North Wales Tech JVM Languages tasks

northwalestech.kt

import com.eclipsesource.json.Json
import com.eclipsesource.json.JsonObject
import com.eclipsesource.json.JsonValue

// Recursively get JSON values for all properties in a JSON object with the given key name
fun getAllWithName(jsonObj : JsonObject, keyName : String) : ArrayList<JsonValue> {
    val keys = jsonObj.names()
    return keys.fold(ArrayList<JsonValue>()) { acc, key ->
        val value = jsonObj.get(key)
        if (key == keyName) {
            acc.add(value)
        } else if (value.isObject) {
            acc.addAll(getAllWithName(value.asObject(), keyName))
        } else if (value.isArray) {
            val items = value.asArray()
            acc.addAll(items.filter { item -> item.isObject }
                            .map { item -> getAllWithName(item.asObject(), keyName) }
                            .flatten())
        }
        acc
    }
}

fun main(args: Array<String>) {
    if (args.size != 1) throw IllegalArgumentException("Too many arguments")

    val pageTitle = args[0]
    val apiUrl = "https://en.wikipedia.org/w/api.php?action=query&prop=revisions&rvprop=content&format=json&titles=$pageTitle"

    val jsonText = java.net.URL(apiUrl).readText()
    val jsonObj = Json.parse(jsonText).asObject()
    val texts = getAllWithName(jsonObj, "*").map(JsonValue::asString)

    val wordsRegex = Regex("\\p{L}+") // All letters as defined by Unicode
    val wordsMap = texts.fold(HashMap<String, Int>()) { map, text ->
        val words = wordsRegex.findAll(text).map { x -> x.value.toLowerCase() }
        words.forEach { word -> map[word] = map.getOrDefault(word, 0) + 1 }
        map
    }

    val allWords = wordsMap.keys
    val lengthOfLongestWords = allWords.fold(-1) { maxLength, word -> if (word.length > maxLength) word.length else maxLength }
    val longestWords = allWords.filter { word -> word.length == lengthOfLongestWords }

    val wordsToExclude = sortedSetOf( "the", "to", "of", "and", "a", "is", "that", "in", "be", "for")
    val mostCommonWords = allWords
            .filter { word -> !wordsToExclude.contains(word.toLowerCase()) }
            .sortedByDescending { word -> wordsMap.getOrDefault(word, -1) }
            .take(10)

    println("Longest word(s): ${longestWords.joinToString(", ")}")
    println("Most common words: ${mostCommonWords.joinToString(", ")}")
}