20,
// Whether to check the required amount of words before the first item.
'require_before_first' => false,
// Whether to check the required amount of words after the last item.
'require_after_last' => false,
'debug' => false,
'before' => false,
'paragraph_select_from_bottom' => false,
);
/**
* Amount of words between items.
*
* @var array
*/
protected $words_between = array();
/**
* Previous offset.
*
* @var false/int
*/
protected $previous_offset = false;
const START_EXISTING_AD = 'advads_amount_of_words_s';
const END_EXISTING_AD = 'advads_amount_of_words_e';
const INSERTION_POINT = 'advads_amount_of_words_i';
const SPLIT_REGEXP = '/(advads_amount_of_words_.)/';
const PREV_WORDS = 'prev_words';
const PREV_NUMBER = 'prev_number';
const NEXT_WORDS = 'next_words';
const NEXT_NUMBER = 'next_number';
const PREV_IS_AD = 'prev_is_ad';
const NEXT_IS_AD = 'next_is_ad';
/**
* Create an object of the class.
*
* @param string $html HTML string.
* @param array $options Options.
* @return object Object of this class.
*/
public static function from_html( $html, $options ) {
$libxml_previous_state = libxml_use_internal_errors( true );
$dom = new DOMDocument( '1.0', 'UTF-8' );
$success = $dom->loadHtml( '' . $html );
libxml_clear_errors();
libxml_use_internal_errors( $libxml_previous_state );
$xpath = new DOMXPath( $dom );
return new static( $dom, $xpath, $options );
}
/**
* Constructor.
*
* @param DOMDocument $dom DOMDocument object.
* @param DOMXPath $xpath DOMXpath object.
* @param array $options Options.
*/
public function __construct( DOMDocument $dom, DOMXPath $xpath, array $options ) {
$this->dom = $dom;
$this->xpath = $xpath;
$this->options = array_merge( $this->default_options, $options );
}
/**
* Prepare HTML for parsing.
*
* @param array $items Existing selected items.
* @return string $r String with injected patterns.
*/
private function prepare_for_parsing( $items ) {
$expr = $this->get_expression_for_existing_ads();
$existing = $this->xpath->query( $expr );
$created = array();
// Excel existing ads.
foreach ( $existing as $node ) {
$start = $this->dom->createTextNode( self::START_EXISTING_AD );
$node->parentNode->insertBefore( $start, $node ); // phpcs:ignore WordPress.NamingConventions.ValidVariableName.UsedPropertyNotSnakeCase
$created[] = $start;
$end = $this->dom->createTextNode( self::END_EXISTING_AD );
$node->parentNode->insertBefore( $end, $node->nextSibling ); // phpcs:ignore WordPress.NamingConventions.ValidVariableName.UsedPropertyNotSnakeCase
$created[] = $end;
}
// Excel items, e.g. paragraphs.
foreach ( $items as $node ) {
$point = $this->dom->createTextNode( self::INSERTION_POINT );
$created[] = $point;
if ( $this->options['before'] ) {
$node->parentNode->insertBefore( $point, $node ); // phpcs:ignore WordPress.NamingConventions.ValidVariableName.UsedPropertyNotSnakeCase
} else {
$node->parentNode->insertBefore( $point, $node->nextSibling ); // phpcs:ignore WordPress.NamingConventions.ValidVariableName.UsedPropertyNotSnakeCase
}
}
// Select all text nodes.
$nodes = $this->xpath->query( '//text()[not(parent::script or parent::style)]' );
$r = '';
foreach ( $nodes as $n ) {
$r .= $n->data;
}
foreach ( $created as $node ) {
$node->parentNode->removeChild( $node ); // phpcs:ignore WordPress.NamingConventions.ValidVariableName.UsedPropertyNotSnakeCase
}
return $r;
}
/**
* Prepare the array that contains calculated amount of words between items.
*
* @param str $parts Post text splited by regexp.
* @return $r Word amounts between items.
