTest extends MediaWikiTestCase {
1, 1,
'P.J. Harvey is a singer.', 'P.J. Harvey is a singer.',
), ),
// Bug T115817 - Non-breaking space is not a delimiter
array(
html_entity_decode( 'Pigeons (lat. Columbidae) are birds. They primarily feed on seeds.' ),
1,
html_entity_decode( 'Pigeons (lat. Columbidae) are birds.' ),
),
); );
} }