parse a HTML string
| public parseString ( string $html ) : string | ||
| $html | string | |
| return | string | markdown formatted |
/**
* @inheritdoc
*/
public function transform($value)
{
if (is_null($value)) {
return $value;
}
if (is_scalar($value)) {
$value = (string) $value;
}
if (!is_string($value)) {
throw new TransformationFailedException(sprintf('Expected a string to transform, got %s instead', json_encode($value)));
}
// replace non-breaking spaces, somehow this results in a question mark when markdownifying
$value = str_replace([' ', " "], ' ', $value);
// remove leading spaces/tabs
$value = preg_replace('/^[ \\t]+/m', '', $value);
// purify the html first
$value = $this->purifier->purify($value);
// perform some replacements...
$replacements = [[['/>\\s+</', '/\\s+<\\//'], ['><', '</']], [['/\\s+<br\\/?>/', '/<br\\/?>\\s+/'], '<br>'], ['/([^>])\\n([^<])/', '\\1<br>\\2'], ['/(<(p|li)>)<br\\s?\\/?>/i', '\\1'], ['/<br\\s?\\/?>(<\\/(p|li)>)/i', '\\1']];
foreach ($replacements as list($search, $replace)) {
$value = preg_replace($search, $replace, $value);
}
// strip tags in headings
foreach (range(1, 6) as $headingSize) {
$value = preg_replace_callback('/(<h' . $headingSize . '>)(.*)(<\\/h' . $headingSize . '>)/iU', function ($matches) {
if (count($matches) !== 4) {
return $matches[0];
}
return $matches[1] . trim(strip_tags(str_replace('<br>', ' ', $matches[2]))) . $matches[3];
}, $value);
}
// remove any double bullets
$value = preg_replace('/(<li>\\s*)[\\*|\\-]{1}/im', '\\1', $value);
// convert to markdown
$value = @$this->converter->parseString($value);
// Fix different types of bullets. What this does is check each line if it starts with any of "-ו○",
// not followed by another bullet, and normalizes it to "* text".
$value = preg_replace('/^[\\-ו○]\\s*([^\\-ו○])/mu', '* $1', $value);
// Now make sure there's a newline before 2 consecutive lines that start with a bullet.
// This could lead to superfluous newlines, but they will be corrected later on.
$value = preg_replace('/(\\n\\* [^\\n]+){2,}/', "\n\$0", "\n" . $value);
// remove trailing spaces/tabs
$value = preg_replace('/[ \\t]+$/m', '', $value);
// remove excessive newlines
$value = preg_replace('/\\n{3,}/m', "\n\n", $value);
return trim($value);
}
/**
* @inheritdoc
*/
public function transform($value)
{
if (is_null($value)) {
return $value;
}
if (is_scalar($value)) {
$value = (string) $value;
}
if (!is_string($value)) {
throw new TransformationFailedException(sprintf('Expected a string to transform, got %s instead', json_encode($value)));
}
// replace non-breaking spaces, somehow this results in a question mark when markdownifying
$value = str_replace([' ', " "], ' ', $value);
// remove leading spaces/tabs
$value = preg_replace('/^[ \\t]+/m', '', $value);
// purify to remove really obscure html
$value = $this->purifier->purify($value);
// remove whitespace/newlines between tags: this can cause trailing
// whitespace after markdownifying
$value = preg_replace(['/>\\s+</', '/\\s+<\\//'], ['><', '</'], $value);
// also remove whitespace/newlines around <br> tags
$value = preg_replace(['/\\s+<br\\/?>/', '/<br\\/?>\\s+/'], '<br>', $value);
// Replace newlines with <br> if the newline is not between 2 tags
$value = preg_replace('/([^>])\\n([^<])/', '\\1<br>\\2', $value);
// Remove <br>'s at the beginning of a paragraph
$value = preg_replace('/(<(p|li)>)<br\\s?\\/?>/i', '\\1', $value);
// Remove <br>'s at the end of a paragraph
$value = preg_replace('/<br\\s?\\/?>(<\\/(p|li)>)/i', '\\1', $value);
// replace •-bullets
$value = preg_replace('/•/', '*', $value);
// convert to markdown
$value = @$this->converter->parseString($value);
// remove trailing spaces/tabs
$value = preg_replace('/[ \\t]+$/m', '', $value);
// remove excessive newlines
$value = preg_replace('/\\n{3,}/m', "\n\n", $value);
return $value;
}
public function testResetState()
{
// Broken (unclosed) tags cause properties (such as indents) to run onto subsequent strings,
$blockquote = 'Test blockquote <blockquote>Here it is';
$linebreaks = 'Test<br /><br />Linebreaks';
$converter = new Converter();
$bqOutput = $converter->parseString($blockquote);
$this->assertContains('>', $bqOutput);
$lbOutput = $converter->parseString($linebreaks);
$this->assertNotContains('>', $lbOutput);
}
