. * * Returns a boolean indicator whether the rendering of the full text * in the available space succeeded or not. * * @param ezcDocumentPdfPage $page * @param ezcDocumentPdfHyphenator $hyphenator * @param ezcDocumentPdfTokenizer $tokenizer * @param ezcDocumentLocateableDomElement $text * @param ezcDocumentPdfMainRenderer $mainRenderer * @return bool */ public function renderNode( ezcDocumentPdfPage $page, ezcDocumentPdfHyphenator $hyphenator, ezcDocumentPdfTokenizer $tokenizer, ezcDocumentLocateableDomElement $text, ezcDocumentPdfMainRenderer $mainRenderer ) { // Inference page styles $styles = $this->styles->inferenceFormattingRules( $text ); $width = $page->innerWidth / $styles['text-columns']->value - ( $styles['text-column-spacing']->value * ( $styles['text-columns']->value - 1 ) ); // Evaluate available space if ( ( $space = $this->evaluateAvailableBoundingBox( $page, $styles, $width ) ) === false ) { return false; } // Iterate over tokens and try to fit them in the current line, use // hyphenator to split words. $tokens = $this->tokenize( $text, $tokenizer ); $lines = $this->fitTokensInLines( $tokens, $hyphenator, $space->width ); // Evaluate required space by text box $required = 0; foreach ( $lines as $line ) { $required += $line['height'] * $styles['line-height']->value; } // Check that enough space is available to render text box if ( $required > $space->height ) { return false; } $space->height = $required; $this->renderBoxBackground( $space, $styles ); $this->renderBoxBorder( $space, $styles ); $this->renderTextBox( $lines, $space, $styles ); $this->setBoxCovered( $page, $space, $styles ); return true; } /** * Render text box * * Render a single text box, specified by the given lines array, * containing tokens and their styles, the available space and * the styles array for the currently rendered element. * * Returns false, if the box size was not sufficant for the * given text, and the covered vertical area otherwise. * * @param array $lines * @param ezcDocumentPdfBoundingBox $space * @param array $styles * @return boolean */ protected function renderTextBox( array $lines, ezcDocumentPdfBoundingBox $space, array $styles ) { $yPos = $space->y; foreach ( $lines as $nr => $line ) { $yPos += $this->renderLine( $yPos, $nr, $line, $space, $styles ) * $styles['line-height']->value; // Check if we run out of vertical space if ( $yPos > ( $space->y + $space->height ) ) { return false; } } return $yPos - $space->y; } /** * Reverses a string * * Similar to PHPs strrev() function, but also works for UTF-8 strings. * * @param string $string * @return string */ protected function strrev( $string ) { if ( !is_string( $string ) || empty( $string ) ) { return $string; } if ( strlen( $string ) === ( $length = iconv_strlen( $string, 'UTF-8' ) ) ) { // String only contains of single-byte characters return strrev( $string ); } $reverted = ''; for ( $c = $length; $c > 0; --$c ) { $reverted .= iconv_substr( $string, $c - 1, 1, 'UTF-8' ); } return $reverted; } /** * Render a single line and return the used height * * @param float $position * @param int $number * @param array $line * @param ezcDocumentPdfBoundingBox $space * @param array $styles * @return void */ protected function renderLine( $position, $number, array $line, ezcDocumentPdfBoundingBox $space, array $styles ) { $spaceWidth = $this->driver->calculateWordWidth( ' ' ); $lineWidth = 0; foreach ( $line['tokens'] as $token ) { if ( !is_int( $token['word'] ) ) { $lineWidth += $token['width']; } } // Reverse alignement, if direction is set to "right-to-left" $align = $styles['text-align']->value; if ( $styles['direction']->value === 'rtl' ) { switch ( $align ) { case 'right': $align = 'left'; default: $align = 'right'; } } switch ( $align ) { case 'center': $offset = ( $space->width - $lineWidth - ( $line['spaces'] * $spaceWidth ) ) / 2; break; case 'right': $offset = $space->width - $lineWidth - ( $line['spaces'] * $spaceWidth ); break; case 'justify': $offset = 0; switch ( true ) { case $number === $line['words']: // Just use common space width in last line of a // paragraph $spaceWidth = $this->driver->calculateWordWidth( ' ' ); break; case $line['words'] <= 1: // Space width is irrelevant, if only one token is // in the line break; default: $spaceWidth = ( $space->width - $lineWidth ) / ( $line['spaces'] - 1 ); } default: $offset = 0; } // Reverse tokens and words, if direction is set to "right-to-left" $tokens = $line['tokens']; if ( $styles['direction']->value === 'rtl' ) { $tokens = array_reverse( $tokens ); foreach ( $tokens as $nr => $token ) { if ( is_string( $token['word'] ) ) { $tokens[$nr]['word'] = $this->strrev( $token['word'] ); } } } // Default to left alignement $xPos = $space->x + $offset; foreach ( $tokens as $token ) { if ( $token['word'] === ezcDocumentPdfTokenizer::SPACE ) { $this->renderTextDecoration( $token['style'], $xPos, $position, $spaceWidth, $line['height'] ); $this->handleLinks( $token, $xPos, $position, $spaceWidth, $line['height'] ); $xPos += $spaceWidth; } else if ( is_string( $token['word'] ) ) { $this->renderTextDecoration( $token['style'], $xPos, $position, $token['width'], $line['height'] ); // Apply current styles foreach ( $token['style'] as $style => $value ) { $this->driver->setTextFormatting( $style, $value->value ); } // Render word $this->driver->drawWord( $xPos, $position + $line['height'], $token['word'] ); $this->handleLinks( $token, $xPos, $position, $token['width'], $line['height'] ); $xPos += $token['width']; } } return $line['height']; } /** * Handle links * * Handle embedded link markup for current token and perform the * appropriate calls to the driver. * * @param array $token * @param float $x * @param float $y * @param float $width * @param float $height * @return void */ protected function handleLinks( array $token, $x, $y, $width, $height ) { if ( $token['url'] !== null ) { $this->driver->addExternalLink( $x, $y, $width, $height * 1.1, $token['url'] ); } if ( $token['target'] !== null ) { $this->driver->addInternalLink( $x, $y, $width, $height * 1.1, $token['target'] ); } } /** * Render text decoration * * Render text decoration, like by a assigned text-decoration setting, or * background-colors, and similar. * * @param array $styles * @param float $x * @param float $y * @param float $width * @param float $height * @return void */ protected function renderTextDecoration( array $styles, $x, $y, $width, $height ) { // Directly exit, if there are no decorations to render if ( ( $styles['text-decoration']->value === 'none' ) && ( !isset( $styles['background-color'] ) || ( $styles['background-color']->value['alpha'] >= 1 ) ) ) { return; } if ( isset( $styles['background-color'] ) && ( $styles['background-color']->value['alpha'] < 1 ) ) { $this->driver->drawPolygon( array( array( $x, $y ), array( $x + $width, $y ), array( $x + $width, $y + $height * $styles['line-height']->value ), array( $x, $y + $height * $styles['line-height']->value ), ), $styles['background-color']->value ); } if ( strpos( $styles['text-decoration'], 'line-through' ) !== false ) { $this->driver->drawPolyline( array( array( $x, $y + $height - $styles['font-size']->value / 3 ), array( $x + $width, $y + $height - $styles['font-size']->value / 3 ), ), $styles['color']->value, // @todo: How thick should line-throughs be? ezcDocumentPcssMeasure::create( '1px' )->get(), false ); } if ( strpos( $styles['text-decoration'], 'overline' ) !== false ) { $this->driver->drawPolyline( array( array( $x, $y ), array( $x + $width, $y ), ), $styles['color']->value, // @todo: How thick should overlines be? ezcDocumentPcssMeasure::create( '1px' )->get(), false ); } if ( strpos( $styles['text-decoration'], 'underline' ) !