[RESOLVED] extract the source of an <img> tag from html ?

steamPunk

I was wondering if any of you have an idea for how I can do this :

Imagine I have a chunk of html like this :

<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Phasellus ultricies tempus neque quis fermentum. Suspendisse viverra tincidunt lorem, at commodo nunc aliquam in. Morbi ipsum metus, placerat vitae accumsan ut, viverra eu nulla. Mauris interdum mollis vulputate. 
<img src="http://www.mysite.com/images/myImage.jpg"  class="" />
Sed et massa vel enim sagittis condimentum. Fusce molestie nisl vitae risus accumsan imperdiet. Mauris non fringilla quam. Nunc ultrices lacinia justo et gravida. Proin vel mi vel erat adipiscing lacinia vitae vitae libero. Integer sollicitudin sem eu diam ultricies viverra. Ut ultrices vulputate purus, at dapibus ligula dictum hendrerit. Mauris tincidunt blandit neque.</p>

what I need to do is

find if there's an image in the html
get the source of the image
remove all tags from the html (I think there's a function that does that)
get the size of the image so I can reduce the displayed size if it's wider than 200px (I know I can do this with getimagesize() )
get a substring of of the text
re-display the html like this

<img src="http://www.mysite.com/images/myImage.jpg"  class="" />
<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Phasellus ultricies tempus neque quis fermentum. Suspendisse viverra ... <br />
<a href="[url]">Click for more info</a></p>

the problem I have is how to find if there's an image in the html and then how to extract the image src attribute

is there a php function that will do this ?

thanks

bradgrafelman

If you've only got the chunk of HTML (e.g. not an actual HTML document, in which case I would suggest using [man]DOM[/man]), you could try using [man]preg_match/man (or [man]preg_match_all/man) with a regexp such as:

/<img [^>]*src=["'](.+?)["'][^>]*>/i

steamPunk

yes - that's good - thanks 🙂

when I tested it the result was it captured the whole <img> tag like this :

<img src="http://www.mysite.com/images/myImage.jpg" class="img-left" style="width:200px;" />

but I just needed the src, so I did a bit of research and came across named subpatterns, which I'd never heard of, like this :

if (preg_match("/<img [^>]*src=[\"'](?P<url>.+?)[\"'][^>]*>/i", $str, $matches)) {
    echo "A match was found: ".$matches['url'];
} else {
    echo "A match was not found.";
}

which outputs :

http://www.mysite.com/images/myImage.jpg

now this is very nice, but what I'd really like would be to get just the filename (and not the whole path with the http etc) so that I can use file_exists() with the server path to the images directory

is there a simple-ish way to change the regex so that it will do this ?
or should I just use explode("/",$matches['url']) to get the filename ?

steamPunk

actually, this is not elegant, but works ok

if (preg_match("/<img [^>]*src=[\"'](?P<url>.+?)[\"'][^>]*>/i", $str, $matches)) {
    $f = explode("/",$matches['url']);
    $c = array_reverse($f);
    echo "A match was found: ".$c[0];
} else {
    echo "A match was not found.";
}

so that's nice !

thanks for your help

bradgrafelman

For one, while named subpatterns might make the resulting code easier to read ($matches['url'] does indeed look prettier than $matches[1]), they are certainly not necessary; my regexp pattern above already groups the URL in a subgroup, meaning you simply need to access the appropriate numerical index (rather than just $matches[0] which I'm betting you were doing before).

As for the 'filename' part, you might try using [man]parse_url/man to get the PHP_URL_PATH component from the URL and then use something like [man]basename/man to get the "filename". Note that I say "filename" because URLs really don't have filenames - a URL has no direct correlation to a file path on a filesystem, thus there is no such thing as the "filename" part of a URL.

steamPunk

bradgrafelman;10991547 wrote:
my regexp pattern above already groups the URL in a subgroup, meaning you simply need to access the appropriate numerical index

ok - I hadn't realized that - that's great!

bradgrafelman;10991547 wrote:
As for the 'filename' part, you might try using [man]parse_url/man to get the PHP_URL_PATH component from the URL and then use something like [man]basename/man to get the "filename". Note that I say "filename" because URLs really don't have filenames - a URL has no direct correlation to a file path on a filesystem, thus there is no such thing as the "filename" part of a URL.

I see what you mean and I will try that

thanks again