Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- <?
- function curl($url) {
- $ch = curl_init();
- curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
- curl_setopt($ch, CURLOPT_VERBOSE, 0);
- curl_setopt($ch, CURLOPT_USERAGENT, 'Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.0.3705; .NET CLR 1.1.4322)');
- curl_setopt($ch, CURLOPT_FOLLOWLOCATION, TRUE);
- curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
- curl_setopt($ch, CURLOPT_COOKIEJAR, 'cookie.txt');
- curl_setopt($ch, CURLOPT_REFERER, $referer);
- curl_setopt($ch, CURLOPT_URL, $url);
- $return = curl_exec($ch);
- return $return;
- }
- // EKSEMPEL PÅ BRUK:
- // sett $søkelink under til kategorien som trengs.
- // så er det bare å bytte ut insertProduct funksjonen helt nederst med logikken for å putte det i databasen deres.
- $søkelink = "http://www.focalprice.com/categorylist/001029?orderby=newarrivals&pagesize=72&page=1";
- $html = curl($søkelink, "http://www.focalprice.com");
- preg_match('~<span>Pages: 1/([0-9]{1,})</span>~', $html, $out);
- $pages = $out[1];
- $page = 0;
- while ($page < $pages) {
- $page++;
- $søkelink = preg_replace("/page=[0-9]{1,}/", "page=$page", $søkelink);
- echo "FOCALPRICE SØKELINK: $søkelink<br>";
- if ($page > 1) $html = curl($søkelink);
- $products = explode('<ul class="infobox">', $html);
- array_shift($products);
- while (count($products) > 72) {
- array_pop($products);
- }
- echo "found " . count($products) . " products on page $page..<br>";
- foreach ($products as $product) {
- if (!preg_match('~href="(http://www.focalprice.com/(.*)\.html)"~', $product, $out)) {
- var_dump($product);
- echo "found no url in product<br>";
- continue;
- }
- $producturl = $out[1];
- preg_match("#(.*?//.*?/.*?/)#", $producturl, $match);
- $producturl = $match[1];
- if (productExists($producturl)) {
- echo "$producturl exists. continue<br>";
- continue;
- }
- echo "Producturl[$producturl]<br>";
- $html = curl($producturl);
- preg_match_all('~(http://img.focalprice.com//550x426/(.*?)-(.*?))"~', $html, $out);
- $bilder = array();
- foreach ($out[1] as $tmp) {
- $bilder[] = str_replace("//550", "/550", $tmp);
- }
- var_dump($bilder);
- preg_match('~<h1 id="productName" class="c00">(.*?)<~s', $html, $out);
- $produktnavn = $out[1];
- preg_match('/ecomm_totalvalue: (\d{1,3}\.\d{2}),/', $html, $out);
- $kinapris = $out[1];
- preg_match('~<div class="description_m">(.*?)<div class="description_m hide">~s', $html, $out);
- $beskrivelse = $out[1];
- reporting($producturl, $bilder, $produktnavn, $kinapris, $beskrivelse);
- insertProduct(mysql_real_escape_string($produktnavn), $kinapris, $producturl, $storeurl, mysql_real_escape_string($beskrivelse), $kategori, $subkategori, $butikk, $bilder);
- $focal_product_count++;
- }
- }
- ?>
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement