diff options
Diffstat (limited to 'src/Nodes')
-rw-r--r-- | src/Nodes/DOMAttr.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMCdataSection.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMCharacterData.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMComment.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMDocument.php | 26 | ||||
-rw-r--r-- | src/Nodes/DOMDocumentFragment.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMDocumentType.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMElement.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMNode.php | 14 | ||||
-rw-r--r-- | src/Nodes/DOMNotation.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMProcessingInstruction.php | 8 | ||||
-rw-r--r-- | src/Nodes/DOMText.php | 8 | ||||
-rw-r--r-- | src/Nodes/NodeTrait.php | 426 | ||||
-rw-r--r-- | src/Nodes/NodeUtility.php | 155 |
14 files changed, 701 insertions, 0 deletions
diff --git a/src/Nodes/DOMAttr.php b/src/Nodes/DOMAttr.php new file mode 100644 index 0000000..c31517a --- /dev/null +++ b/src/Nodes/DOMAttr.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMAttr extends \DOMAttr +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMCdataSection.php b/src/Nodes/DOMCdataSection.php new file mode 100644 index 0000000..f3a56f0 --- /dev/null +++ b/src/Nodes/DOMCdataSection.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMCdataSection extends \DOMCdataSection +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMCharacterData.php b/src/Nodes/DOMCharacterData.php new file mode 100644 index 0000000..e5087d9 --- /dev/null +++ b/src/Nodes/DOMCharacterData.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMCharacterData extends \DOMCharacterData +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMComment.php b/src/Nodes/DOMComment.php new file mode 100644 index 0000000..fd2b8b5 --- /dev/null +++ b/src/Nodes/DOMComment.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMComment extends \DOMComment +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMDocument.php b/src/Nodes/DOMDocument.php new file mode 100644 index 0000000..510cdf2 --- /dev/null +++ b/src/Nodes/DOMDocument.php @@ -0,0 +1,26 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMDocument extends \DOMDocument +{ + use NodeTrait; + + public function __construct($version, $encoding) + { + parent::__construct($version, $encoding); + + $this->registerNodeClass('DOMAttr', DOMAttr::class); + $this->registerNodeClass('DOMCdataSection', DOMCdataSection::class); + $this->registerNodeClass('DOMCharacterData', DOMCharacterData::class); + $this->registerNodeClass('DOMComment', DOMComment::class); + $this->registerNodeClass('DOMDocument', self::class); + $this->registerNodeClass('DOMDocumentFragment', DOMDocumentFragment::class); + $this->registerNodeClass('DOMDocumentType', DOMDocumentType::class); + $this->registerNodeClass('DOMElement', DOMElement::class); + $this->registerNodeClass('DOMNode', DOMNode::class); + $this->registerNodeClass('DOMNotation', DOMNotation::class); + $this->registerNodeClass('DOMProcessingInstruction', DOMProcessingInstruction::class); + $this->registerNodeClass('DOMText', DOMText::class); + } +} diff --git a/src/Nodes/DOMDocumentFragment.php b/src/Nodes/DOMDocumentFragment.php new file mode 100644 index 0000000..d5f013e --- /dev/null +++ b/src/Nodes/DOMDocumentFragment.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMDocumentFragment extends \DOMDocumentFragment +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMDocumentType.php b/src/Nodes/DOMDocumentType.php new file mode 100644 index 0000000..81e426b --- /dev/null +++ b/src/Nodes/DOMDocumentType.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMDocumentType extends \DOMDocumentType +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMElement.php b/src/Nodes/DOMElement.php new file mode 100644 index 0000000..6ca0a29 --- /dev/null +++ b/src/Nodes/DOMElement.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMElement extends \DOMElement +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMNode.php b/src/Nodes/DOMNode.php new file mode 100644 index 0000000..79a352b --- /dev/null +++ b/src/Nodes/DOMNode.php @@ -0,0 +1,14 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +/** + * Class DOMNode + * + * @method getAttribute($attribute) + * @package andreskrey\Readability\Nodes + */ +class DOMNode extends \DOMNode +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMNotation.php b/src/Nodes/DOMNotation.php new file mode 100644 index 0000000..a4802e0 --- /dev/null +++ b/src/Nodes/DOMNotation.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMNotation extends \DOMNotation +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMProcessingInstruction.php b/src/Nodes/DOMProcessingInstruction.php new file mode 100644 index 0000000..bd80997 --- /dev/null +++ b/src/Nodes/DOMProcessingInstruction.