Upload the script below and see what happens.
That might give you a nice start on it.
If that script yields good results, you can tweak it and work on the csv part.
PHP Code:
<?php
$url="http://www.ebay.com/sch/Diamonds-Natural-/3824/i.html?_dcat=3824&Clarity=IF%7CVVS1%7CVVS2&_ssn=gemstoneking&_pgn=1&_skc=200&rt=nc";
$data=file_get_contents($url);
// pattern
$pattern = '{<div\s+class="ittl"\s*>((?:(?:(?!<div[^>]*>|</div>).)++|<div[^>]*>(?1)</div>)*)</div>}si';
$matchcount = preg_match_all($pattern, $data, $matches);
echo("<pre>\n");
if ($matchcount > 0) {
echo("$matchcount matches found.\n");
for($i = 0; $i < $matchcount; $i++) {
echo("\nMatch #" . ($i + 1) . ":\n");
echo($matches[1][$i]);
}
} else {
echo('No matches');
}
echo("\n</pre>");
?>
This is not my script, by the way ... I could never figure out that regex by myself.
Just something I remembered that I had in my pile of snippets.
The key to the script, is the fact that all descriptions are contained in this tag: <div class="ittl"> blah blah blah </div>
.