using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;

namespace PBS.WinFrmUI.Hydro 
{
    public class DynamicThresholdProcessor
    {  
        public static List<Yw.Geometry.Point2d> FilterWithDynamicSigma(List<Yw.Geometry.Point2d> rawData)
        {
            return ProcessGroup(rawData);
        }

        private static List<Yw.Geometry.Point2d> ProcessGroup(List<Yw.Geometry.Point2d> group)
         {
            const int minSize = 3;
            if (group.Count < minSize) 
                return group;

            var pressures = group.Select(p => p.Y).ToList();

            // 计算统计量
            var (mean, stdDev) = CalculateStats(pressures);
            double skewness = CalculateSkewness(pressures);
            if (group.Count > 5)
            {

            }
            // 动态调整σ倍数
            double sigmaMultiplier = CalculateSigmaMultiplier(skewness);
            if (sigmaMultiplier==3)
            {

            }
          

            // 计算边界
            double lower = mean - sigmaMultiplier * stdDev;
            double upper = mean + sigmaMultiplier * stdDev;

            return group.Where(p => p.Y >= lower && p.Y <= upper).ToList();
        }

        // 核心统计计算
        private static (double mean, double stdDev) CalculateStats(List<double> values)
        {
            double mean = values.Average();
            double stdDev = Math.Sqrt(values.Sum(v => Math.Pow(v - mean, 2)) / (values.Count - 1));
            return (mean, stdDev);
        }

        // 偏度计算（Pearson's moment coefficient）
        private static double CalculateSkewness(List<double> values)
        {
            double mean = values.Average();
            double std = CalculateStats(values).stdDev;
            double sum = values.Sum(v => Math.Pow((v - mean) / std, 3));
            return (sum * values.Count) / ((values.Count - 1) * (values.Count - 2));
        }

        // 动态σ倍数计算规则
        private static double CalculateSigmaMultiplier(double skewness)
        {
        
            return skewness switch
            {
                > 1.0 => 2.0,    // 强正偏态
                > 0.5 => 2.5,
                > -0.5 => 3.0,   // 近似正态
                > -1.0 => 3.5,
                _ => 4.0    // 强负偏态
            };
        }

    }
}