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMProcessingInstruction extends \DOMProcessingInstruction +{ + use NodeTrait; +} diff --git a/src/Nodes/DOMText.php b/src/Nodes/DOMText.php new file mode 100644 index 0000000..43d2ba9 --- /dev/null +++ b/src/Nodes/DOMText.php @@ -0,0 +1,8 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +class DOMText extends \DOMText +{ + use NodeTrait; +} diff --git a/src/Nodes/NodeTrait.php b/src/Nodes/NodeTrait.php new file mode 100644 index 0000000..5847178 --- /dev/null +++ b/src/Nodes/NodeTrait.php @@ -0,0 +1,426 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +trait NodeTrait +{ + /** + * Content score of the node. Used to determine the value of the content. + * + * @var int + */ + public $contentScore = 0; + + /** + * Flag for initialized status. + * + * @var bool + */ + private $initialized = false; + + /** + * Flag data tables. + * + * @var bool + */ + private $readabilityDataTable = false; + + /** + * @var array + */ + private $divToPElements = [ + 'a', + 'blockquote', + 'dl', + 'div', + 'img', + 'ol', + 'p', + 'pre', + 'table', + 'ul', + 'select', + ]; + + /** + * initialized getter. + * + * @return bool + */ + public function isInitialized() + { + return $this->initialized; + } + + /** + * @return bool + */ + public function isReadabilityDataTable() + { + return $this->readabilityDataTable; + } + + /** + * @param bool $param + */ + public function setReadabilityDataTable($param) + { + $this->readabilityDataTable = $param; + } + + /** + * Initializer. Calculates the current score of the node and returns a full Readability object. + * + * @ TODO: I don't like the weightClasses param. How can we get the config here? + * + * @param $weightClasses bool Weight classes? + * + * @return static + */ + public function initializeNode($weightClasses) + { + if (!$this->isInitialized()) { + $contentScore = 0; + + switch ($this->nodeName) { + case 'div': + $contentScore += 5; + break; + + case 'pre': + case 'td': + case 'blockquote': + $contentScore += 3; + break; + + case 'address': + case 'ol': + case 'ul': + case 'dl': + case 'dd': + case 'dt': + case 'li': + case 'form': + $contentScore -= 3; + break; + + case 'h1': + case 'h2': + case 'h3': + case 'h4': + case 'h5': + case 'h6': + case 'th': + $contentScore -= 5; + break; + } + + $this->contentScore = $contentScore + ($weightClasses ? $this->getClassWeight() : 0); + + $this->initialized = true; + } + + return $this; + } + + /** + * Override for native getAttribute method. Some nodes have the getAttribute method, some don't, so we need + * to check first the existence of the attributes property. + * + * @param $attributeName string Attribute to retrieve + * + * @return string + */ + public function getAttribute($attributeName) + { + if (!is_null($this->attributes)) { + return parent::getAttribute($attributeName); + } + + return ''; + } + + /** + * Get the ancestors of the current node. + * + * @param int|bool $maxLevel Max amount of ancestors to get. False for all of them + * + * @return array + */ + public function getNodeAncestors($maxLevel = 3) + { + $ancestors = []; + $level = 0; + + $node = $this->parentNode; + + while ($node) { + $ancestors[] = $node; + $level++; + if ($level === $maxLevel) { + break; + } + $node = $node->parentNode; + } + + return $ancestors; + } + + /** + * Returns all links from the current element. + * + * @return array + */ + public function getAllLinks() + { + return iterator_to_array($this->getElementsByTagName('a')); + } + + /** + * Get the density of links as a percentage of the content + * This is the amount of text that is inside a link divided by the total text in the node. + * + * @return int + */ + public function getLinkDensity() + { + $linkLength = 0; + $textLength = mb_strlen($this->getTextContent(true)); + + if (!$textLength) { + return 0; + } + + $links = $this->getAllLinks(); + + if ($links) { + /** @var DOMElement $link */ + foreach ($links as $link) { + $linkLength += mb_strlen($link->getTextContent(true)); + } + } + + return $linkLength / $textLength; + } + + /** + * Calculates the weight of the class/id of the current element. + * + * @return int + */ + public function getClassWeight() + { + $weight = 0; + + // Look for a special classname + $class = $this->getAttribute('class'); + if (trim($class)) { + if (preg_match(NodeUtility::$regexps['negative'], $class)) { + $weight -= 25; + } + + if (preg_match(NodeUtility::$regexps['positive'], $class)) { + $weight += 25; + } + } + + // Look for a special ID + $id = $this->getAttribute('id'); + if (trim($id)) { + if (preg_match(NodeUtility::$regexps['negative'], $id)) { + $weight -= 25; + } + + if (preg_match(NodeUtility::$regexps['positive'], $id)) { + $weight += 25; + } + } + + return $weight; + } + + /** + * Returns the full text of the node. + * + * @param bool $normalize Normalize white space? + * + * @return string + */ + public function getTextContent($normalize = false) + { + $nodeValue = $this->nodeValue; + if ($normalize) { + $nodeValue = trim(preg_replace('/\s{2,}/', ' ', $nodeValue)); + } + + return $nodeValue; + } + + /** + * Returns the children of the current node. + * + * @param bool $filterEmptyDOMText Filter empty DOMText nodes? + * + * @return array + */ + public function getChildren($filterEmptyDOMText = false) + { + $ret = iterator_to_array($this->childNodes); + if ($filterEmptyDOMText) { + // Array values is used to discard the key order. Needs to be 0 to whatever without skipping any number + $ret = array_values(array_filter($ret, function ($node) { + return $node->nodeName !