From b8d2ca802004d2ba2ec592d62f2c76c953abc360 Mon Sep 17 00:00:00 2001
From: =?utf8?q?Gerg=C5=91=20Tisza?= <tgr.huwiki@gmail.com>
Date: Wed, 12 Nov 2014 19:10:05 +0000
Subject: [PATCH] FormatMetadata::fetchExtendedMetadata: Ignore multiple
 EXIF/XMP values

Ignore all but the first value of multivalued metadata fields. Before
this, the function returned an array of values for such fields,
causing various errors.

Bug: 72652
Change-Id: I0fd22b78e4938eecc506d870e352b3b196ee3ace
Mingle: https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/973
---
 includes/media/FormatMetadata.php             | 50 +++++++++++++++++++
 .../includes/media/FormatMetadataTest.php     | 32 ++++++++++++
 2 files changed, 82 insertions(+)

diff --git a/includes/media/FormatMetadata.php b/includes/media/FormatMetadata.php
index 4356953997..07d7618c60 100644
--- a/includes/media/FormatMetadata.php
+++ b/includes/media/FormatMetadata.php
@@ -1624,6 +1624,7 @@ class FormatMetadata extends ContextSource {
 			if ( $this->singleLang ) {
 				$this->resolveMultilangMetadata( $extendedMetadata );
 			}
+			$this->discardMultipleValues( $extendedMetadata );
 			// Make sure the metadata won't break the API when an XML format is used.
 			// This is an API-specific function so it would be cleaner to call it from
 			// outside fetchExtendedMetadata, but this way we don't need to redo the
@@ -1776,6 +1777,32 @@ class FormatMetadata extends ContextSource {
 		return null;
 	}
 
+	/**
+	 * Turns an XMP-style multivalue array into a single value by dropping all but the first value.
+	 * If the value is not a multivalue array (or a multivalue array inside a multilang array), it is returned unchanged.
+	 * See mediawiki.org/wiki/Manual:File_metadata_handling#Multi-language_array_format
+	 * @param mixed $value
+	 * @return mixed The value, or the first value if there were multiple ones
+	 * @since 1.25
+	 */
+	protected function resolveMultivalueValue( $value ) {
+		if ( !is_array( $value ) ) {
+			return $value;
+		} elseif ( isset( $value['_type'] ) && $value['_type'] === 'lang' ) { // if this is a multilang array, process fields separately
+			$newValue = array();
+			foreach ( $value as $k => $v ) {
+				$newValue[$k] = $this->resolveMultivalueValue( $v );
+			}
+			return $newValue;
+		} else { // _type is 'ul' or 'ol' or missing in which case it defaults to 'ul'
+			list( $k, $v ) = each( $value );
+			if ( $k === '_type' ) {
+				$v = current( $value );
+			}
+			return $v;
+		}
+	}
+
 	/**
 	 * Takes an array returned by the getExtendedMetadata* functions,
 	 * and resolves multi-language values in it.
@@ -1793,6 +1820,29 @@ class FormatMetadata extends ContextSource {
 		}
 	}
 
+	/**
+	 * Takes an array returned by the getExtendedMetadata* functions,
+	 * and turns all fields into single-valued ones by dropping extra values.
+	 * @param array $metadata
+	 * @since 1.25
+	 */
+	protected function discardMultipleValues( &$metadata ) {
+		if ( !is_array( $metadata ) ) {
+			return;
+		}
+		foreach ( $metadata as $key => &$field ) {
+			if ( $key === 'Software' || $key === 'Contact' ) {
+				// we skip some fields which have composite values. They are not particularly interesting
+				// and you can get them via the metadata / commonmetadata APIs anyway.
+				continue;
+			}
+			if ( isset( $field['value'] ) ) {
+				$field['value'] = $this->resolveMultivalueValue( $field['value'] );
+			}
+		}
+
+	}
+
 	/**
 	 * Makes sure the given array is a valid API response fragment
 	 * (can be transformed into XML)
diff --git a/tests/phpunit/includes/media/FormatMetadataTest.php b/tests/phpunit/includes/media/FormatMetadataTest.php
index 002e2cb987..97aa0a3dce 100644
--- a/tests/phpunit/includes/media/FormatMetadataTest.php
+++ b/tests/phpunit/includes/media/FormatMetadataTest.php
@@ -68,4 +68,36 @@ class FormatMetadataTest extends MediaWikiMediaTestCase {
 			// TODO: more test cases
 		);
 	}
+
+	/**
+	 * @param mixed $input
+	 * @param mixed $output
+	 * @dataProvider provideResolveMultivalueValue
+	 * @covers FormatMetadata::resolveMultivalueValue
+	 */
+	public function testResolveMultivalueValue( $input, $output ) {
+		$formatMetadata = new FormatMetadata();
+		$class = new ReflectionClass( 'FormatMetadata' );
+		$method = $class->getMethod( 'resolveMultivalueValue' );
+		$method->setAccessible( true );
+		$actualInput = $method->invoke( $formatMetadata, $input );
+		$this->assertEquals( $output, $actualInput );
+	}
+
+	public function provideResolveMultivalueValue() {
+		return array(
+			'nonArray' => array( 'foo', 'foo' ),
+			'multiValue' => array( array( 'first', 'second', 'third', '_type' => 'ol' ), 'first' ),
+			'noType' => array( array( 'first', 'second', 'third' ), 'first' ),
+			'typeFirst' => array( array( '_type' => 'ol', 'first', 'second', 'third' ), 'first' ),
+		    'multilang' => array(
+			    array( 'en' => 'first', 'de' => 'Erste', '_type' => 'lang' ),
+			    array( 'en' => 'first', 'de' => 'Erste', '_type' => 'lang' ),
+		    ),
+		    'multilang-multivalue' => array(
+			    array( 'en' => array( 'first', 'second' ), 'de' => array( 'Erste', 'Zweite' ), '_type' => 'lang' ),
+			    array( 'en' => 'first', 'de' => 'Erste', '_type' => 'lang' ),
+		    ),
+		);
+	}
 }
-- 
2.20.1