Getting image source in php is extremely slow - php

Question

I need to get the image src based on the class of the image.
This is the code I wrote.
It works but it is extremely slow.
$url='https://' . $_SERVER['HTTP_HOST'].$_SERVER['REQUEST_URI'];
$html= file_get_contents($url);
$doc = new DOMDocument();
libxml_use_internal_errors(true);
$doc->loadHTML($html);
$xpath = new DOMXPath($doc);
$nodes = $xpath->query("//img[#class='imgbanner']");
if ($nodes->length > 0) {
$src = $nodes->item(0)->getAttribute('src');
}
else {
$src = null;
}
Any clues on how to improve speed?

Answer 1

Assuming you're trying to parse a somewhat convoluted HTML document, and especially considering your rather limited use case, you might be better off resorting to regular expressions and some string parsing (again, in this concrete circumstances, cf. this post's closing remarks).
For testing purposes, let's set up an HTML document with 10,000 image tags, each of them looking like this one:
<img class="imgbanner" src="a49851fb74.jpg">
To benchmark both approaches more easily (XPath vs. regular expression + string parsing), let's wrap them in two functions (the first one is pretty much the same as the sample code you've provided):
function xpath(string $html): array {
libxml_use_internal_errors(true);
$doc = new DOMDocument();
$doc->loadHTML($html);
$xpath = new DOMXPath($doc);
$nodes = $xpath->query("//img[#class='imgbanner']");
$src = [];
if ($nodes) {
foreach ($nodes as $node) {
$src[] = $node->getAttribute('src');
}
}
libxml_clear_errors(); // Free up memory
return $src;
}
function regex(string $html): array {
preg_match_all("/<img[^>]+src=[\"']([^\"']+)[\"'][^>]*>/i", $html, $matches);
$matches = array_combine($matches[0], $matches[1]);
$filtered = [];
foreach ($matches as $key => $value) {
if (strpos($key, 'class="imgbanner"') || strpos($key, "class='imgbanner'")) {
$filtered[] = $value;
}
}
return $filtered;
}
Since the HTML document doesn't contain much else but the image tags, XPath is pretty fast (~0.06 seconds over the course of ten runs):
$start = microtime(true);
$html = file_get_contents('pics.html'); // 10,000 random image tags
$src = xpath($html);
$time_elapsed_secs = (microtime(true) - $start);
echo "Total execution time: {$time_elapsed_secs}\n"; // ~0.06 sec
Nevertheless, the second approach turned out to be about ten times faster (~0.005 seconds over the course of ten runs):
$start = microtime(true);
$html = file_get_contents('pics.html'); // 10,000 random image tags
$src = regex($html);
$time_elapsed_secs = (microtime(true) - $start);
echo "Total execution time: {$time_elapsed_secs}\n"; // ~0.005 sec
While the second approach is obviously faster for this very limited use case, bear in mind it's usually a bad idea to parse HTML using regular expressions:
https://blog.codinghorror.com/parsing-html-the-cthulhu-way/
If your parsing needs grow in complexity (i.e., anything above the case at hand), you should consider factoring out the parsing into a dedicated command line script and cache its results.