using CommonLib.LOG;
using NWaves.FeatureExtractors;
using NWaves.FeatureExtractors.Multi;
using NWaves.FeatureExtractors.Options;
using System;
using System.Data.SqlTypes;
using System.IO;
using System.Linq;
using System.Text;
using System.Windows.Forms;

namespace DataCollectionSystem.CommonLib.Data
{
    public class DataProcess
    {
        alglib.clusterizerstate s1;//释放聚类空间
        alglib.kmeansreport rep1;//聚类结果报告
        alglib.clusterizerstate s2;
        alglib.kmeansreport rep2;
        public int repeat = 0;//次数

        //定义固定窗
        int MatrixLen = 20; //数据窗的长度(行)
        int MaxtrixDimension =4;//数据的特征维度（列）
        double[,] fixed_window = null;//new double[MatrixLen, MaxtrixDimension];

        //定义特征展示窗
        double[] features = null;//new double[35];

        //int Len = 3600; //回溯1小时Mel
        int Len = 300; //回溯1小时Mel

        public double[] Mel1 = null;//new double[Len];//特征窗
        public double[] Mel2 = null;//new double[Len];//特征窗
        public double[] Time1 = null;//new double[Len];//特征窗
        public double[] Time2 = null;//new double[Len];//特征窗
        public double[] spectral1 = null;//new double[Len];//特征窗
        public double[] spectral2 = null;//new double[Len];//特征窗
        public double[,] all_features = new double[35, 300];//new double[Len];//特征窗

        public int pos = 0;
        //定义一个数据窗

        int DataLen = 20; //数据窗的长度(行)
        int DataDimension = 4;//数据的特征维度（列）
        //double[,] Data_window = null;// new double[DataLen, DataDimension];
        double[,] Data_window = null;//new double[DataLen, DataDimension];


        //固定莱特窗
        int WrightWindowLen =20;
        double[] Wright = null;//new double[WrightWindowLen];//莱特窗

        double k = 8;//倍数

        //定义窗的指标
        double avr = 0;
        //int sum = 0;
        //double sum1 = 0;
        double Standard = 0;
        //double ss = 0;
        double Residual = 0;





        /// <summary>
        /// 把环形缓冲区的数据计算出来
        /// </summary>
        /// <param name="Maxtrix"></param>
        /// <returns></returns>
        public   Boolean checkError(string identName,double[] RingArray1,
            ref Boolean anomaly,
            ref int anomalyInt
            )//double Feature_0, double Feature_1)
        {
            int len = RingArray1.Length;
            float[] RingArray = new float[len];
            for (int i = 0; i < len; i++)
            {
                RingArray[i] = (float)RingArray1[i];
            }
            Boolean result = true;
            alglib.clusterizerstate s1;//释放聚类空间
            alglib.kmeansreport rep1;//聚类结果报告
            alglib.clusterizerstate s2;
            alglib.kmeansreport rep2;
            if (Data_window == null)
            {
                
                
                //定义固定窗
              
                fixed_window = new double[MatrixLen, MaxtrixDimension];

                //定义特征展示窗
                features = new double[35];

               
                Mel1 = new double[Len];//特征窗
                Mel2 = new double[Len];//特征窗
                Time1 = new double[Len];//特征窗
                Time2 = new double[Len];//特征窗
                spectral1 = new double[Len];//特征窗
                spectral2 = new double[Len];//特征窗

                //定义一个数据窗

               
                Data_window = new double[DataLen, DataDimension];


                //固定莱特窗

                Wright = new double[WrightWindowLen];//莱特窗

            }

            #region 特征提取

            var mfccOptions = new MfccOptions //梅尔倒谱系数提取
            {
                SamplingRate = 32000,
                FeatureCount = 13,
                FrameDuration = 1,
                HopDuration = 1,
                FilterBankSize = 26,
                // PreEmphasis = 0.97,
            };
            var mfccExtractor = new MfccExtractor(mfccOptions);
            //var mfccVectors = mfccExtractor.ComputeFrom(fArr); //得到mel系数
            var mfccVectors = (mfccExtractor.ComputeFrom(RingArray))[0]; //得到mel系数
                                                                       //float[] mfccvectors = mfccVectors[0];
                                                                       //保存mel特征
            for (int i = 0; i < 13; ++i)
            {
                features[i] = mfccVectors[i];
                //fArr[i] = f;
            }

