text. * * @param \DOMXPath $xpath Query interface. * * @return string */ private function extract_title( \DOMXPath $xpath ) { $nodes = $xpath->query( '//title' ); if ( ! $nodes || 0 === $nodes->length ) { return ''; } return trim( $nodes->item( 0 )->textContent ); } /** * Collect every attribute of every node matching $query, in document * order, as a list of assoc arrays. * * @param \DOMXPath $xpath Query interface. * @param string $query XPath expression. * * @return array[] */ private function extract_attribute_list( \DOMXPath $xpath, $query ) { $results = array(); $nodes = $xpath->query( $query ); if ( ! $nodes ) { return $results; } foreach ( $nodes as $node ) { $results[] = $this->node_attributes( $node ); } return $results; } /** * <script> tags: attributes, plus inline text content when there is no * src attribute. * * @param \DOMXPath $xpath Query interface. * * @return array[] */ private function extract_scripts( \DOMXPath $xpath ) { $results = array(); $nodes = $xpath->query( '//script' ); if ( ! $nodes ) { return $results; } foreach ( $nodes as $node ) { $attrs = $this->node_attributes( $node ); if ( ! isset( $attrs['src'] ) ) { $attrs['inline_content'] = $node->textContent; } $results[] = $attrs; } return $results; } /** * Raw outerHTML for every <nav> element, document order. * * @param \DOMXPath $xpath Query interface. * * @return string[] */ private function extract_nav_html( \DOMXPath $xpath ) { $results = array(); $nodes = $xpath->query( '//nav' ); if ( ! $nodes ) { return $results; } foreach ( $nodes as $node ) { $results[] = $node->ownerDocument->saveHTML( $node ); } return $results; } /** * Inner HTML of <body>, with <script>/<link> descendants always * removed (re-emitted via wp_enqueue_*) and <nav> removed too when * $strip_nav is true — see parse()'s docblock for when it isn't. * * @param \DOMXPath $xpath Query interface. * @param bool $strip_nav Whether to also remove <nav> descendants. * * @return string */ private function extract_body_html( \DOMXPath $xpath, $strip_nav ) { $body_nodes = $xpath->query( '//body' ); if ( ! $body_nodes || 0 === $body_nodes->length ) { return ''; } $body = $body_nodes->item( 0 ); $query = $strip_nav ? './/script | .//link | .//nav' : './/script | .//link'; // Snapshot into a plain array first: DOMNodeList is live, so // removing nodes while iterating it directly would skip siblings. $to_remove = array(); foreach ( $xpath->query( $query, $body ) as $node ) { $to_remove[] = $node; } foreach ( $to_remove as $node ) { $node->parentNode->removeChild( $node ); } return $this->inner_html( $body ); } /** * Serialize a node's children (not the node itself) back to HTML. * * @param \DOMNode $node Node whose children should be serialized. * * @return string */ private function inner_html( \DOMNode $node ) { $html = ''; foreach ( $node->childNodes as $child ) { $html .= $node->ownerDocument->saveHTML( $child ); } return $html; } /** * All attributes of a node as an assoc array, in attribute order. * * @param \DOMNode $node Node to read attributes from. * * @return array<string,string> */ private function node_attributes( \DOMNode $node ) { $attrs = array(); if ( ! $node->hasAttributes() ) { return $attrs; } foreach ( $node->attributes as $attribute ) { $attrs[ $attribute->name ] = $attribute->value; } return $attrs; } }

descendants from * body_html. True (the default) is correct * for the single-page pipeline, where * Nav_Mapper turns $doc->nav_html into a * wp_nav_menu() call in header.php instead * — leaving the original markup in * body_html too would render the nav * twice. A multi-page conversion * (Multi_Page_Rewriter) passes false: its * navs vary structurally page to page (in * the real export this was built against, * the homepage's nav used in-page anchor * links while every other page's used real * page links) and don't fit Nav_Mapper's *