== '#text' || mb_strlen(trim($node->nodeValue)); + })); + } + + return $ret; + } + + /** + * Return an array indicating how many rows and columns this table has. + * + * @return array + */ + public function getRowAndColumnCount() + { + $rows = $columns = 0; + $trs = $this->getElementsByTagName('tr'); + foreach ($trs as $tr) { + /** @var \DOMElement $tr */ + $rowspan = $tr->getAttribute('rowspan'); + $rows += ($rowspan || 1); + + // Now look for column-related info + $columnsInThisRow = 0; + $cells = $tr->getElementsByTagName('td'); + foreach ($cells as $cell) { + /** @var \DOMElement $cell */ + $colspan = $cell->getAttribute('colspan'); + $columnsInThisRow += ($colspan || 1); + } + $columns = max($columns, $columnsInThisRow); + } + + return ['rows' => $rows, 'columns' => $columns]; + } + + /** + * Creates a new node based on the text content of the original node. + * + * @param $originalNode DOMNode + * @param $tagName string + * + * @return DOMElement + */ + public function createNode($originalNode, $tagName) + { + $text = $originalNode->getTextContent(); + $newNode = $originalNode->ownerDocument->createElement($tagName, $text); + + return $newNode; + } + + /** + * Check if a given node has one of its ancestor tag name matching the + * provided one. + * + * @param DOMElement $node + * @param string $tagName + * @param int $maxDepth + * + * @return bool + */ + public function hasAncestorTag($node, $tagName, $maxDepth = 3) + { + $depth = 0; + while ($node->parentNode) { + if ($maxDepth > 0 && $depth > $maxDepth) { + return false; + } + if ($node->parentNode->nodeName === $tagName) { + return true; + } + $node = $node->parentNode; + $depth++; + } + + return false; + } + + /** + * Checks if the current node has a single child and if that child is a P node. + * Useful to convert <div><p> nodes to a single <p> node and avoid confusing the scoring system since div with p + * tags are, in practice, paragraphs. + * + * @param DOMNode $node + * + * @return bool + */ + public function hasSinglePNode() + { + // There should be exactly 1 element child which is a P: + if (count($children = $this->getChildren(true)) !== 1 || $children[0]->nodeName !== 'p') { + return false; + } + + // And there should be no text nodes with real content (param true on ->getChildren) + foreach ($children as $child) { + /** @var $child DOMNode */ + if ($child->nodeType === XML_TEXT_NODE && !preg_match('/\S$/', $child->getTextContent())) { + return false; + } + } + + return true; + } + + /** + * Check if the current element has a single child block element. + * Block elements are the ones defined in the divToPElements array. + * + * @return bool + */ + public function hasSingleChildBlockElement() + { + $result = false; + if ($this->hasChildNodes()) { + foreach ($this->getChildren() as $child) { + if (in_array($child->nodeName, $this->divToPElements)) { + $result = true; + } else { + // If any of the hasSingleChildBlockElement calls return true, return true then. + /** @var $child DOMElement */ + $result = ($result || $child->hasSingleChildBlockElement()); + } + } + } + + return $result; + } + + /** + * Determines if a node has no content or it is just a bunch of dividing lines and/or whitespace. + * + * @return bool + */ + public function isElementWithoutContent() + { + return $this instanceof DOMElement && + mb_strlen(preg_replace(NodeUtility::$regexps['onlyWhitespace'], '', $this->textContent)) === 0 && + ($this->childNodes->length === 0 || + $this->childNodes->length === $this->getElementsByTagName('br')->length + $this->getElementsByTagName('hr')->length + /* + * Special PHP DOMDocument case: We also need to count how many DOMText we have inside the node. + * If there's an empty tag with an space inside and a BR (for example "<p> <br/></p>) counting only BRs and + * HRs will will say that the example has 2 nodes, instead of one. This happens because in DOMDocument, + * DOMTexts are also nodes (which doesn't happen in JS). So we need to also count how many DOMText we + * are dealing with (And at this point we know they are empty or are just whitespace, because of the + * mb_strlen in this chain of checks). + */ + + count(array_filter(iterator_to_array($this->childNodes), function ($child) { + return $child instanceof DOMText; + })) + + ); + } +} diff --git a/src/Nodes/NodeUtility.php b/src/Nodes/NodeUtility.php new file mode 100644 index 0000000..8938a49 --- /dev/null +++ b/src/Nodes/NodeUtility.php @@ -0,0 +1,155 @@ +<?php + +namespace andreskrey\Readability\Nodes; + +/** + * Class NodeUtility. + */ +class NodeUtility +{ + /** + * Collection of regexps to check the node usability. + * + * @var array + */ + public static $regexps = [ + 'unlikelyCandidates' => '/banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|foot|header|legends|menu|modal|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i', + 'okMaybeItsACandidate' => '/and|article|body|column|main|shadow/i', + 'extraneous' => '/print|archive|comment|discuss|e[\-]?mail|share|reply|all|login|sign|single|utility/i', + 'byline' => '/byline|author|dateline|writtenby|p-author/i', + 'replaceFonts' => '/<(\/?)font[^>]*>/gi', + 'normalize' => '/\s{2,}/', + 'videos' => '/\/\/(www\.)?(dailymotion|youtube|youtube-nocookie|player\.vimeo)\.com/i', + 'nextLink' => '/(next|weiter|continue|>([^\|]|$)|»([^\|]|$))/i', + 'prevLink' => '/(prev|earl|old|new|<|«)/i', + 'whitespace' => '/^\s*$/', + 'hasContent' => '/\S$/', + 'positive' => '/article|body|content|entry|hentry|h-entry|main|page|pagination|post|text|blog|story/i', + 'negative' => '/hidden|^hid$| hid$| hid |^hid |banner|combx|comment|com-|contact|foot|footer|footnote|masthead|media|meta|modal|outbrain|promo|related|scroll|share|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|tool|widget/i', + // \x{00A0} is the unicode version of + 'onlyWhitespace' => '/\x{00A0}|\s+/u' + ]; + + /** + * Imported from the Element class on league\html-to-markdown. + * + * @param $node + * + * @return DOMElement + */ + public static function nextElement($node) + { + $next = $node; + while ($next + && $next->nodeName !== '#text' + && trim($next->textContent)) { + $next = $next->nextSibling; + } + + return $next; + } + + /** + * Changes the node tag name. Since tagName on DOMElement is a read only value, this must be done creating a new + * element with the new tag name and importing it to the main DOMDocument. + * + * @param string $value + * @param bool $importAttributes + * + * @return DOMNode + */ + public static function setNodeTag($node, $value, $importAttributes = false) + { + $new = new DOMDocument('1.0', 'utf-8'); + $new->appendChild($new->createElement($value)); + + $children = $node->childNodes; + /** @var $children \DOMNodeList $i */ + for ($i = 0; $i < $children->length; $i++) { + $import = $new->importNode($children->item($i), true); + $new->firstChild->appendChild($import); + } + + if ($importAttributes) { + // Import attributes from the original node. + foreach ($node->attributes as $attribute) { + $new->firstChild->setAttribute($attribute->nodeName, $attribute->nodeValue); + } + } + + // The import must be done on the firstChild of $new, since $new is a DOMDocument and not a DOMElement. + $import = $node->ownerDocument->importNode($new->firstChild, true); + $node->parentNode->replaceChild($import, $node); + + return $import; + } + + /** + * Removes the current node and returns the next node to be parsed (child, sibling or parent). + * + * @param DOMNode $node + * + * @return DOMNode + */ + public static function removeAndGetNext($node) + { + $nextNode = self::getNextNode($node, true); + $node->parentNode->removeChild($node); + + return $nextNode; + } + + /** + * Remove the selected node. + * + * @param $node DOMElement + * + * @return void + **/ + public static function removeNode($node) + { + $parent = $node->parentNode; + if ($parent) { + $parent->removeChild($node); + } + } + + /** + * Returns the next node. First checks for children (if the flag allows it), then for siblings, and finally + * for parents. + * + * @param DOMNode $originalNode + * @param bool $ignoreSelfAndKids + * + * @return DOMNode + */ + public static function getNextNode($originalNode, $ignoreSelfAndKids = false) + { + /* + * Traverse the DOM from node to node, starting at the node passed in. + * Pass true for the second parameter to indicate this node itself + * (and its kids) are going away, and we want the next node over. + * + * Calling this in a loop will traverse the DOM depth-first. + */ + + // First check for kids if those aren't being ignored + if (!$ignoreSelfAndKids && $originalNode->firstChild) { + return $originalNode->firstChild; + } + + // Then for siblings... + if ($originalNode->nextSibling) { + return $originalNode->nextSibling; + } + + // And finally, move up the parent chain *and* find a sibling + // (because this is depth-first traversal, we will have already + // seen the parent nodes themselves). + do { + $originalNode = $originalNode->parentNode; + } while ($originalNode && !$originalNode->nextSibling); + + return ($originalNode) ? $originalNode->nextSibling : $originalNode; + } +} |