            var lpccOptions = new LpccOptions //LPCC系数提取
            {
                SamplingRate = 32000,
                FeatureCount = 10,
                FrameDuration = 1,
                HopDuration = 1,
                LpcOrder = 1,
                //LifterSize = _lifterSize,
                // PreEmphasis = _preEmphasis,
                //Window = _window
            };
            var LpccExtractor = new LpccExtractor(lpccOptions);
            var LpccVectors = (LpccExtractor.ComputeFrom(RingArray))[0]; //得到前10维LPCC系数 
                                                                       //float[] Lpccvectors = LpccVectors[0];

            

            for (int i = 0; i < 10; ++i)
            {
                features[i + 13] = LpccVectors[i];
                //fArr[i] = f;
            }

            var spectralOptions = new MultiFeatureOptions //频域特征提取
            {
                SamplingRate = 32000,
                //FeatureList = spectralFeatureSet,
                FrameDuration = 1,
                HopDuration = 1,
                //FftSize = _blockSize,
                //Frequencies = _frequencies,
                //PreEmphasis = _preEmphasis,
                //Window = _window,
                //Parameters = _parameters
            };


            var Spectral_featureExtractor = new SpectralFeaturesExtractor(spectralOptions);
            var Spectral_featureVectors = (Spectral_featureExtractor.ComputeFrom(RingArray))[0]; //频域特征向量 

            for (int i = 0; i < 8; i++)
            {
                features[i + 23] = Spectral_featureVectors[i];
                //fArr[i] = f;
            }

            var timeOptions = new MultiFeatureOptions //时域特征提取
            {
                SamplingRate = 31998,
                FrameDuration = 1,
                HopDuration = 1,
                //FeatureList = string.Join(",", FeatureDescriptions),
                //Parameters = _parameters
            };

            var Time_featureExtractor = new TimeDomainFeaturesExtractor(timeOptions);

            //float[] Time_HardArr = new float[10001];
            //for (int i = 0; i < 10000; i++)
            //{
                //Time_HardArr[i] = HardArr[i];
                //fArr[i] = f;
            //}
            //Time_HardArr[10000] = HardArr[9999];

            var timeVectors = (Time_featureExtractor.ComputeFrom(RingArray))[0]; //时域特征向量 
                                                                                 //var timeVectors = Time_featureExtractor.ComputeFrom(HardArr); //时域特征向量

            for (int i = 0; i < 4; ++i)
            {
                features[i + 31] = timeVectors[i];
                //fArr[i] = f;
            }


            //all
            //int Len = 300; //回溯1小时Mel
            //features = new double[35];
            // public double[,] all_features = new double[35, 300];//new double[Len];//特征窗





            #endregion

            #region 特征窗更新

            double m1 = features[0];
            for (int i = 1; i < Len; i++)//窗更新
            {
                Mel1[i - 1] = Mel1[i];
            }

            Mel1[Len - 1] = m1;

            double m2 = features[15];
            for (int i = 1; i < Len; i++)//窗更新
            {
                Mel2[i - 1] = Mel2[i];
            }

            Mel2[Len - 1] = m2;

            double t1 = features[31];
            for (int i = 1; i < Len; i++)//窗更新
            {
                Time1[i - 1] = Time1[i];
            }

            Time1[Len - 1] = t1;

            double t2 = features[21];
            for (int i = 1; i < Len; i++)//窗更新
            {
                Time2[i - 1] = Time2[i];
            }

            Time2[Len - 1] = t2;

            double sp1 = features[23];
            for (int i = 1; i < Len; i++)//窗更新
            {
                spectral1[i - 1] = spectral1[i];
            }

            spectral1[Len - 1] = sp1;

            double sp2 = features[25];
            for (int i = 1; i < Len; i++)//窗更新
            {
                spectral2[i - 1] = spectral2[i];
            }

            spectral2[Len - 1] = sp2;


