Thijs/wiki.techinc.nl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Port categories dump header fix

Browse source 
This fix is ported from https://gerrit.wikimedia.org/r/#/c/372905
since it does not belong to SPARQL updates really.

It changes main node for category dump to
<http://acme.test/wiki/Special:CategoryDump>
thus allowing to support multiple wikis living on the same domain.w

Change-Id: Ie1ac5ddf6b3b73f3966274f90abc1db50061b494
This commit is contained in:
Stanislav Malyshev 2017-11-17 16:17:09 -08:00
parent ce42fdf151
commit bb2453fa7c
4 changed files with 36 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

28
includes/CategoriesRdf.php
View file

@ -37,7 +37,13 @@ class CategoriesRdf {
/**
* Current version of the dump format.
*/
const FORMAT_VERSION = "1.0";
const FORMAT_VERSION = "1.1";
/**
* Special page for Dump identification.
* Used as head URI for each wiki's category dump, e.g.:
* https://en.wikipedia.org/wiki/Special:CategoryDump
*/
const SPECIAL_DUMP = 'Special:CategoryDump';
/**
* @var RdfWriter
*/
@ -84,12 +90,30 @@ class CategoriesRdf {
$this->rdfWriter->say( 'rdfs', 'label' )->value( $titletext );
}
/**
* Make URL from title label
* @param string $titleLabel Short label (without namespace) of the category
* @return string URL for the category
*/
public function labelToUrl( $titleLabel ) {
return $this->titleToUrl( Title::makeTitle( NS_CATEGORY, $titleLabel ) );
}
/**
* Convert Title to link to target page.
* @param Title $title
* @return string
* @return string URL for the category
*/
private function titleToUrl( Title $title ) {
return $title->getFullURL( '', false, PROTO_CANONICAL );
}
/**
* Get URI of the dump for this particular wiki.
* @return false|string
*/
public function getDumpURI() {
return $this->titleToUrl( Title::makeTitle( NS_MAIN, self::SPECIAL_DUMP ) );
}
}

2
maintenance/dumpCategoriesAsRdf.php
View file

@ -96,7 +96,7 @@ class DumpCategoriesAsRdf extends Maintenance {
if ( substr( $licenseUrl, 0, 2 ) == '//' ) {
$licenseUrl = 'https:' . $licenseUrl;
}
$this->rdfWriter->about( wfExpandUrl( '/categoriesDump', PROTO_CANONICAL ) )
$this->rdfWriter->about( $this->categoriesRdf->getDumpURI() )
->a( 'schema', 'Dataset' )
->a( 'owl', 'Ontology' )
->say( 'cc', 'license' )->is( $licenseUrl )

14
tests/phpunit/data/categoriesrdf/categoriesRdf-out.nt
View file

@ -1,10 +1,10 @@
<http://acme.test/categoriesDump> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://schema.org/Dataset> .
<http://acme.test/categoriesDump> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2002/07/owl#Ontology> .
<http://acme.test/categoriesDump> <http://creativecommons.org/ns#license> <https://creativecommons.org/licenses/by-sa/3.0/> .
<http://acme.test/categoriesDump> <http://schema.org/softwareVersion> "1.0" .
<http://acme.test/categoriesDump> <http://schema.org/dateModified> "{DATE}"^^<http://www.w3.org/2001/XMLSchema#dateTime> .
<http://acme.test/categoriesDump> <http://schema.org/isPartOf> <http://acme.test/> .
<http://acme.test/categoriesDump> <http://www.w3.org/2002/07/owl#imports> <https://www.mediawiki.org/ontology/ontology.owl> .
<http://acme.test/wiki/Special:CategoryDump> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://schema.org/Dataset> .
<http://acme.test/wiki/Special:CategoryDump> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2002/07/owl#Ontology> .
<http://acme.test/wiki/Special:CategoryDump> <http://creativecommons.org/ns#license> <https://creativecommons.org/licenses/by-sa/3.0/> .
<http://acme.test/wiki/Special:CategoryDump> <http://schema.org/softwareVersion> "1.1" .
<http://acme.test/wiki/Special:CategoryDump> <http://schema.org/dateModified> "{DATE}"^^<http://www.w3.org/2001/XMLSchema#dateTime> .
<http://acme.test/wiki/Special:CategoryDump> <http://schema.org/isPartOf> <http://acme.test/> .
<http://acme.test/wiki/Special:CategoryDump> <http://www.w3.org/2002/07/owl#imports> <https://www.mediawiki.org/ontology/ontology.owl> .
<http://acme.test/wiki/Category:Category_One> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <https://www.mediawiki.org/ontology#Category> .
<http://acme.test/wiki/Category:Category_One> <http://www.w3.org/2000/01/rdf-schema#label> "Category One" .
<http://acme.test/wiki/Category:2_Category_Two> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <https://www.mediawiki.org/ontology#Category> .

4
tests/phpunit/maintenance/categoriesRdfTest.php
View file

@ -60,8 +60,8 @@ class CategoriesRdfTest extends MediaWikiLangTestCase {
$dumpScript->execute();
$actualOut = file_get_contents( $outFileName );
$actualOut = preg_replace(
'|<http://acme.test/categoriesDump> <http://schema.org/dateModified> "[^"]+?"|',
'<http://acme.test/categoriesDump> <http://schema.org/dateModified> "{DATE}"',
'|<http://acme.test/wiki/Special:CategoryDump> <http://schema.org/dateModified> "[^"]+?"|',
'<http://acme.test/wiki/Special:CategoryDump> <http://schema.org/dateModified> "{DATE}"',
$actualOut
);