MarketAlly.AIPlugin.Extensions/MarketAlly.AIPlugin.Context/Search/FuzzyMatcher.cs

using System.Text;

namespace MarketAlly.AIPlugin.Context.Search
{
	/// <summary>
	/// Provides fuzzy string matching capabilities for search operations
	/// </summary>
	public class FuzzyMatcher
	{
		private readonly double _threshold;

		public FuzzyMatcher(double threshold = 0.7)
		{
			_threshold = threshold;
		}

		/// <summary>
		/// Calculates similarity between two strings using multiple algorithms
		/// </summary>
		public double CalculateSimilarity(string source, string target)
		{
			if (string.IsNullOrEmpty(source) || string.IsNullOrEmpty(target))
				return 0.0;

			// Normalize strings
			var normalizedSource = NormalizeString(source);
			var normalizedTarget = NormalizeString(target);

			// If strings are identical after normalization, return perfect match
			if (normalizedSource.Equals(normalizedTarget, StringComparison.OrdinalIgnoreCase))
				return 1.0;

			// Combine multiple similarity algorithms for better accuracy
			var levenshteinRatio = CalculateLevenshteinRatio(normalizedSource, normalizedTarget);
			var jaroWinklerScore = CalculateJaroWinkler(normalizedSource, normalizedTarget);
			var tokenSetRatio = CalculateTokenSetRatio(normalizedSource, normalizedTarget);

			// Weighted combination of different algorithms
			var combinedScore = (levenshteinRatio * 0.4) + (jaroWinklerScore * 0.4) + (tokenSetRatio * 0.2);

			return Math.Min(1.0, combinedScore);
		}

		/// <summary>
		/// Checks if two strings are similar based on the configured threshold
		/// </summary>
		public bool AreSimilar(string source, string target)
		{
			return CalculateSimilarity(source, target) >= _threshold;
		}

		/// <summary>
		/// Finds the best match for a query in a collection of candidates
		/// </summary>
		public FuzzyMatchResult FindBestMatch(string query, IEnumerable<string> candidates)
		{
			var bestMatch = new FuzzyMatchResult { Query = query };

			foreach (var candidate in candidates)
			{
				var similarity = CalculateSimilarity(query, candidate);
				if (similarity > bestMatch.Score)
				{
					bestMatch.Match = candidate;
					bestMatch.Score = similarity;
				}
			}

			bestMatch.IsMatch = bestMatch.Score >= _threshold;
			return bestMatch;
		}

		/// <summary>
		/// Finds all matches above the threshold
		/// </summary>
		public List<FuzzyMatchResult> FindMatches(string query, IEnumerable<string> candidates)
		{
			var matches = new List<FuzzyMatchResult>();

			foreach (var candidate in candidates)
			{
				var similarity = CalculateSimilarity(query, candidate);
				if (similarity >= _threshold)
				{
					matches.Add(new FuzzyMatchResult
					{
						Query = query,
						Match = candidate,
						Score = similarity,
						IsMatch = true
					});
				}
			}

			return matches.OrderByDescending(m => m.Score).ToList();
		}

		/// <summary>
		/// Normalizes a string for better matching
		/// </summary>
		private string NormalizeString(string input)
		{
			if (string.IsNullOrEmpty(input))
				return string.Empty;

			return input.ToLowerInvariant()
						.Trim()
						.Replace("  ", " ") // Remove double spaces
						.Replace("-", " ")
						.Replace("_", " ");
		}

		/// <summary>
		/// Calculates Levenshtein distance ratio (1 - distance/max_length)
		/// </summary>
		private double CalculateLevenshteinRatio(string source, string target)
		{
			var distance = CalculateLevenshteinDistance(source, target);
			var maxLength = Math.Max(source.Length, target.Length);
			
			if (maxLength == 0)
				return 1.0;

			return 1.0 - (double)distance / maxLength;
		}

		/// <summary>
		/// Calculates Levenshtein distance between two strings
		/// </summary>
		private int CalculateLevenshteinDistance(string source, string target)
		{
			if (string.IsNullOrEmpty(source))
				return target?.Length ?? 0;

			if (string.IsNullOrEmpty(target))
				return source.Length;

			var sourceLength = source.Length;
			var targetLength = target.Length;
			var matrix = new int[sourceLength + 1, targetLength + 1];