            #endregion

            // 选取特征 阈值a b c d e
            double a = features[31];
            double b = features[32];
            double c = features[33];
            //double d = features[23];//对质心不用归一
            double d = features[23] / 32000;//对质心归一
            double e = features[29];


            try
            {



                //方法1 阈值

                //if(repeat > 1 )
                //{
                //string Info1 = $"查看特征：由于短时能量:{a},RMS:{b},过零率:{c},质心:{d},谱熵:{e}";
                //LogStorage.Logs.Add(LogLevel.Error, Info1);
                //}

                // 


                //if (a > 0.015 || b > 0.12 || c > 0.14 || d > 0.3)
                if (a > 1.5 && b > 1.2 && c > 1.4 && d > 3)
                //if (a > 0.015 && b > 0.12 && c > 0.14 && d > 1600 )//残差减去5倍标准差
                //if (a > 0.15 || b > 0.12 || c > 0.14 || d > 1600 )//残差减去5倍标准差
                //if(j>3&&laiTe[j-1]> laiTe[j]&& laiTe[j - 1]> laiTe[j - 2]&& ((laiTe_2[j-1]/ laiTe[j - 1])>0.1))
                {
                    /* if(a > 0.015)
                     {
                         string Info1 = $"法1（阈值判别）：由于C大于设定的阈值(0.015)，在第{repeat}点出现了疑似异常";
                         LogStorage.Logs.Add(LogLevel.Error, Info1);
                     }
                     if (b > 0.12)
                     {
                         string Info1 = $"法1（阈值判别）：由于RMS:{b}大于设定的阈值(0.12)，在第{repeat}点出现了疑似异常";
                         LogStorage.Logs.Add(LogLevel.Error, Info1);
                     }
                     if (c > 0.14)
                     {
                         string Info1 = $"法1（阈值判别）：由于过零率:{c}大于设定的阈值(0.14)，在第{repeat}点出现了疑似异常";
                         LogStorage.Logs.Add(LogLevel.Error, Info1);
                     }
                     if (d > 0.3)
                     {
                         string Info1 = $"法1（阈值判别）：由于谱熵:{e}大于设定的阈值(0.3)，在第{repeat}点出现了疑似异常";
                         LogStorage.Logs.Add(LogLevel.Error, Info1);
                     }*/
                    /*
                    Console.ForegroundColor = ConsoleColor.Red;
                    Console.WriteLine("******************************************************异常出现******************************************************");
                    Console.WriteLine("法1（阈值判别）：由于短时能量:{0},RMS:{1},过零率:{2},质心:{3},谱熵:{4} 之中有大于设定的阈值，在第{5}点出现了疑似异常", a, b, c, d, e, j);
                    Console.WriteLine("******************************************************异常出现******************************************************");
                    */
                    string Info1 = $"{identName} 法1（阈值判别）：由于短时能量:{a}大于设定的阈值(0.015)，在第{repeat}点出现了疑似异常";
                    LogStorage.Logs.Add(LogLevel.Error1, Info1);
                    anomaly = true;
                    anomalyInt = 1;

                    //string Info1 = $"法1（阈值判别）：由于短时能量:{a},RMS:{b},过零率:{c},质心:{d},谱熵:{e} 之中有大于设定的阈值，在第{repeat}点出现了疑似异常";
                    //LogStorage.Logs.Add(LogLevel.Error, Info1);
                }
                //Console.WriteLine("短时能量:{0},RMS:{1},过零率:{2},质心:{3},谱熵:{4} ，目前运行到了第{5}点  运行正常！！！", a, b, c, d, e,j);
                //Console.WriteLine("目前运行到了第{0}点", j);
                //Console.ForegroundColor = ConsoleColor.White;
                //Console.ReadKey();


                //方法2 均值聚类中心差异判别

                //固定窗 fixed_window
                //int MatrixLen = 20; //数据窗的长度(行)
                //int MaxtrixDimension = 4;//数据的特征维度（列）
                //double[,] fixed_window = new double[MatrixLen, MaxtrixDimension];
                for (int z = 1; z < MatrixLen; z++)//窗更新
                {
                    fixed_window[z - 1, 0] = fixed_window[z, 0];
                    fixed_window[z - 1, 1] = fixed_window[z, 1];
                    fixed_window[z - 1, 2] = fixed_window[z, 2];
                    fixed_window[z - 1, 3] = fixed_window[z, 3];
                }
                //Wright[Wright.Length - 1] = E;
                fixed_window[MatrixLen - 1, 0] = a;
                fixed_window[MatrixLen - 1, 1] = b;
                fixed_window[MatrixLen - 1, 2] = c;
                fixed_window[MatrixLen - 1, 3] = d;