-or-direct--children * assumptions at all, so each page keeps * its own nav embedded in its own content * instead of one shared wp_nav_menu() call. * Silently stripping it there — as * discovered against a real export — also * corrupts the exported HTML enough to * break the page's own client-side * hydration, not just visually drop the nav. * * @return Parsed_Document */ public function parse( $html, $strip_nav = true ) { $dom = new \DOMDocument(); $previous_setting = libxml_use_internal_errors( true ); // The XML prolog trick forces DOMDocument to treat the input as // UTF-8 without requiring mbstring-based re-encoding. $dom->loadHTML( '' . $html, LIBXML_NOWARNING | LIBXML_NOERROR ); libxml_clear_errors(); libxml_use_internal_errors( $previous_setting ); $xpath = new \DOMXPath( $dom ); $doc = new Parsed_Document(); $doc->title = $this->extract_title( $xpath ); $doc->meta = $this->extract_attribute_list( $xpath, '//meta' ); $doc->links = $this->extract_attribute_list( $xpath, '//link' ); $doc->scripts = $this->extract_scripts( $xpath ); $doc->nav_html = $this->extract_nav_html( $xpath ); $doc->body_html = $this->extract_body_html( $xpath, $strip_nav ); $doc->stylesheets = array_values( array_filter( $doc->links, function ( $attrs ) { return isset( $attrs['rel'] ) && 'stylesheet' === strtolower( $attrs['rel'] ); } ) ); return $doc; } /** * Extract the document text. * * @param \DOMXPath $xpath Query interface. * * @return string */ private function extract_title( \DOMXPath $xpath ) { $nodes = $xpath->query( '//title' ); if ( ! $nodes || 0 === $nodes->length ) { return ''; } return trim( $nodes->item( 0 )->textContent ); } /** * Collect every attribute of every node matching $query, in document * order, as a list of assoc arrays. * * @param \DOMXPath $xpath Query interface. * @param string $query XPath expression. * * @return array[] */ private function extract_attribute_list( \DOMXPath $xpath, $query ) { $results = array(); $nodes = $xpath->query( $query ); if ( ! $nodes ) { return $results; } foreach ( $nodes as $node ) { $results[] = $this->node_attributes( $node ); } return $results; } /** * <script> tags: attributes, plus inline text content when there is no * src attribute. * * @param \DOMXPath $xpath Query interface. * * @return array[] */ private function extract_scripts( \DOMXPath $xpath ) { $results = array(); $nodes = $xpath->query( '//script' ); if ( ! $nodes ) { return $results; } foreach ( $nodes as $node ) { $attrs = $this->node_attributes( $node ); if ( ! isset( $attrs['src'] ) ) { $attrs['inline_content'] = $node->textContent; } $results[] = $attrs; } return $results; } /** * Raw outerHTML for every <nav> element, document order. * * @param \DOMXPath $xpath Query interface. * * @return string[] */ private function extract_nav_html( \DOMXPath $xpath ) { $results = array(); $nodes = $xpath->query( '//nav' ); if ( ! $nodes ) { return $results; } foreach ( $nodes as $node ) { $results[] = $node->ownerDocument->saveHTML( $node ); } return $results; } /** * Inner HTML of <body>, with <script>/<link> descendants always * removed (re-emitted via wp_enqueue_*) and <nav> removed too when * $strip_nav is true — see parse()'s docblock for when it isn't. * * @param \DOMXPath $xpath Query interface. * @param bool $strip_nav Whether to also remove <nav> descendants. * * @return string */ private function extract_body_html( \DOMXPath $xpath, $strip_nav ) { $body_nodes = $xpath->query( '//body' ); if ( ! $body_nodes || 0 === $body_nodes->length ) { return ''; } $body = $body_nodes->item( 0 ); $query = $strip_nav ? './/script | .//link | .//nav' : './/script | .//link'; // Snapshot into a plain array first: DOMNodeList is live, so // removing nodes while iterating it directly would skip siblings. $to_remove = array(); foreach ( $xpath->query( $query, $body ) as $node ) { $to_remove[] = $node; } foreach ( $to_remove as $node ) { $node->parentNode->removeChild( $node ); } return $this->inner_html( $body ); } /** * Serialize a node's children (not the node itself) back to HTML. * * @param \DOMNode $node Node whose children should be serialized. * * @return string */ private function inner_html( \DOMNode $node ) { $html = ''; foreach ( $node->childNodes as $child ) { $html .= $node->ownerDocument->saveHTML( $child ); } return $html; } /** * All attributes of a node as an assoc array, in attribute order. * * @param \DOMNode $node Node to read attributes from. * * @return array<string,string> */ private function node_attributes( \DOMNode $node ) { $attrs = array(); if ( ! $node->hasAttributes() ) { return $attrs; } foreach ( $node->attributes as $attribute ) { $attrs[ $attribute->name ] = $attribute->value; } return $attrs; } }