			// Initialize first row and column
			for (int i = 0; i <= sourceLength; i++)
				matrix[i, 0] = i;

			for (int j = 0; j <= targetLength; j++)
				matrix[0, j] = j;

			// Calculate distances
			for (int i = 1; i <= sourceLength; i++)
			{
				for (int j = 1; j <= targetLength; j++)
				{
					var cost = source[i - 1] == target[j - 1] ? 0 : 1;
					matrix[i, j] = Math.Min(
						Math.Min(matrix[i - 1, j] + 1, matrix[i, j - 1] + 1),
						matrix[i - 1, j - 1] + cost);
				}
			}

			return matrix[sourceLength, targetLength];
		}

		/// <summary>
		/// Calculates Jaro-Winkler similarity
		/// </summary>
		private double CalculateJaroWinkler(string source, string target)
		{
			var jaroScore = CalculateJaro(source, target);
			
			if (jaroScore < 0.7) // Jaro-Winkler only applies prefix scaling if Jaro > 0.7
				return jaroScore;

			// Calculate common prefix (up to 4 characters)
			var prefix = 0;
			var minLength = Math.Min(source.Length, Math.Min(target.Length, 4));
			
			for (int i = 0; i < minLength; i++)
			{
				if (source[i] == target[i])
					prefix++;
				else
					break;
			}

			return jaroScore + (0.1 * prefix * (1 - jaroScore));
		}

		/// <summary>
		/// Calculates Jaro similarity
		/// </summary>
		private double CalculateJaro(string source, string target)
		{
			if (source.Length == 0 && target.Length == 0)
				return 1.0;

			if (source.Length == 0 || target.Length == 0)
				return 0.0;

			var matchWindow = Math.Max(source.Length, target.Length) / 2 - 1;
			if (matchWindow < 0)
				matchWindow = 0;

			var sourceMatches = new bool[source.Length];
			var targetMatches = new bool[target.Length];
			var matches = 0;
			var transpositions = 0;

			// Find matches
			for (int i = 0; i < source.Length; i++)
			{
				var start = Math.Max(0, i - matchWindow);
				var end = Math.Min(i + matchWindow + 1, target.Length);

				for (int j = start; j < end; j++)
				{
					if (targetMatches[j] || source[i] != target[j])
						continue;

					sourceMatches[i] = true;
					targetMatches[j] = true;
					matches++;
					break;
				}
			}

			if (matches == 0)
				return 0.0;

			// Count transpositions
			var k = 0;
			for (int i = 0; i < source.Length; i++)
			{
				if (!sourceMatches[i])
					continue;

				while (!targetMatches[k])
					k++;

				if (source[i] != target[k])
					transpositions++;

				k++;
			}

			return (matches / (double)source.Length + 
					matches / (double)target.Length + 
					(matches - transpositions / 2.0) / matches) / 3.0;
		}

		/// <summary>
		/// Calculates token set ratio for better handling of word order differences
		/// </summary>
		private double CalculateTokenSetRatio(string source, string target)
		{
			var sourceTokens = new HashSet<string>(source.Split(' ', StringSplitOptions.RemoveEmptyEntries));
			var targetTokens = new HashSet<string>(target.Split(' ', StringSplitOptions.RemoveEmptyEntries));

			if (sourceTokens.Count == 0 && targetTokens.Count == 0)
				return 1.0;

			if (sourceTokens.Count == 0 || targetTokens.Count == 0)
				return 0.0;

			var intersection = sourceTokens.Intersect(targetTokens).Count();
			var union = sourceTokens.Union(targetTokens).Count();

			return (double)intersection / union;
		}
	}

	/// <summary>
	/// Result of a fuzzy matching operation
	/// </summary>
	public class FuzzyMatchResult
	{
		public string Query { get; set; } = "";
		public string Match { get; set; } = "";
		public double Score { get; set; }
		public bool IsMatch { get; set; }
	}
}