                if (repeat >= 20 && repeat % 20 == 0) //满足取窗条件
                {
                    //开始聚类
                    alglib.clusterizercreate(out s2);
                    alglib.clusterizersetpoints(s2, fixed_window, 2);
                    alglib.clusterizersetkmeanslimits(s2, 2000, 0);
                    alglib.clusterizerrunkmeans(s2, 2, out rep2);//聚2类

                    double a1 = rep2.c[0, 0];
                    double a2 = rep2.c[0, 1];
                    double a3 = rep2.c[0, 2];
                    double a4 = rep2.c[0, 3];

                    double b1 = rep2.c[1, 0];
                    double b2 = rep2.c[1, 1];
                    double b3 = rep2.c[1, 2];
                    double b4 = rep2.c[1, 3];

                    double U1 = Math.Abs(a1 - b1); // + Math.Abs(a2 - b2) + Math.Abs(a3 - b3);
                    double U2 = Math.Abs(a2 - b2);
                    double U3 = Math.Abs(a3 - b3);
                    double U4 = Math.Abs(a4 - b4);

                    int sum = 0;
                    for (int i = 0; i < rep2.cidx.Length; i++)
                    {
                        //Console.WriteLine(Wright[i]);
                        sum += rep2.cidx[i];
                    }

                    // int sum = rep2.cidx.Sum();

                    if (U1 > 1 && U2 > 5 && U3 > 7 && U4 > 1.2 && sum > 9)
                    //if ((U1 > 0.01 || U2 > 0.05 || U3 > 0.07 || U4 > 0.12) && sum > 9)
                    //if (U1 > 0.01 && U2 > 0.05 && U3 > 0.07 && U4 > 800 && sum > 9)//残差减去5倍标准差
                    //if (U1 > 0.01 || U2 > 0.05 || U3 > 0.07 || U4 > 800)//残差减去5倍标准差
                    //if (a > 0.15 || b > 0.12 || c > 0.14 || d > 1600 )//残差减去5倍标准差
                    //if(j>3&&laiTe[j-1]> laiTe[j]&& laiTe[j - 1]> laiTe[j - 2]&& ((laiTe_2[j-1]/ laiTe[j - 1])>0.1))
                    {
                        /*
                        Console.ForegroundColor = ConsoleColor.Blue;
                        Console.WriteLine("******************************************************异常出现******************************************************");
                        Console.WriteLine("法2（均值聚类）：由于聚类中心c1({0},{1},{2},{3})与聚类中心c2({4},{5},{6},{7})相差过大，在第{8}点出现了疑似异常", a1, a2, a3, a4, b1, b2, b3, b4, j);
                        Console.WriteLine("第{0}聚类，异常的个数为{1}", j, sum);
                        Console.WriteLine("******************************************************异常出现******************************************************");
                        */
                        string Info1 = $"{identName}  法2（均值聚类）：由于聚类中心c1({a1},{a2},{a3},{a4})与聚类中心c2({b1},{b2},{b3},{b4})相差过大，在第{repeat}点出现了疑似异常,异常个数{sum}";
                        LogStorage.Logs.Add(LogLevel.Error2, Info1);
                        anomaly = true;
                        anomalyInt = 1;
                    }
                    //Console.WriteLine("短时能量:{0},RMS:{1},过零率:{2},质心:{3},谱熵:{4} ，目前运行到了第{5}点  运行正常！！！", a, b, c, d, e, j);
                    //Console.ForegroundColor = ConsoleColor.White;
                    //Console.ReadKey();

