Regular expression to replace an <a> with respective <img>

Question

I'm looking for a PHP preg_replace() solution find links to images and replace them with respective image tags.

Find:

<a href="http://www.domain.tld/any/valid/path/to/imagefile.ext"&gt;This will be ignored.</a>

Replace with:

<img src="http://www.domain.tld/any/valid/path/to/imagefile.ext" alt="imagefile" />

Where the protocol MUST be http://, the .ext MUST be a valid image format (.jpg, .jpeg, .gif, .png, .tif), and the base file name becomes the alt="" value.

I know preg_replace() is the right function for the job, but I suck with regex, so any help is greatly appreciated! THANKS!

Answer 1

Ahh, my daily DOM practice. You should use DOM to parse HTML and regex to parse strings such as html attributes.

Note: I have some basic regexes that could surely be improved upon by some wizards :)

Note #2: Though it might be extra overhead you could use something like curl to thoroughly check if the href is an actual image by sending a HEAD request and looking at the Content-Type, but this would work in 80-90% of cases.

<?php

$content = '

<a href="http://www.domain.tld/any/valid/path/to/imagefile.ext"&gt;This will be ignored.</a>
<br>

<a href="http://col.stb.s-msn.com/i/43/A4711309495C88F8CD154C99FCE.jpg"&gt;this will not be ignored</a>

<br>

<a href="http://col.stb.s-msn.com/i/A0/8E9A454F701E4F5F89E58E14B532C.jpg"&gt;bah&lt;/a&gt;
';

$dom = new DOMDocument();
$dom->loadHTML($content);

$anchors = $dom->getElementsByTagName('a');

$i = $anchors->length-1;

$protocol = '/^http:\/\//';
$ext = '/([\w+]+)\.(?:gif|jpg|jpeg|png)$/';

if ( count($anchors->length) > 0 ) {
    while( $i > -1 ) {
    $anchor = $anchors->item($i);
    if ( $anchor->hasAttribute('href') ) {
        $link = $anchor->getAttribute('href');

        if ( 
     preg_match ( $protocol , $link ) &&
     preg_match ( $ext, $link )
        ) {
     //echo 'replacing this one.';
     $image = $dom->createElement('img');

     if ( preg_match( $ext, $link, $matches ) ) {
         if ( count($matches) ) {
      $altName = $matches[1];
      $image->setAttribute('alt', $altName);
         }
         $image->setAttribute('src', $link);
         $anchor->parentNode->replaceChild( $image, $anchor );
     }
        }

    }
    $i--;
    }
}

echo $dom->saveHTML();

Answer 2

Congratulations, you are the one millionth customer to ask Stack Overflow how to parse HTML with regex!

[X][HT]ML is not a regular language and cannot reliably be parsed with regex. Use an HTML parser. PHP itself gives you DOMDocument, or you may prefer simplehtmldom.

Incidentally, you cannot tell what type a file is by looking at its URL. There is no reason a JPEG has to have ‘.jpeg’ as its extension — and indeed, no guarantee that a file with ‘.jpeg’ extension will actually be JPEG. The only way to be certain is to fetch the resource (eg. using a HEAD request) and look at the Content-Type header.

Answer 3

I would suggest using this more flexible non-greddy regex:

<a[^>]+?href=\"(http:\/\/[^\"]+?\/([^\"]*?)\.(jpg|jpeg|png|gif))[^>]*?>[^<]*?<\/a>

And a more complex regex (including PHP test code) to hopefully please Gumbo :)

<?php
$test_data = <<<END
<a blabla="asldlsaj" alksjada="aslkdj" href="http://www.domain.tld/any/valid/path/to/imagefile.jpg" lkjasd=""asdlaskjd>This will be ignored.</a>
Lorem ipsum..
<a    blabla=asldlsaj alksjada="aslkdj" href="http://www.domain.tld/any/valid/path/to/imagefile.jpg" lkjasd=""asdlaskjd>This will be ignored.</a>
<a lkjafs='asdsa> ' blabla="asldlksjada=>"aslkdj" href="http://www.domain.tld/any/valid/path/to/imagefile.jpg" lkjasd=""asdlaskjd>This will be ignored.</a>
<a    blabla="ajada="aslk href="http://www.domain.tld/any/valid/path&gt;/to/imagefile.jpg" lkjasd>asdlaskjd>This will be ignored.</a>
<a    blabla="asldlsaj>" aslkdj href="http://www.domain.tld/any/valid/path/ to/imagefile.jpg" lkjasd=""asdlaskjd>This will be ignored.</a>
Something:
<a    blabla='asldls<ajslkdj' href="http://www.domain.tld/any/valid'/path/to/imagefile.jpg" lkjasd=""asdlaskjd>This will be ignored.</a>
<a    blabla=  asldlsadj href="http://www.domain.tld/any/valid/path/to/imagefile.jpg" lkjasd>This will be ignored.</a>
<a blabla="asldlsaj" alksjslkdj" href='http://www.domain.tld/any/valid/path/to/imagefile.jpg' lkjasdskjd>This will be ignored.</a>
Something else...
<a    blabla="asldlsaj" alksjslkdj" href='http://www.domain.tld/any/valid/path/to/imagefile.jpg' lkjasdskjd>This will be ignored.</a>
<a    blabla="asldlsaj" alksjada="aslkdj" href=http://www.domain.tld/any/valid/path/to/imagefile.jpg lkjdlaskjdll> be ignored.</a>
END;
$regex = "/<a\s(\s*\w+(\s*=\s*(\".*?\"|'.*?'|[^'\">\s]+))?)+?\s+href\s*=\s*(\"(http:\/\/[^\"]+\/(.*?)\.(jpg|jpeg|png|gif))\"|'(http:\/\/[^']+\/(.*?)\.(jpg|jpeg|png|gif))'|(http:\/\/[^'\">\s]+\/([^'\">\s]+)\.(jpg|jpeg|png|gif)))\s(\s*\w+(\s*=\s*(\".*?\"|'.*?'|[^'\">\s]+))?)+>[^<]*?<\/a>/i";
$replaced = preg_replace($regex, '<img src="$5$8$11" alt="$6$9$12" />', $test_data);

echo '<pre>'.htmlentities($replaced);
?>