Umlaut in URL fails Codable (Swift) - what to do? - json

Question

Using Swift-4.1, Xcode-9.3.1, iOS-11.3.1
I use the Codable protocol to decode a JSON-file. Everything works, except until the moment where I have an Internationalised Domain-Name (in this case, with a German Umlaut "ä") in a URL (example: http://www.rhätische-zeitung.ch).
This leads to a decoder-error inside the following code:
func loadJSON(url: URL) -> Media? {
do {
let data = try Data(contentsOf: url)
let decoder = JSONDecoder()
let media = try decoder.decode(Media.self, from: data)
return media
} catch {
print("error:\(error)")
}
return nil
}
The error-message is:
The Codable protocol does not seem to be able to decode this URL form
my JSON-file into the needed Struct.
Here is the Struct:
struct Media: Codable {
var publisher: [MediaPublisher]
}
struct MediaPublisher: Codable {
var title: String?
var homepage_url: URL?
}
And here is the JSON-excerpt:
{
"publisher": [
{
"title" : "Rhätische-Zeitung",
"homepage_url" : "http://www.rhätische-zeitung.ch",
}
]
}
Since the JSON-file is coming from the outside, I have no control over the content. And therefore, replacing the URL inside the JSON is not an option !
(Therefore, I cannot replace the URL inside the JSON to an accepted Internationalized Form suche as: www.xn--rhtische-zeitung-wnb.ch) !!
I know that there are techniques to place a custom initialiser into the Struct-definition (see my trials below...) - but since new to Codable, I don't know how to do that for this current URL-Umlaut problem. The custom-initialiser I placed below does return nil for the URL at question. What do I need to change ??
Or is there another way of making this JSON-decoding of an URL with Umlaut work ??
Here is the Struct, this time with a custom initialiser:
(at least with this, I can get rid of the error-message above... But the URL is now nil it seems and that is not what I want either)
struct Media: Codable {
var publisher: [MediaPublisher]
}
struct MediaPublisher: Codable {
var title: String?
var homepage_url: URL?
// default initializer
init(title: String?, homepage_url: URL?) {
self.title = title
self.homepage_url = homepage_url
}
// custom initializer
init(from decoder: Decoder) throws {
let map = try decoder.container(keyedBy: CodingKeys.self)
self.title = try? map.decode(String.self, forKey: .title)
self.homepage_url = try? map.decode(URL.self, forKey: .homepage_url)
}
private enum CodingKeys: CodingKey {
case title
case homepage_url
}
}

Answer 1

I just stumbled over the same problem. The solution is actually fairly simple and works well in my brief testing.
The idea is to not let the decoder decode the value as an URL, because it expects the string behind it to be in of a certain format. What we can do to circumvent this is to decode the value as a string directly and convert that manually into a URL.
I wrote a little extension that does that job for me:
extension KeyedDecodingContainer {
/// Decodes the string at the given key as a URL. This allows for special characters like umlauts to be decoded correctly.
func decodeSanitizedURL(forKey key: KeyedDecodingContainer<K>.Key) throws -> URL {
let urlString = try self.decode(String.self, forKey: key)
// Sanitize string and attempt to convert it into a valid url
if let urlString = urlString.addingPercentEncoding(withAllowedCharacters: .urlQueryAllowed), let url = URL(string: urlString) {
return url
}
// Throw an error as the URL could not be decoded
else {
throw DecodingError.dataCorruptedError(forKey: key, in: self, debugDescription: "Could not decode \(urlString)")
}
}
}
This allows for a streamlined use in the init method.
required init(from decoder: Decoder) throws {
let values = try decoder.container(keyedBy: CodingKeys.self)
self.url = try values.decodeSanitizedURL(forKey: .url)
}
Hope that helps, even if the question is a little bit older.