//+------------------------------------------------------------------+
//|                                     EGARCH_InnovationZscore.mq5 |
//|                                  Copyright 2025, MetaQuotes Ltd. |
//|                                             https://www.mql5.com |
//+------------------------------------------------------------------+
#property copyright "Copyright 2025, MetaQuotes Ltd."
#property link      "https://www.mql5.com"
#property version   "1.00"

// Preprocessor directives and indicator properties
#define __SLSQP__                                 // Enable Sequential Least Squares Programming optimization algorithm
#property indicator_separate_window               // Render indicator in an independent subwindow

#include "Arch\Univariate\mean.mqh"         // Include statistical mean model header dependencies

// Define fixed indicator visual threshold levels
#property indicator_level1  2.0                   // Upper extreme threshold line
#property indicator_level2  0.0                   // Zero baseline
#property indicator_level3 -2.0                   // Lower extreme threshold line

// Define buffer and plotting counts
#property indicator_buffers 1
#property indicator_plots   1

// --- Plot 1: Standardized Innovation Z-Score
#property indicator_label1  "Eiz"
#property indicator_type1   DRAW_LINE
#property indicator_color1  clrRed
#property indicator_style1  STYLE_SOLID
#property indicator_width1  1

//+------------------------------------------------------------------+
//| INPUT PARAMETERS                                                 |
//+------------------------------------------------------------------+
input int                   BarsToDraw        = 500;           // Number of historical bars to calculate and render
input ulong                 HistoryLen        = 200;           // Rolling sample size (lookback period) for EGARCH fitting
input ulong                 WindowLen         = 20;            // Rolling window length for Z-score standardizing residuals
input double                ScaleFactor       = 100.;          // Multiplier applied to log returns (improves numerical optimizer stability)
input ENUM_MEAN_MODEL       MeanModel         = MEAN_CONSTANT; // Model type for time-series conditional mean
input bool                  MeanConstant      = true;          // Include intercept constant in mean specification
input string                MeanLags          = "";            // Comma-separated list of AR lag terms (e.g. "1,2")
ENUM_VOLATILITY_MODEL       VolatilityModel   = VOL_EGARCH;    // Volatility process specified as Exponential GARCH
input ulong                 _P_               = 1;             // GARCH order (lagged conditional log-variances)
input ulong                 _O_               = 1;             // Asymmetry order (leverage/asymmetric magnitude terms)
input ulong                 _Q_               = 1;             // ARCH order (lagged innovations)
input int                   Volatility_Seed   = 0;             // Seed for volatility model random generator initialization
input ENUM_DISTRIBUTION_MODEL ErrorDistribution = DIST_NORMAL; // Innovation error distribution assumption
input int                   Distribution_Seed = 0;             // Seed for distribution model random generator initialization

//+------------------------------------------------------------------+
//| GLOBAL INDICATOR BUFFERS AND STATE VARIABLES                     |
//+------------------------------------------------------------------+
// Dynamic array mapped to indicator plot buffer
double         EizBuffer[];

// Computational vectors and model pointers
vector         returns = vector::Zeros(HistoryLen); // Rolling logarithmic returns vector
vector         stdresid = returns;                  // Model-standardized residuals vector
vector         window;                              // Sliced sub-vector of standardized residuals for local normalization
ArchParameters model_spec;                          // Struct holding model specifications and hyperparameters
HARX*          full_model;                          // Polymorphic pointer to mean model instance

//+------------------------------------------------------------------+
//| Custom indicator initialization function                         |
//+------------------------------------------------------------------+
int OnInit()
  {
// Validate lookback window length
   if(HistoryLen < 30)
     {
      Print("Invalid input value for HistoryLen");
      return INIT_FAILED;
     }

// --- Map dynamic array to primary indicator buffer
   SetIndexBuffer(0, EizBuffer, INDICATOR_DATA);

// --- Set plot properties and drawing offset
   PlotIndexSetInteger(0, PLOT_DRAW_BEGIN, BarsToDraw);
   PlotIndexSetDouble(0, PLOT_EMPTY_VALUE, EMPTY_VALUE);

// --- Configure model specifications
   model_spec.mean_model_type = MeanModel;

