Jdlrobson has uploaded a new change for review. (
https://gerrit.wikimedia.org/r/363083 )
Change subject: WIP: more test cases
......................................................................
WIP: more test cases
Change-Id: I928299d5f008c37d3bc7d28ce226bf4f6be02bcc
---
M includes/ExtractFormatter.php
M tests/phpunit/ExtractFormatterTest.php
2 files changed, 8 insertions(+), 25 deletions(-)
git pull ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/TextExtracts
refs/changes/83/363083/1
diff --git a/includes/ExtractFormatter.php b/includes/ExtractFormatter.php
index 48a4e88..dd714bb 100644
--- a/includes/ExtractFormatter.php
+++ b/includes/ExtractFormatter.php
@@ -212,31 +212,8 @@
* @return string
*/
protected static function getFirstSentencesPlain( $text,
$requestedSentenceCount ) {
- // Based on code from OpenSearchXml by Brion Vibber
- $endchars = [
- '[^\p{Lu}]\.(?:[ \n]|$)', '[\!\?](?:[ \n]|$)', //
regular ASCII
- '。', // full-width ideographic full-stop
- '．', '！', '？', // double-width roman forms
- '｡', // half-width ideographic full stop
- ];
-
- $endgroup = implode( '|', $endchars );
- $regexp = "/($endgroup)+/u";
-
- $matches = [];
- $res = preg_match_all( $regexp, $text, $matches,
PREG_OFFSET_CAPTURE );
-
- if ( $res ) {
- $index = min( $requestedSentenceCount, $res ) - 1;
- list( $tail, $length ) = $matches[0][ $index ];
- // PCRE returns raw offsets, so using substr() instead
of mb_substr()
- $text = substr( $text, 0, $length ) . trim( $tail );
- } else {
- // Just return the first line
- $lines = explode( "\n", $text, 2 );
- $text = trim( $lines[0] );
- }
- return $text;
+ $sentences = self::explodeSentences( $text );
+ return implode( '', array_slice( $sentences, 0,
$requestedSentenceCount ) );
}
/**
diff --git a/tests/phpunit/ExtractFormatterTest.php
b/tests/phpunit/ExtractFormatterTest.php
index 97826ec..4c0318c 100644
--- a/tests/phpunit/ExtractFormatterTest.php
+++ b/tests/phpunit/ExtractFormatterTest.php
@@ -110,6 +110,12 @@
1,
"<p>It's good to stay at the Y.M.C.A.
establishment.</p>"
],
+ // brackets are ignored
+ [
+ "<p><span>Jon Robson (b. 1985) wrote this
test.</span><span>Feel free to refactor it.</span><span>honestly</span></p>",
+ 1,
+ "<p><span>Jon Robson (b. 1985) wrote this
test.</span></p>"
+ ],
// Inappropriate use of `...` is fine
[
'<p>a. b... c. d. e. f. g.</p>',
--
To view, visit https://gerrit.wikimedia.org/r/363083
To unsubscribe, visit https://gerrit.wikimedia.org/r/settings
Gerrit-MessageType: newchange
Gerrit-Change-Id: I928299d5f008c37d3bc7d28ce226bf4f6be02bcc
Gerrit-PatchSet: 1
Gerrit-Project: mediawiki/extensions/TextExtracts
Gerrit-Branch: master
Gerrit-Owner: Jdlrobson <jrob...@wikimedia.org>
_______________________________________________
MediaWiki-commits mailing list
MediaWiki-commits@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/mediawiki-commits