*/
protected function prepare_words_between( $parts ) {
$r = array();
$prev_is_ad = false;
$doing_ad = 0;
$prev_text = array();
$prev_ip = false;
foreach ( $parts as $part ) {
if ( self::START_EXISTING_AD === $part ) {
if ( $prev_ip ) {
$r = $this->add_after_insertion_point( $r, $prev_text, true );
$prev_ip = false;
}
$prev_text = array();
$prev_is_ad = true;
$doing_ad++;
continue;
} elseif ( self::END_EXISTING_AD === $part ) {
$doing_ad--;
continue;
} elseif ( $doing_ad ) {
continue;
} elseif ( self::INSERTION_POINT === $part ) {
if ( $prev_ip ) {
$r = $this->add_after_insertion_point( $r, $prev_text );
$prev_ip = false;
}
$data = array(
self::PREV_NUMBER => self::calc_words( $prev_text ),
self::PREV_IS_AD => $prev_is_ad,
);
if ( $this->options['debug'] ) {
$data[ self::PREV_WORDS ] = $prev_text;
}
$r[] = $data;
$prev_text = array();
$prev_is_ad = false;
$prev_ip = true;
} else {
$prev_text[] = self::prepare_text( $part );
}
}
if ( $prev_ip ) {
$r = $this->add_after_insertion_point( $r, $prev_text );
$prev_ip = false;
}
return $r;
}
/**
* Add words after insertion points.
*
* @param array $r Word amounts between items.
* @param array $prev_text Word amounts between items.
* @param bool $next_is_ad Whether or an ad is placed after the point.
* @return $r Word amounts between items.
*/
private function add_after_insertion_point( $r, $prev_text, $next_is_ad = false ) {
$el = array_pop( $r );
$el[ self::NEXT_NUMBER ] = self::calc_words( $prev_text );
$el[ self::NEXT_IS_AD ] = $next_is_ad;
if ( $this->options['debug'] ) {
$el[ self::NEXT_WORDS ] = $prev_text;
}
$r[] = $el;
return $r;
}
/**
* Get amount of words between items.
*
* @param string/array $items An array of `DOMElement`s or `XPath` query.
* @return array
*/
public function get_words_between( $items ) {
if ( is_string( $items ) ) {
$items = iterator_to_array( $this->xpath->query( $items ) );
}
$text = $this->prepare_for_parsing( $items );
$parts = preg_split( self::SPLIT_REGEXP, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
$this->words_between = $this->prepare_words_between( $parts );
return $this->words_between;
}
/**
* Calculate new offsets.
*
* They may be shifted to provide the minimum amount of words before and after or deleted,
* if shifting is not possible
*
* @param array $offsets Existing offsets.
* @param string/array $items An array of `DOMElement`s or `XPath` query.
* @return array $offsets New offsets.
*/
public function calc_offsets( array $offsets, $items ) {
$words_between = $this->get_words_between( $items );
$from_bottom = $this->options['paragraph_select_from_bottom'];
if ( $from_bottom ) {
rsort( $offsets );
} else {
sort( $offsets );
}
$new_offsets = array();
$this->previous_offset = false;
$this->debug( '///// debug start: ' . implode( ', ', $offsets ) . ' /////' );
foreach ( $offsets as $k => $offset ) {
if ( $from_bottom ) {
$new_offset = $this->calc_offset_from_bottom( $offset );
} else {
$new_offset = $this->calc_offset( $offset );
}
if ( false === $new_offset ) {
break;
}
$new_offsets[] = $new_offset;
}
$this->debug( '///// debug end /////' );
return $new_offsets;
}
/**
* Calculate offset.
*
* @param int $offset Offset.
* @return int/false New shifted offset or false.
