PHP simplexml_load_file with special chars in URL - php

Question

I'm attempting to retrieve a local weather forecast, based on the IP of the user.
I'm using geoplugin.net to get the user location and feed the city and country name to the Google Weather API.
//Get user IP
$ip = $_SERVER['REMOTE_ADDR'];
$geolocation = unserialize(file_get_contents('http://www.geoplugin.net/php.gp?ip='.$ip));
$geo_city = $geolocation['geoplugin_city'];
$geo_country = $geolocation['geoplugin_countryName'];
$file = "http://www.google.com/ig/api?weather=".$geo_city.",".$geo_country;
$xml = simplexml_load_file($file);
//Echo content of retrieved XML for debugging purposes
echo "<pre>";
print_r($xml);
echo "</pre>";
It works well for most cases, but when I try it on my own IP, I get Søborg, Denmark (which is not 100% accurate, but close enough) and that gives me an almost empty response from the weather API.
The main suspect in the case, is the dastardly "ø"-character.
The XML that I want can be seen here: http://www.google.com/ig/api?weather=S%C3%B8borg,Denmark
The XML that I'm getting can be seen here: http://www.google.com/ig/api?weather=S
When I type this URL into the browser it works fine:
http://www.google.com/ig/api?weather=Søborg,Denmark
When I use this version it works as well (in the browser):
http://www.google.com/ig/api?weather=S%C3%B8borg,Denmark
but this version returns the forecast for Borg,Syddanmark:
http://www.google.com/ig/api?weather=S%26oslash%3Bborg,Denmark
None of the above returns the desired result, when fed to the simplexml_load_file().
As stated, I suspect that it is a character set issue, but I can't figure out what to do about it.
What is the correct way to solve it?
I know that I can use latitude and longtitude as parameters for Google Weather API instead, but that's just circumventing the problem, not solving it.

Answer 1

If you URL-decode S%26oslash%3Bborg you'll see that this string corresponds to Søborg which gives us Søborg after we decode HTML entities like so:
$city = 'S%26oslash%3Bborg,Denmark';
echo $city = rawurldecode($city);
//prints Søborg,Denmark
echo $city = html_entity_decode($city, 0, 'UTF-8');
//prints Søborg,Denmark
echo $city = rawurlencode($city);
//prints S%C3%B8borg%2CDenmark
And then:
$xml = file_get_contents('http://www.google.com/ig/api?weather='.$city);
$xml = mb_convert_encoding($xml, 'UTF-8');
$xml = simplexml_load_string($xml);
echo $xml->weather->forecast_information->city['data'];
Outputs expected:
Søborg, Capital Region of Denmark

Answer 2

It does indeed sound like a character set issue. Have you tried converting the URL to another encoding, e.g. using iconv, before passing the result into simplexml_load_file()?

Answer 3

Try this out:
$file = "http://www.google.com/ig/api?weather=" . $geo_city . "," . $geo_country;
$data = file_get_contents($file);
$data = mb_convert_encoding($data, "UTF-8", "ISO-8859-2");
$xml = simplexml_load_string($data);
echo "<pre>"; print_r($xml); echo "</pre>";
It's taken from this maybe similar thread: https://stackoverflow.com/a/5136549/949476