Parsing JSON Lines with Alamofire/Codable - ios

Question

Is it possible to parse JSON lines with Alamofire and codable?
Here is my code right now.
Alamofire.request(url, method: .get, parameters: parameters, encoding: URLEncoding.default, headers: headers).responseString {(response) in
switch response.result {
case .success(let value):
print ("response is \(value)")
case .failure(let error):
print ("error is \(error)")
}
}
This prints all the JSON lines as a string but I want to serialize the response as an array of JSON. How would I do that? The problem with JSON lines is that it returns each set of json on a separate line so it is not familiar to alamofire.
Here is what I tried as if this was traditional JSON which obviously it is not so this did not work:
Alamofire.request(url, method: .get, parameters: parameters, encoding: URLEncoding.default, headers: headers).responseJSON {(response) in
switch response.result {
case .success(let value):
print ("response is \(value)")
let decoder = JSONDecoder()
decoder.dateDecodingStrategy = .secondsSince1970
let data = try! JSONSerialization.data(withJSONObject: value)
do {
let logs = try decoder.decode([Logs].self, from: data)
completion(logs)
} catch let error {
print ("error parsing get logs: \(error)")
}
case .failure(let error):
print ("failed get logs: \(error) ** \(response.result.value ?? "")")
}
}
For anybody unfamiliar with json lines here is the official format info: http://jsonlines.org
{"_logtype":"syslogline","_ingester":"agent","_ip":"40.121.203.183","pid":5573,"program":"docker","_host":"k8s-master-5A226838-0","logsource":"k8s-master-5A226838-0","_app":"syslog","_file":"/var/log/syslog","_line":"docker[5573]: I0411 00:18:39.644199 6124 conversion.go:134] failed to handle multiple devices for container. Skipping Filesystem stats","_ts":1491869920198,"timestamp":"2017-04-11T00:18:39.000Z","_id":"804760774821019649"}
{"_logtype":"syslogline","_ingester":"agent","_ip":"40.121.203.183","pid":5573,"program":"docker","_host":"k8s-master-5A226838-0","logsource":"k8s-master-5A226838-0","_app":"syslog","_file":"/var/log/syslog","_line":"docker[5573]: I0411 00:18:39.644167 6124 conversion.go:134] failed to handle multiple devices for container. Skipping Filesystem stats","_ts":1491869920198,"timestamp":"2017-04-11T00:18:39.000Z","_id":"804760774821019648"}
{"_logtype":"syslogline","_ingester":"agent","_ip":"40.121.203.183","pid":5573,"program":"docker","_host":"k8s-master-5A226838-0","logsource":"k8s-master-5A226838-0","_app":"syslog","_file":"/var/log/syslog","_line":"docker[5573]: I0411 00:18:37.053730 6124 operation_executor.go:917] MountVolume.SetUp succeeded for volume \"kubernetes.io/secret/6f322c04-e1d2-11e6-bca0-000d3a111245-default-token-swb07\" (spec.Name: \"default-token-swb07\") pod \"6f322c04-e1d2-11e6-bca0-000d3a111245\" (UID: \"6f322c04-e1d2-11e6-bca0-000d3a111245\").","_ts":1491869917193,"timestamp":"2017-04-11T00:18:37.000Z","_id":"804760762212941824"}

Answer 1

Here is an example of writing custom DataSerializer in Alamofire that you could use for decoding your Decodable object.
I am using an example from random posts json url https://jsonplaceholder.typicode.com/posts
Here is an example of Post class and custom serializer class PostDataSerializer.
struct Post: Decodable {
let userId: Int
let id: Int
let title: String
let body: String
}
struct PostDataSerializer: DataResponseSerializerProtocol {
enum PostDataSerializerError: Error {
case InvalidData
}
var serializeResponse: (URLRequest?, HTTPURLResponse?, Data?, Error?) -> Result<[Post]> {
return { request, response, data, error in
if let error = error {
return .failure(error)
}
guard let data = data else {
return .failure(PostDataSerializerError.InvalidData)
}
do {
let jsonDecoder = JSONDecoder()
let posts = try jsonDecoder.decode([Post].self, from: data)
return .success(posts)
} catch {
return .failure(error)
}
}
}
}
You could simply hook this upto your Alamofire client which sends request to remote url like so,
let request = Alamofire.request("https://jsonplaceholder.typicode.com/posts")
let postDataSerializer = PostDataSerializer()
request.response(responseSerializer: postDataSerializer) { response in
print(response)
}
You could also do additional error checking for the error and http response code in serializeResponse getter of the custom serializer.
For you json line, it seems that each json object is separated with new line character in so called json line format. You could simply split the line with new line characters and decode each line to Log.
Here is Log class I created.
struct Log: Decodable {
let logType: String
let ingester: String
let ip: String
let pid: Int
let host: String
let logsource: String
let app: String
let file: String
let line: String
let ts: Float64
let timestamp: String
let id: String
enum CodingKeys: String, CodingKey {
case logType = "_logtype"
case ingester = "_ingester"
case ip = "_ip"
case pid
case host = "_host"
case logsource
case app = "_app"
case file = "_file"
case line = "_line"
case ts = "_ts"
case timestamp
case id = "_id"
}
}
And custom log serializer that you could use with your Alamofire. I have not handled error in the following serializer, I hope you can do it.
struct LogDataSerializer: DataResponseSerializerProtocol {
enum LogDataSerializerError: Error {
case InvalidData
}
var serializeResponse: (URLRequest?, HTTPURLResponse?, Data?, Error?) -> Result<[Post]> {
return { request, response, data, error in
if let error = error {
return .failure(error)
}
guard let data = data else {
return .failure(LogDataSerializerError.InvalidData)
}
do {
let jsonDecoder = JSONDecoder()
let string = String(data: data, encoding: .utf8)!
let allLogs = string.components(separatedBy: .newlines)
.filter { $0 != "" }
.map { jsonLine -> Log? in
guard let data = jsonLine.data(using: .utf8) else {
return nil
}
return try? jsonDecoder.decode(Log.self, from: data)
}.flatMap { $0 }
return .success(allLogs)
} catch {
return .failure(error)
}
}
}
}

Answer 2

Alamofire is extensible, so I'd suggest writing your own response ResponseSerializer that can parse JSON by line. It seems that each line should parse fine, they're just not parseable together since the whole document isn't valid JSON.