*/
private function calc_offset( $offset ) {
$l = count( $this->words_between );
if ( false !== $this->previous_offset ) {
$offset = max( $offset, $this->previous_offset + 1 );
}
$this->debug( "///// offset: $offset /////" );
for ( $i = $offset; $i < $l; $i++ ) {
if ( ! isset( $this->words_between[ $i ] ) ) {
continue;
}
$prev_num = 0;
$next_num = 0;
$prev_has_ad = false;
for ( $j = $i; $j >= 0; $j-- ) {
if ( $this->options['debug'] ) {
$this->debug( implode( ' / ', $this->words_between[ $j ][ self::PREV_WORDS ] ) );
}
$prev_num += $this->words_between[ $j ][ self::PREV_NUMBER ];
if ( $this->words_between[ $j ][ self::PREV_IS_AD ] ) {
$prev_has_ad = true;
break;
}
}
$this->debug( $i . ': prev: ' . $prev_num );
if ( $prev_num < $this->options['words_between_repeats']
// Check if this is the first item and there are no ads before it.
&& ( $this->options['require_before_first'] || $prev_has_ad || false !== $this->previous_offset ) ) {
continue;
}
$next_has_ad = false;
for ( $k = $i; $k < $l; $k++ ) {
$next_num += $this->words_between[ $k ][ self::NEXT_NUMBER ];
if ( $this->words_between[ $k ][ self::NEXT_IS_AD ] ) {
$next_has_ad = true;
break;
}
}
if ( $next_num < $this->options['words_between_repeats']
// Check if there are no ads after the last item.
&& ( $this->options['require_after_last'] || $next_has_ad ) ) {
continue;
}
$this->previous_offset = $i;
$this->words_between[ $i ][ self::NEXT_IS_AD ] = true;
if ( isset( $this->words_between[ $i + 1 ] ) ) {
$this->words_between[ $i + 1 ][ self::PREV_IS_AD ] = true;
}
$this->debug( "found $i" );
return $i;
}
return false;
}
/**
* Calculate offset from bottom.
*
* @param int $offset Offset.
* @return int/false New shifted offset or false.
*/
private function calc_offset_from_bottom( $offset ) {
$l = count( $this->words_between );
if ( false !== $this->previous_offset ) {
$offset = min( $offset, $this->previous_offset - 1 );
}
$this->debug( "///// offset: $offset /////" );
for ( $i = $offset; $i >= 0; $i-- ) {
if ( ! isset( $this->words_between[ $i ] ) ) {
continue;
}
$prev_num = 0;
$next_num = 0;
$prev_has_ad = false;
for ( $j = $i; $j < $l; $j++ ) {
if ( $this->options['debug'] ) {
$this->debug( implode( ' / ', $this->words_between[ $j ][ self::NEXT_WORDS ] ) );
}
$prev_num += $this->words_between[ $j ][ self::NEXT_NUMBER ];
if ( $this->words_between[ $j ][ self::NEXT_IS_AD ] ) {
$prev_has_ad = true;
break;
}
}
$this->debug( $i . ': prev: ' . $prev_num );
if ( $prev_num < $this->options['words_between_repeats']
// Check if this is the first item and there are no ads before it.
&& ( $this->options['require_before_first'] || $prev_has_ad || false !== $this->previous_offset ) ) {
continue;
}
$next_has_ad = false;
for ( $k = $i; $k >= 0; $k-- ) {
$next_num += $this->words_between[ $k ][ self::NEXT_NUMBER ];
if ( $this->words_between[ $k ][ self::PREV_IS_AD ] ) {
$next_has_ad = true;
break;
}
}
if ( $next_num < $this->options['words_between_repeats']
// Check if there are no ads after the last item.
&& ( $this->options['require_after_last'] || $next_has_ad ) ) {
continue;
}
$this->previous_offset = $i;
$this->words_between[ $i ][ self::PREV_IS_AD ] = true;
if ( isset( $this->words_between[ $i - 1 ] ) ) {
$this->words_between[ $i - 1 ][ self::NEXT_IS_AD ] = true;
}
$this->debug( "found $i" );
return $i;
}
return false;
}
/**
* Get xpath expression for selecting existing ads.
*
* @return string XPath expression.
*/
protected function get_expression_for_existing_ads() {
$expr = array(
// The assumption is that a `div` that has a class starting with the frontend prefix is ad.
"//div[@class and contains(concat(' ', normalize-space(@class), ' '), ' %s')]",
// Waiting for consent ads (Privacy module): `