php only remove wrapper tag - php

Question

How can I remove only the wrapper tag with preg_replace.
For example: I want to remove p tag from this:
$html = "<p><div><p>aaaaaa</p></div></p>";
Output should be: <div><p>aaaaaa</p></div>
If input is
$html = "<p>aaaaaa</p><div>bbbb</div>";
Output should be: <p>aaaaaa</p><div>bbbb</div>
I tried using this regex: '/<p[^>]*>(.*)<\/p[^>]*>/i' but it replaced all p tags.

Answer 1

Here is a regex approach using a recursive pattern.
Code: (Demo)
$htmls = [
"<p><div><p>aaaaaa</p></div></p>",
"<div><p>aaaaaa</p></div>",
"<p>aaaaaa</p><div>bbbbbb</div>",
"<p>aaaaaa</p><div>bbbbbb</div><p>cccccc</p>",
"<p>aaaaaa</p><p>bbbbbb</p>",
"<p>hello<p>aaaaaa</p></p>",
"<p><p>aaaaaa</p></p>"
];
foreach ($htmls as $i => $html) {
$without_ptags = preg_replace('~<p>(?:(?R)|.*?)*</p>~', '', $html,2, $count);
if ($without_ptags === '' && $count == 1) {
echo "$i => ", substr($html, 3, -4);
}else{
echo "$i => not wrapped in p tags";
}
echo "\n---\n";
}
Output:
0 => <div><p>aaaaaa</p></div>
---
1 => not wrapped in p tags
---
2 => not wrapped in p tags
---
3 => not wrapped in p tags
---
4 => not wrapped in p tags
---
5 => hello<p>aaaaaa</p>
---
6 => <p>aaaaaa</p>
---
*Note Parsing HTML with regex is not recommended. If I can come up with a clever DomDocument approach, I'll add it to my answer.
Until then, my code uses a recursive pattern to replace <p>...</p> substrings with an empty string. (Pattern Demo) preg_replace() stores the number of replacements made in $count. If the output string is completely empty and $count is 1 then it can be reasoned that the html string was fully nested in a single, parent <p> tag. After making this determination, substr() is used to remove the leading <p> and the trailing </p>. *note: A replacement limit of 2 is used because 2 or more replacements constitutes a disqualified html string regardless of the output to $without_ptags.