from AlgorithmImports import *
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
import numpy as np
import pandas as pd

class MLTradingAlgorithm(QCAlgorithm):
    
    def Initialize(self):
        # 1. Setup Algorithm Parameters
        self.SetStartDate(2019, 1, 1)
        self.SetEndDate(2024, 12, 31)
        self.SetCash(100000)    # Increase to 100k for more comfortable partial allocation

        # 2. Configurable partial allocation (20% of portfolio by default)
        self.allocation = 0.20  

        # 3. Add Equity
        self.symbol = self.AddEquity("MSTR", Resolution.Daily).Symbol

        # 4. Rolling Window for 200 Days
        self.data = RollingWindow[TradeBar](200)
        self.SetWarmUp(200)

        # 5. Random Forest Model
        self.model = RandomForestClassifier(n_estimators=100, random_state=42)
        self.training_count = 0
        self.is_model_trained = False

        # 6. Schedule Weekly Retraining
        self.Schedule.On(self.DateRules.Every(DayOfWeek.Monday),
                         self.TimeRules.At(10, 0),
                         self.TrainModel)
        
        # 7. Trailing Stop Tracking
        self.highestPrice = 0
        self.trailStopTicket = None
        self.trailing_stop_pct = 1  # Example: 5% trailing stop below highest price

    def OnData(self, data):
        # Check for valid data
        if not data.ContainsKey(self.symbol):
            return
        
        trade_bar = data[self.symbol]
        if trade_bar is None:
            return
        
        # Update rolling window
        self.data.Add(trade_bar)
        if not self.data.IsReady or self.data.Count < 200:
            return
        
        # Skip if model not trained
        if not self.is_model_trained:
            return
        
        # Build features for the latest bar
        df = self.GetFeatureDataFrame()
        if df is None or len(df) == 0:
            return
        
        latest_features = df.iloc[-1, :-1].values.reshape(1, -1)
        try:
            prediction = self.model.predict(latest_features)[0]  # 1 = Buy, 0 = Sell
        except:
            return
        
        holdings = self.Portfolio[self.symbol].Quantity

        # -- Trading Logic --
        # If prediction = 1 (bullish), we want to go long (to 'allocation') if not already
        if prediction == 1 and holdings <= 0:
            # Liquidate any short holdings just in case
            if holdings < 0:
                self.Liquidate(self.symbol)
                self.ResetTrailingStop()
            
            # Go long with partial allocation
            self.SetHoldings(self.symbol, self.allocation)
            
            # Reset trailing stop logic for a new position
            self.highestPrice = trade_bar.Close
            self.PlaceOrUpdateTrailingStop()

        # If prediction = 0 (bearish), we want to close any existing position
        elif prediction == 0 and holdings > 0:
            self.Liquidate(self.symbol)
            self.ResetTrailingStop()
        
        # -- Update Trailing Stop if holding a long position --
        if holdings > 0:
            # If price made a new high, adjust the trailing stop upward
            if trade_bar.Close > self.highestPrice:
                self.highestPrice = trade_bar.Close
                self.PlaceOrUpdateTrailingStop()


    def TrainModel(self):
        # Prepare training data
        df = self.GetFeatureDataFrame()
        if df is None or len(df) < 50:
            self.Debug("Insufficient data for training.")
            return
        
        X = df.iloc[:, :-1]
        y = df.iloc[:, -1]

        # 80/20 split (time-based)
        X_train, X_test, y_train, y_test = train_test_split(
            X, y, test_size=0.2, shuffle=False, random_state=42
        )

        # Fit the model
        self.model.fit(X_train, y_train)
        self.is_model_trained = True

        # Evaluate
        y_train_pred = self.model.predict(X_train)
        train_accuracy = accuracy_score(y_train, y_train_pred)
        y_test_pred = self.model.predict(X_test)
        test_accuracy = accuracy_score(y_test_pred, y_test)

        self.training_count += 1
        self.Debug(f"Training #{self.training_count}: "
                   f"Train Accuracy: {train_accuracy:.2%}, "
                   f"Test Accuracy: {test_accuracy:.2%}")

    def GetFeatureDataFrame(self):
        """ Build features: Bollinger Bands + Historical Volatility, target = next-day close up/down """
        if self.data.Count < 200:
            return None
        
        close_prices = [bar.Close for bar in self.data]
        df = pd.DataFrame(close_prices, columns=["Close"])
        
        # Bollinger Bands (20-day)
        period = 20
        df["BB_mid"] = df["Close"].rolling(period).mean()
        df["BB_std"] = df["Close"].rolling(period).std()
        df["BB_upper"] = df["BB_mid"] + 2 * df["BB_std"]
        df["BB_lower"] = df["BB_mid"] - 2 * df["BB_std"]
        
        # Historical Volatility (30-day)
        df["daily_returns"] = df["Close"].pct_change()
        df["HV_30"] = df["daily_returns"].rolling(window=30).std() * np.sqrt(252)

        # Target
        df["Target"] = (df["Close"].shift(-1) > df["Close"]).astype(int)
        
        # Drop NaN
        df.dropna(inplace=True)
        
        # Drop intermediate columns not used as features
        df.drop(columns=["daily_returns"], inplace=True)
        
        return df


    # ----------------------------------------------------------------
    # Trailing Stop Logic 
    # ----------------------------------------------------------------
    def PlaceOrUpdateTrailingStop(self):
        """
        Places a stop-market order ticket if we don't have one,
        or updates the stop price if we do.
        """
        quantity = self.Portfolio[self.symbol].Quantity
        if quantity <= 0:
            return
        
        # Trailing stop is 5% below highestPrice, for example
        newStopPrice = self.highestPrice * self.trailing_stop_pct
        
        if not self.trailStopTicket or self.trailStopTicket.Status in [OrderStatus.Filled, OrderStatus.Canceled, OrderStatus.Invalid]:
            # Create a new stop-market ticket (negative quantity = sell)
            self.trailStopTicket = self.StopMarketOrder(self.symbol, -quantity, newStopPrice)
        else:
            # Update existing stop ticket
            updateFields = UpdateOrderFields()
            updateFields.StopPrice = newStopPrice
            self.trailStopTicket.Update(updateFields)
            

    def ResetTrailingStop(self):
        """
        Resets tracking variables and cancels any active stop ticket
        when we exit or reverse position.
        """
        self.highestPrice = 0
        if self.trailStopTicket is not None:
            # Cancel the trailing stop ticket if it is open
            if self.trailStopTicket.Status not in [OrderStatus.Filled, OrderStatus.Canceled, OrderStatus.Invalid]:
                self.trailStopTicket.Cancel("Exiting position or reversing trade.")
        self.trailStopTicket = None