XMLProcessor: Support namespaces

components/DataLiberation/DataFormatConsumer/MarkupProcessorConsumer.php

@@ -60,7 +60,7 @@ public function consume() {
 							break;
 						}
 						$this->append_rich_text( htmlspecialchars( $this->markup_processor->get_modifiable_text() ) );
-						if ( in_array( $this->markup_processor->get_tag(), array( 'H1', 'H2', 'H3', 'H4', 'H5', 'H6' ) ) ) {
+						if ( in_array( $this->get_tag_name(), array( 'H1', 'H2', 'H3', 'H4', 'H5', 'H6' ) ) ) {
 							$this->on_title_candidate( $this->markup_processor->get_modifiable_text() );
 						}
 						break;
@@ -90,7 +90,11 @@ public function consume() {
 
 	private function handle_tag() {
 		$html          = $this->markup_processor;
-		$tag           = strtoupper( $html->get_tag() );
+		if($html instanceof WP_HTML_Processor) {
+			$tag = strtoupper( $html->get_tag() );
+		} else {
+			$tag = strtoupper( $html->get_tag_local_name() );
+		}
 		$tag_lowercase = strtolower( $tag );
 
 		$is_void_tag = ! $html->expects_closer() && ! $html->is_tag_closer();
@@ -100,14 +104,14 @@ private function handle_tag() {
 					$this->on_title_candidate( $html->get_modifiable_text() );
 					break;
 				case 'META':
-					$key   = $html->get_attribute( 'name' );
-					$value = $html->get_attribute( 'content' );
+					$key   = $this->get_attribute( 'name' );
+					$value = $this->get_attribute( 'content' );
 					if ( ! array_key_exists( $key, $this->metadata ) ) {
 						if ( $key ) {
 							$this->metadata[ $key ] = array();
 						}
 					}
-					switch ( $html->get_attribute( 'type' ) ) {
+					switch ( $this->get_attribute( 'type' ) ) {
 						case 'integer':
 							$value = (int) $value;
 							break;
@@ -124,8 +128,8 @@ private function handle_tag() {
 					$template = new WP_HTML_Tag_Processor( '<img>' );
 					$template->next_tag();
 					foreach ( array( 'alt', 'title', 'src' ) as $attr ) {
-						if ( $html->get_attribute( $attr ) ) {
-							$template->set_attribute( $attr, $html->get_attribute( $attr ) );
+						if ( $this->get_attribute( $attr ) ) {
+							$template->set_attribute( $attr, $this->get_attribute( $attr ) );
 						}
 					}
 					/**
@@ -211,8 +215,8 @@ private function handle_tag() {
 				case 'A':
 					$template = new WP_HTML_Tag_Processor( '<a>' );
 					$template->next_tag();
-					if ( $html->get_attribute( 'href' ) ) {
-						$template->set_attribute( 'href', $html->get_attribute( 'href' ) );
+					if ( $this->get_attribute( 'href' ) ) {
+						$template->set_attribute( 'href', $this->get_attribute( 'href' ) );
 					}
 					/**
 					 *
@@ -297,6 +301,22 @@ private function handle_tag() {
 		}
 	}
 
+	private function get_tag_name() {
+		if($this->markup_processor instanceof WP_HTML_Processor) {
+			return $this->markup_processor->get_tag();
+		} else {
+			return $this->markup_processor->get_tag_local_name();
+		}
+	}
+
+	private function get_attribute($key) {
+		if($this->markup_processor instanceof WP_HTML_Processor) {
+			return $this->markup_processor->get_attribute($key);
+		} else {
+			return $this->markup_processor->get_attribute('', $key);
+		}
+	}
+
 	private function on_title_candidate( $text ) {
 		if ( ! array_key_exists( 'post_title', $this->metadata ) ) {
 			$this->metadata['post_title'] = array(

components/DataLiberation/EntityReader/EPubEntityReader.php

@@ -54,6 +54,7 @@ public function next_entity() {
 				return false;
 			}
 
+			$this->remaining_html_files = [];
 			foreach ( $this->manifest['items'] as $item ) {
 				if ( $item['media-type'] !== 'application/xhtml+xml' ) {
 					continue;
@@ -137,11 +138,12 @@ private function parse_manifest() {
 		$xml = XMLProcessor::create_from_string(
 			$this->zip->get_contents( 'META-INF/container.xml' )
 		);
-		if ( false === $xml->next_tag( 'rootfile' ) ) {
+
+		if ( false === $xml->next_tag( ['urn:oasis:names:tc:opendocument:xmlns:container', 'rootfile'] ) ) {
 			return false;
 		}
 
-		$full_path = $xml->get_attribute( 'full-path' );
+		$full_path = $xml->get_attribute( '', 'full-path' );
 		if ( ! $full_path ) {
 			return false;
 		}
@@ -161,16 +163,16 @@ private function parse_manifest() {
 		);
 		while ( $xml->next_tag() ) {
 			$parsed_entry = array();
-			$keys         = $xml->get_attribute_names_with_prefix( '' );
-			foreach ( $keys as $key ) {
-				$parsed_entry[ $key ] = $xml->get_attribute( $key );
+			$keys         = $xml->get_attribute_names_with_prefix( '', '' );
+			foreach ( $keys as list($ns, $key) ) {
+				$parsed_entry[ $key ] = $xml->get_attribute( $ns, $key );
 			}
-			if ( $xml->matches_breadcrumbs( array( 'metadata', '*' ) ) ) {
+			if ( $xml->matches_breadcrumbs( array( 'package', 'metadata', '*' ) ) ) {
 				$parsed['metadata'][] = array(
-					'tag'        => $xml->get_tag(),
+					'tag'        => $xml->get_tag_local_name(),
 					'attributes' => $parsed_entry,
 				);
-			} elseif ( $xml->matches_breadcrumbs( array( 'manifest', 'item' ) ) ) {
+			} elseif ( $xml->matches_breadcrumbs( array( 'package', 'manifest', 'item' ) ) ) {
 				$parsed_entry['type'] = 'item';
 				$parsed['items'][]    = $parsed_entry;
 			}