kmoroz / wanakanashaapu

🐊 C# port of WanaKana JS 🦀

ワナカナ ↔️ WanaKana ↔️ わなかな

🐊🦀 Basic Overview

WanaKanaShaapu is a utility library for converting Japanese characters to the Latin ones and detecting the Japanese language in a given input. The C# version is ported from WanaKana JS v5.0.0.

🐊🦀 Original Documentation

WanaKana API Documentation

🐊🦀 Differences in Implementation

Method	WanaKana JS	WanaKanaShaapu
IsMixed	isMixed(input, { passKanji: true })	IsMixed(input, passKanji)
StripOkurigana	stripOkurigana(input, { leading: false, matchKanji: '' })	StripOkurigana(input, leading, matchKanji)
Tokenize	1. tokenize(input, { detailed: true, compact: false }) 2. Returns either a string array or an object	1. Tokenize(input, compact) 2. Returns Tokenization object

wanakanashaapu's Issues

"chuu" is translated into "ちゅ" instead of "ちゅう"

Hi,

When trying to transcript "chuu" into hiragana, I get the result "ちゅ" instead of the expected "ちゅう".

This seems to happen when adding the "ya" parts after consonants by calling this method inTreeBuilder.cs during the call line 70:

private static void AddNode(Node node, string romaji, string kana)
{
    if (romaji.Length == 1 && !node.Children.ContainsKey(romaji))
        node.Children.Add(romaji, new Node(kana));
    else if (romaji.Length == 1 && node.Children.ContainsKey(romaji))
        node.Children[romaji].Data = kana;
    else
    {
        foreach (char c in romaji)
        {
            if (!node.Children.ContainsKey(c.ToString()))
                node.Children.Add(c.ToString(), new Node(string.Empty));
            node = node.Children[c.ToString()];
            AddNode(node, romaji[1..], kana);
        }
    }
}

When trying to add ya in the k node, we get to foreach loop:

Current node is `k`, romaji = "ya", kana = "きゃ", current character is "y"
`k -> y` does not exist
    we create `k -> y`
node = `k -> y`
    ->  Add node recursively (node: `k -> y`, romaji: "a", kana: "きゃ")
        romaji.Length is 1 and `k -> y` does not have any child yet
            `k -> y -> a` is created with value "きゃ"
    <-

// next loop, node `k -> y`, romaji = "ya", kana = "きゃ", current character is "a"
`k -> y -> a` exists
node = `k -> y -> a`
    -> Add node recursively (node: `k -> y -> a`, romaji: "a", kana: "きゃ")
        romaji.Length is 1 and `k -> y -> a` does not have any child yet
            `k -> y -> a -> a` is created with value "きゃ"

It seems that there is a mix of recursion and iteration here which mixes things up.

I can see two solutions to that problem.

By fixing the iterations:

private static void AddNode(Node node, string romaji, string kana)
{
    if (romaji.Length == 1 && !node.Children.ContainsKey(romaji))
        node.Children.Add(romaji, new Node(kana));
    else if (romaji.Length == 1 && node.Children.ContainsKey(romaji))
        node.Children[romaji].Data = kana;
    else
    {
        var remaining = romaji;
        foreach (char c in romaji)
        {
            if (!node.Children.ContainsKey(c.ToString()))
                node.Children.Add(c.ToString(), new Node(string.Empty));
            node = node.Children[c.ToString()];
            AddNode(node, remaining[1..], kana);
            remaining = remaining[1..];
        }
    }
}

By fixing the recursion: (this seems more elegant to me)

private static void AddNode(Node node, string romaji, string kana)
{
    if (romaji.Length == 0)
        return;
    else if (romaji.Length == 1 && !node.Children.ContainsKey(romaji))
        node.Children.Add(romaji, new Node(kana));
    else if (romaji.Length == 1 && node.Children.ContainsKey(romaji))
        node.Children[romaji].Data = kana;
    else
    {
        var firstChar = romaji.First().ToString();
        if (!node.Children.ContainsKey(firstChar))
            node.Children.Add(firstChar.First().ToString(), new Node(string.Empty));
        node = node.Children[firstChar.ToString()];
        AddNode(node, romaji[1..], kana);
    }
}