== false ) { $this->driver->drawPolyline( array( array( $x, $y + $height * 1.1 ), array( $x + $width, $y + $height * 1.1 ), ), $styles['color']->value, // @todo: How thick should underlines be? ezcDocumentPcssMeasure::create( '1px' )->get(), false ); } } /** * Tokenize the input string * * For proper word wrapping in the paragraph the strng needs to be * tokenized, while each token has to maintain its stack of assigned * formats. * * This method should return an array of tokens, also maintaining the * included whitespace characters, each associated with its markup * elements. * * @param ezcDocumentLocateableDomElement $element * @param ezcDocumentPdfTokenizer $tokenizer * @param bool $recursed * @return array */ protected function tokenize( ezcDocumentLocateableDomElement $element, ezcDocumentPdfTokenizer $tokenizer, $recursed = false ) { $tokens = array(); $rules = $this->styles->inferenceFormattingRules( $element, ezcDocumentPcssStyleInferencer::TEXT ); // Do not inherit background and border rules from paragraph if ( !$recursed ) { $rules = array_diff_key( $rules, array( 'background-color' => true, 'border' => true, ) ); } $url = $element->tagName === 'ulink' && $element->hasAttribute( 'url' ) ? $element->getAttribute( 'url' ) : null; $target = $element->tagName === 'link' && $element->hasAttribute( 'linked' ) ? $element->getAttribute( 'linked' ) : null; foreach ( $element->childNodes as $child ) { switch ( $child->nodeType ) { // case XML_CDATA_SECTION_NODE: case XML_TEXT_NODE: $words = $tokenizer->tokenize( $child->textContent ); foreach ( $words as $word ) { $tokens[] = array( 'word' => $word, 'style' => $rules, 'url' => $url, 'target' => $target, ); } break; case XML_ELEMENT_NODE: $tokens = array_merge( $tokens, $this->tokenize( $child, $tokenizer, true ) ); break; } } if ( !$recursed ) { // Remove double spaces foreach ( $tokens as $nr => $token ) { if ( ( $token['word'] === ezcDocumentPdfTokenizer::SPACE ) && isset( $tokens[$nr + 1] ) && ( $tokens[$nr + 1]['word'] === ezcDocumentPdfTokenizer::SPACE ) ) { $i = 1; do { unset( $tokens[$nr + $i] ); } while ( isset( $tokens[$nr + ( ++$i )] ) && ( $tokens[$nr + $i]['word'] === ezcDocumentPdfTokenizer::SPACE ) ); } } $tokens = array_values( $tokens ); // Remove optional starting spaces if ( count( $tokens ) && ( $tokens[0]['word'] === ezcDocumentPdfTokenizer::SPACE ) ) { $tokens = array_slice( $tokens, 1 ); } } return $tokens; } /** * Force split a word. * * Force the splitting of a word, which did not fit in a line alone and * could not be splitted using the hyphenator. We just search for the * maximum word part length which fits the available space. * * Could be improved to use a binary search on the word length, but this * shouldn't happen too often anyways. * * @param string $word * @param float $available * @return array */ protected function forceSplit( $word, $available ) { $length = iconv_strlen( $word ) - 1; while ( $this->driver->calculateWordWidth( iconv_substr( $word, 0, $length ) ) > $available ) { --$length; } return array( iconv_substr( $word, 0, $length ), iconv_substr( $word, $length ) ); } /** * Try to match tokens into lines * * Try to match tokens into lines of the given width. Returns an array with * words for each line. The words might already be split up by the * hyphenator. * * @param array $tokens * @param ezcDocumentPdfHyphenator $hyphenator * @param float $available * @return array */ protected function fitTokensInLines( array $tokens, ezcDocumentPdfHyphenator $hyphenator, $available ) { $lines = array( array( 'tokens' => array(), 'height' => 0, 'words' => 0, 'spaces' => 0, ) ); $line = 0; $consumed = 0; while ( $token = array_shift( $tokens ) ) { // Handle forced line breaks if ( $token['word'] === ezcDocumentPdfTokenizer::FORCED ) { // Continue rendering in next line $consumed = 0; $lines[++$line] = array( 'tokens' => array(), 'height' => 0, 'words' => 0, 'spaces' => 0, ); continue; } // Pure wrapping tokens are irrleveant to width calculation if ( $token['word'] === ezcDocumentPdfTokenizer::WRAP ) { continue; } // Apply current styles foreach ( $token['style'] as $style => $value ) { // Only pass whitelist of properties to the backend. This is not really // the right place to do this, but it reduces the amount of calls to // the backend and the size of the call log massively. if ( !