// Parse comma-separated mean lag configuration string if provided
   if(StringLen(MeanLags))
     {
      string lag_info[];
      int nlags = StringSplit(MeanLags, StringGetCharacter(",", 0), lag_info);
      if(nlags > 0)
        {
         for(uint i = 0; i < uint(nlags); ++i)
           {
            if(StringLen(lag_info[i]) > 0)
              {
               if(model_spec.mean_lags.Resize(model_spec.mean_lags.Size() + 1, 3))
                  model_spec.mean_lags[model_spec.mean_lags.Size() - 1] = StringToDouble(lag_info[i]);
               else
                 {
                  Print(" error ", GetLastError());
                  return INIT_FAILED;
                 }
              }
           }
        }
     }

// Set EGARCH structural dynamics and seeds
   model_spec.vol_rng_seed   = Volatility_Seed;
   model_spec.garch_o        = _O_;
   model_spec.garch_p        = _P_;
   model_spec.garch_q        = _Q_;
   model_spec.dist_type      = ErrorDistribution;
   model_spec.dist_rng_seed  = Distribution_Seed;

// Factory instantiation of the requested mean model class
   switch(MeanModel)
     {
      case MEAN_CONSTANT:
         full_model = new ConstantMean();
         break;
      case MEAN_ZERO:
         full_model = new ZeroMean();
         break;
      case MEAN_AR:
         full_model = new AR();
         break;
      default:
         full_model = new ConstantMean();
         break;
     }

// Ensure memory allocation for mean model succeeded
   if(CheckPointer(full_model) == POINTER_INVALID)
      return INIT_FAILED;

   model_spec.vol_model_type = VolatilityModel;

   return(INIT_SUCCEEDED);
  }

//+------------------------------------------------------------------+
//| Custom indicator deinitialization function                       |
//+------------------------------------------------------------------+
void OnDeinit(const int reason)
  {
// Free dynamic memory allocated for the mean model instance
   if(CheckPointer(full_model) == POINTER_DYNAMIC)
      delete full_model;
  }

//+------------------------------------------------------------------+
//| Custom indicator iteration function                              |
//+------------------------------------------------------------------+
int OnCalculate(const int32_t rates_total,
                const int32_t prev_calculated,
                const datetime &time[],
                const double &open[],
                const double &high[],
                const double &low[],
                const double &close[],
                const long &tick_volume[],
                const long &volume[],
                const int32_t &spread[])
  {
   int32_t limit = 0;

// Verify total bars available meet the required lookback + rendering window
   if(rates_total < int32_t(HistoryLen + BarsToDraw))
     {
      Print("Not enough bars for indicator calculation");
      return -1;
     }

// Determine starting index for incremental bar calculation
   if(prev_calculated <= 0)
     {
      limit = rates_total - int32_t(fabs(BarsToDraw)); // First run: calculate specified historical depth
      ArrayInitialize(EizBuffer, EMPTY_VALUE);
     }
   else
      limit = prev_calculated - 1;                     // Subsequent runs: update only latest bar(s)

// Main calculation loop iterating through historical price bars
   for(int32_t shift = limit; shift < rates_total; ++shift)
     {
      int32_t from = (shift - int32_t(HistoryLen)) + 1;

      // Calculate logarithmic returns over the rolling lookback window
      for(int32_t i = from, k = 0; k < int32_t(HistoryLen); ++i, ++k)
         returns[k] = log(close[i] / close[i - 1]);

      // Scale returns to assist optimizer convergence
      returns *= fabs(ScaleFactor);

      // Load current sample window into specification structure
      model_spec.observations = returns;

      // Re-initialize model state with new sample window
      if(!full_model.initialize(model_spec))
        {
         Print(" initialization error ");
         return 0;
        }

      // Fit EGARCH model via maximum likelihood optimization
      ArchModelResult result = full_model.fit();
      if(!result.conditional_volatility.Size())
        {
         Print(" model fit error ");
         return 0;
        }

      // Extract full vector of standardized residuals (innovations divided by fitted volatility)
      stdresid = result.std_resid();

      // Slice out the dynamic evaluation window (last WindowLen entries) from standardized residuals
      window = np::sliceVector(stdresid, long(stdresid.Size() - WindowLen));

      // Calculate local Z-score of the most recent innovation over the sliced window (with epsilon to prevent division by zero)
      EizBuffer[shift] = (window[window.Size() - 1] - window.Mean()) / (window.Std() + 1.e-8);
     }

// Return calculated count to optimize subsequent iteration calls
   return(rates_total);
  }
//+------------------------------------------------------------------+
//+------------------------------------------------------------------+