                }


                // 滑动窗口判别

                //对数据窗更新
                //int DataLen = 20; //数据窗的长度(行)
                //int DataDimension = 4;//数据的特征维度（列）
                //double[,] Data_window = new double[DataLen, DataDimension];

                for (int z = 1; z < DataLen; z++)//窗更新
                {
                    Data_window[z - 1, 0] = Data_window[z, 0];
                    Data_window[z - 1, 1] = Data_window[z, 1];
                    Data_window[z - 1, 2] = Data_window[z, 2];
                    Data_window[z - 1, 3] = Data_window[z, 3];
                }
                //Wright[Wright.Length - 1] = E;
                Data_window[DataLen - 1, 0] = a;
                Data_window[DataLen - 1, 1] = b;
                Data_window[DataLen - 1, 2] = c;
                Data_window[DataLen - 1, 3] = d;

                alglib.clusterizercreate(out s1);
                alglib.clusterizersetpoints(s1, Data_window, 2);
                alglib.clusterizersetkmeanslimits(s1, 10, 0);
                alglib.clusterizerrunkmeans(s1, 1, out rep1);//聚1类



                alglib.clusterizercreate(out s2);
                alglib.clusterizersetpoints(s2, Data_window, 2);
                alglib.clusterizersetkmeanslimits(s2, 2000, 0);
                alglib.clusterizerrunkmeans(s2, 2, out rep2);//聚2类

                double E = Math.Abs(rep1.energy - rep2.energy);
                //double E = Math.Abs(rep1.energy - rep2.energy);//只检测突然增大

                for (int i = 1; i < Wright.Length; i++)//窗更新
                {
                    Wright[i - 1] = Wright[i];
                }

                Wright[Wright.Length - 1] = E;

                double sum1 = 0;

                for (int i = 0; i < Wright.Length; i++)
                {
                    //Console.WriteLine(Wright[i]);
                    sum1 += Wright[i];
                }

                //sum1 = Wright.Sum();
                avr = sum1 / Wright.Length;//均值

                double ss = 0;

                for (int i = 0; i < Wright.Length; i++)
                {
                    ss += Math.Pow((Wright[i] - avr), 2);
                }
                //biaoZhun = Math.Sqrt(s/(laiTe.Length-1));//除以N-1
                Standard = Math.Sqrt(ss / (Wright.Length));//除以N

                Residual = Math.Abs(Wright[Wright.Length - 1] - avr);

                if (Residual > k * Standard && a > 1 && repeat > 40)
                {/*
                    Console.ForegroundColor = ConsoleColor.Yellow;
                    Console.WriteLine("******************************************************异常出现******************************************************");
                    Console.WriteLine("法3（自适应聚类）：由于残差{0}大于标准差{1},在第{2}点出现了疑似异常", Residual, Standard, j);
                    //Console.WriteLine("第{0}聚类，异常的个数为{1}", j, sum);
                    Console.WriteLine("******************************************************异常出现******************************************************");
                    */
                    string Info1 = $"{identName}  法3（自适应聚类）：由于残差{Residual}大于{k}倍标准差{Standard},在第{repeat}点出现了疑似异常";
                    LogStorage.Logs.Add(LogLevel.Error, Info1);
                    anomaly = true;
                    anomalyInt = 1;
                    /*
                    Info1 = $"第{j}聚类，异常的个数为{sum}";
                    LogStorage.Logs.Add(LogLevel.Error, Info1);
                    */
                }
                //if(repeat>20)
                //{
                    //string Info1 = $"查看计算值：和1{sum1}、和2{ss}、平均值{avr}、第{repeat}点";
                    //LogStorage.Logs.Add(LogLevel.Error, Info1);
                //}
                repeat++;

                if (repeat > 8640000)
                {
                    repeat = 0;
                }
                //保存特征值
                if(pos<300)
                {                // all_features features
                    int vector_size = features.Count();
                    for (int i = 0; i < vector_size; i++)
                    {
                        all_features[i, pos] = features[i];
                    }
                    pos++;
                }
                else
                {
                    if (pos == 300) 
                    { 
                        SaveCSV(all_features);
                        pos++;
                    }
                }

            }
            catch (Exception ex)
            {
                LogStorage.Logs.Add(LogLevel.Error, ex.Message + ex.StackTrace);
            }



            // }
            return result;
        }

        private void SaveCSV(double[,] all_features)
        {
            int len0 = all_features.GetLength(0);
            int len1 =  all_features.GetLength(1);
            StringBuilder sb = new StringBuilder();
            for(int i=0; i<len0; i++)
            {
                for(int j=0;j<len1;j++)
                {
                    sb.Append(all_features[i, j].ToString());
                    sb.Append(",");
                }
                sb.Append("\r\n");
            }
            File.WriteAllText("all_features.csv", sb.ToString());
        }
    }
}