( ( $style === 'font-style' ) || ( $style === 'font-weight' ) || ( $style === 'color' ) || ( $style === 'font-size' ) || ( $style === 'font-family' ) ) ) { continue; } $this->driver->setTextFormatting( $style, $value->value ); } // Just add space to consumed wird width if ( $token['word'] === ezcDocumentPdfTokenizer::SPACE ) { if ( $consumed > 0 ) { $consumed += $width = $this->driver->calculateWordWidth( ' ' ); $token['width'] = $width; $lines[$line]['tokens'][] = $token; $lines[$line]['spaces']++; } continue; } $wordStack = array( 'tokens' => array(), 'height' => 0, 'words' => 0, 'spaces' => 0, ); $wConsumed = 0; do { if ( ( $consumed + $wConsumed + ( $width = $this->driver->calculateWordWidth( $token['word'] ) ) ) < $available ) { // The word just fits into the current line $token['width'] = $width; $wordStack['tokens'][] = $token; $wordStack['height'] = max( $lines[$line]['height'], $this->driver->getCurrentLineHeight() ); $wordStack['words']++; $wConsumed += $width; if ( !isset( $tokens[0] ) || ( $tokens[0]['word'] === ezcDocumentPdfTokenizer::WRAP ) || ( $tokens[0]['word'] === ezcDocumentPdfTokenizer::FORCED ) || ( $tokens[0]['word'] === ezcDocumentPdfTokenizer::SPACE ) ) { // We are allowed to wrap, so we can continue with the // next iteration and merge the current word stack with // the line array. $lines[$line]['tokens'] = array_merge( $lines[$line]['tokens'], $wordStack['tokens'] ); $lines[$line]['height'] = max( $lines[$line]['height'], $wordStack['height'] ); $lines[$line]['words'] += $wordStack['words']; $consumed += $wConsumed; continue 2; } continue; } // Try to hyphenate the current word $hyphens = array_reverse( $hyphenator->splitWord( $token['word'] ) ); foreach ( $hyphens as $hyphen ) { if ( ( $consumed + $wConsumed + ( $width = $this->driver->calculateWordWidth( $hyphen[0] ) ) ) < $available ) { $second = $token; $second['word'] = $hyphen[1]; array_unshift( $tokens, $second ); $token['width'] = $width; $token['word'] = $hyphen[0]; $lines[$line]['tokens'] = array_merge( $lines[$line]['tokens'], $wordStack['tokens'], array( $token ) ); $lines[$line]['height'] = max( $lines[$line]['height'], $wordStack['height'], $this->driver->getCurrentLineHeight() ); $lines[$line]['words'] += $wordStack['words'] + 1; // Continue rendering in next line $consumed = 0; $lines[++$line] = array( 'tokens' => array(), 'height' => 0, 'words' => 0, 'spaces' => 0, ); continue 3; } } if ( ( $consumed + $wConsumed ) <= 0 ) { // If we are already at the beginning of the line, and the // word does still not fit, we forcefully split the word. $hyphen = $this->forceSplit( $token['word'], $available ); $second = $token; $second['word'] = $hyphen[1]; array_unshift( $tokens, $second ); $token['width'] = $this->driver->calculateWordWidth( $hyphen[0] ); $token['word'] = $hyphen[0]; $lines[$line]['tokens'] = array( $token ); $lines[$line]['height'] = $this->driver->getCurrentLineHeight(); $lines[$line]['words'] = 1; // Continue rendering in next line $consumed = 0; $lines[++$line] = array( 'tokens' => array(), 'height' => 0, 'words' => 0, 'spaces' => 0, ); continue 2; } // Did not fit using hyphenation, so retry after wrapping // the current word stack into the next line. array_unshift( $tokens, $token ); // Still does not fit, move whole word stack into the next line // and try again $token['width'] = $width = $this->driver->calculateWordWidth( $token['word'] ); $lines[++$line] = $wordStack; $consumed = $wConsumed; continue 2; } while ( $token = array_shift( $tokens ) ); } return $lines; } } ?>