Question

I am trying to get the source for images on some pages but there are some differences between the code of the two pages.

Page 1 Code:

<img class="thumb thumb_0" onclick="setImage(0); return false;" src="http://example.com/b1.jpg">

Page 2 Code:

<img style="width: 46px ! important; height: 46px ! important;" class="thumb thumb_0" onclick="setImage(0); return false;" src="http://example.com/image4.jpg">

Notice the difference between the 2 pages… Page 2 has a stupid style at the beginning of the img tag. Also, the “onclick” is located in a different position. The only thing I need to snag is the image location.

Here is the code that I have thus far… which only works for page 1 scenario:

preg_match_all("/<img\s*?class='thumb.*?'.*?src='(.*?)'.*?\/>/is", $hotelPage, $thumbs, PREG_PATTERN_ORDER);

Ideally, I would like to be able to keep it in one php line. How can I do an “or” in preg_replace and how can I get the regex to work for page 2 as well?

Thank you in advance!

UPDATE: The pages have other images, I am only looking for the ones that have a class that contains “thumb”. I apologize for leaving out that heavy detail.

Answer 1

This should work as you intended – if your html is in $html, the regular expression should look like the one $reg :

$html='some html <img class="thumb thumb_0" onclick="setImage(0); return false;"
   src="http://example.com/b1.jpg"> xxx yyy <img class="bummer thumb_0"
   onclick="setImage(0); return false;" src="http://example.com/bummer.jpg">
   <img style="width: 46px ! important; height: 46px ! important;"
   class="thumb thumb_0" onclick="setImage(0); return false;"
   src="http://example.com/image4.jpg"> some html';

$reg = ' <img .+?                # img tag
         class="thumb .+?        # class tag
         src="([^"]+)            # capture src
       ';

preg_match_all("/$reg/xis", $html, $thumbs, PREG_SET_ORDER);

foreach($thumbs as $t) echo $t[1]."\n";

It matches only if the order of attributes is {class, src} and if it found both the img-tag, and the correct class “thumb”. Here we go:

http://example.com/b1.jpg
http://example.com/image4.jpg

Only two of three img entries match (I included a third, wrong link in the test set).

Regards

rbo