WinWARP/WinWARP.Core/Helper/FlagEmojiParser.cs

using System;
using System.Collections.Generic;
using System.Text;

namespace WinWARP.Core.Helper;

public static class FlagEmojiParser
{
    private const int RegionalIndicatorStart = 0x1F1E6; // '🇦'
    private const int RegionalIndicatorEnd = 0x1F1FF;   // '🇿'

    /// <summary>
    /// Parses a string and returns a list of ISO 3166-1 alpha-2 country codes found as emojis.
    /// </summary>
    public static List<string> ExtractCountryCodes(string text, out List<string> countryIndicators)
    {
        countryIndicators = [];
        var countryCodes = new List<string>();
        if (string.IsNullOrEmpty(text)) return countryCodes;

        var i = 0;
        int? firstLetterCode = null;

        while (i < text.Length)
        {
            // Read the 32-bit Unicode code point (handles emoji surrogate pairs)
            var codePoint = char.ConvertToUtf32(text, i);

            // Advance the index (surrogate pairs take 2 char slots in .NET strings)
            i += char.IsSurrogatePair(text, i) ? 2 : 1;

            // Check if the character is a Regional Indicator Symbol (A-Z flag component)
            if (codePoint >= RegionalIndicatorStart && codePoint <= RegionalIndicatorEnd)
            {
                var letterOffset = codePoint - RegionalIndicatorStart; // 0 for A, 1 for B...

                if (firstLetterCode == null)
                {
                    // This is the first letter of a flag pair
                    firstLetterCode = letterOffset;
                }
                else
                {
                    // This is the second letter of a flag pair. Construct the ISO string.
                    var c1 = (char)('A' + firstLetterCode.Value);
                    var c2 = (char)('A' + letterOffset);

                    countryCodes.Add($"{c1}{c2}");
                    countryIndicators.Add($"{char.ConvertFromUtf32(RegionalIndicatorStart + firstLetterCode.Value)}{char.ConvertFromUtf32(RegionalIndicatorStart + letterOffset)}");

                    // Reset to look for the next flag
                    firstLetterCode = null;
                }
            }
            else
            {
                // If we hit any other character, break any broken/incomplete flag sequences
                firstLetterCode = null;
            }
        }

        return countryCodes;
    }

    public static List<string> ExtractCountryCodes(string text)
    {
        return ExtractCountryCodes(text, out _);
    }
}