/**
* Prepares the given HTTP-query-string for the HTTP-request.
*
* HTTP-query-strings always should be utf8-encoded and urlencoded afterwards.
* So "/path/file?test=tatütata" will be converted to "/path/file?test=tat%C3%BCtata":
*
* @param stirng The quetry-string (like "/path/file?test=tatütata")
* @return string
*/
protected function prepareHTTPRequestQuery($query)
{
// If string already is a valid URL -> do nothing
if (PHPCrawlerUtils::isValidUrlString($query)) {
return $query;
}
// Decode query-string (for URLs that are partly urlencoded and partly not)
$query = rawurldecode($query);
// if query is already utf-8 encoded -> simply urlencode it,
// otherwise encode it to utf8 first.
if (PHPCrawlerEncodingUtils::isUTF8String($query) == true) {
$query = rawurlencode($query);
} else {
$query = rawurlencode(utf8_encode($query));
}
// Replace url-specific signs back
$query = str_replace("%2F", "/", $query);
$query = str_replace("%3F", "?", $query);
$query = str_replace("%3D", "=", $query);
$query = str_replace("%26", "&", $query);
return $query;
}
/**
* Reconstructs a full qualified and normalized URL from a given link relating to the URL the link was found in.
*
* @param string $link The link (i.e. "../page.htm")
* @param PHPCrawlerUrlPartsDescriptor $BaseUrl The base-URL the link was found in as PHPCrawlerUrlPartsDescriptor-object
*
* @return string The rebuild, full qualified and normilazed URL the link is leading to (i.e. "http://www.foo.com/page.htm"),
* or NULL if the link couldn't be rebuild correctly.
*/
public static function buildURLFromLink($link, PHPCrawlerUrlPartsDescriptor $BaseUrl)
{
$url_parts = $BaseUrl->toArray();
// Dedoce HTML-entities
$link = PHPCrawlerEncodingUtils::decodeHtmlEntities($link);
// Remove anchor ("#..."), but ONLY at the end, not if # is at the beginning !
$link = preg_replace("/^(.{1,})#.{0,}\$/", "\\1", $link);
// Cases
// Strange link like "//foo.htm" -> make it to "http://foo.html"
if (substr($link, 0, 2) == "//") {
$link = "http:" . $link;
} elseif (substr($link, 0, 1) == "/") {
$link = $url_parts["protocol"] . $url_parts["host"] . ":" . $url_parts["port"] . $link;
} elseif (substr($link, 0, 2) == "./") {
$link = $url_parts["protocol"] . $url_parts["host"] . ":" . $url_parts["port"] . $url_parts["path"] . substr($link, 2);
} elseif (preg_match("#^[a-z0-9-]{1,}(:\\/\\/)# i", $link)) {
$link = $link;
} elseif (preg_match("/^[a-zA-Z]{0,}:[^\\/]{0,1}/", $link)) {
$link = "";
} elseif (substr($link, 0, 3) == "../") {
$new_path = $url_parts["path"];
while (substr($link, 0, 3) == "../") {
$new_path = preg_replace('/\\/[^\\/]{0,}\\/$/', "/", $new_path);
$link = substr($link, 3);
}
$link = $url_parts["protocol"] . $url_parts["host"] . ":" . $url_parts["port"] . $new_path . $link;
} elseif (substr($link, 0, 1) == "#") {
$link = "";
} elseif (substr($link, 0, 1) == "?") {
$link = $url_parts["protocol"] . $url_parts["host"] . ":" . $url_parts["port"] . $url_parts["path"] . $url_parts["file"] . $link;
} else {
$link = $url_parts["protocol"] . $url_parts["host"] . ":" . $url_parts["port"] . $url_parts["path"] . $link;
}
if ($link == "") {
return null;
}
// Now, at least, replace all HTMLENTITIES with normal text.
// I.E.: HTML-Code of the link is: <a href="index.php?x=1&y=2">
// -> Link has to be "index.php?x=1&y=2"
//$link = PHPCrawlerEncodingUtils::decodeHtmlEntities($link);
// Replace linebreaks in the link with "" (happens if a link in the sourcecode
// linebreaks)
$link = str_replace(array("\n", "\r"), "", $link);
// "Normalize" URL
$link = self::normalizeUrl($link);
return $link;
}