Hy2026/backend/Application/Text/KeywordExtractor.cs

using System.Text;

namespace TarninaHub.Application.Text;

public static class KeywordExtractor
{
    private static readonly HashSet<string> StopWords = new(StringComparer.OrdinalIgnoreCase)
    {
        "i", "w", "we", "na", "z", "do", "nie", "że", "się", "jest", "to", "dla", "od", "po",
        "o", "a", "ale", "jak", "co", "oraz", "przez", "przy", "tak", "już", "czy", "ich",
        "the", "and", "for", "with", "this", "that", "are", "was",
    };

    public static IReadOnlyList<string> Extract(string text, int max = 8)
    {
        if (string.IsNullOrWhiteSpace(text) || max <= 0)
        {
            return [];
        }

        var counts = new Dictionary<string, int>(StringComparer.OrdinalIgnoreCase);
        var order = new List<string>();

        foreach (var rawToken in Tokenize(text))
        {
            var token = rawToken.ToLowerInvariant();

            if (token.Length < 3 || StopWords.Contains(token))
            {
                continue;
            }

            if (!counts.TryGetValue(token, out var count))
            {
                order.Add(token);
            }

            counts[token] = count + 1;
        }

        return order
            .OrderByDescending(token => counts[token])
            .ThenBy(token => order.IndexOf(token))
            .Take(max)
            .ToList();
    }

    private static IEnumerable<string> Tokenize(string text)
    {
        var buffer = new StringBuilder();

        foreach (var character in text)
        {
            if (char.IsLetter(character))
            {
                buffer.Append(character);
            }
            else if (buffer.Length > 0)
            {
                yield return buffer.ToString();
                buffer.Clear();
            }
        }

        if (buffer.Length > 0)
        {
            yield return buffer.ToString();
        }
    }
}