# source: https://raw.githubusercontent.com/adityabhimesha/DebtInducerBot/667def4a692afcad11adaf4d114d83f98bf232fd/user_data/strategies/TrainDeepQStrategy.py
# pragma pylint: disable=missing-docstring, invalid-name, pointless-string-statement
# flake8: noqa: F401

# --- Do not remove these libs ---
import numpy as np  # noqa
import pandas as pd  # noqa
from pandas import DataFrame  # noqa

from freqtrade.strategy import (BooleanParameter, CategoricalParameter, DecimalParameter,
                                IStrategy, IntParameter)

# --------------------------------
# Add your lib to import here
import talib.abstract as ta
import pandas_ta as pta
import freqtrade.vendor.qtpylib.indicators as qtpylib

from DQN import Agent
from utils import *
import logging

logger = logging.getLogger(__name__)


class github_adityabhimesha_DebtInducerBot__TrainDeepQStrategy__20220722_082505(IStrategy):
    # Strategy interface version - allow new iterations of the strategy interface.
    # Check the documentation or the Sample strategy to get the latest version.
    INTERFACE_VERSION = 3

    # Optimal timeframe for the strategy.
    timeframe = '5m'

    num_episode = 5
    window_size = 10+3 
    agent = Agent(window_size, False) #Training

    action_dict = {0: 'Hold', 1: 'Buy', 2: 'Sell'}

    # Can this strategy go short?
    can_short: bool = False

    # Minimal ROI designed for the strategy.
    # This attribute will be overridden if the config file contains "minimal_roi".
    minimal_roi = {
        "60": 0.01,
        "30": 0.02,
        "0": 0.04
    }

    # Optimal stoploss designed for the strategy.
    # This attribute will be overridden if the config file contains "stoploss".
    stoploss = -0.10

    # Trailing stoploss
    trailing_stop = False
    # trailing_only_offset_is_reached = False
    # trailing_stop_positive = 0.01
    # trailing_stop_positive_offset = 0.0  # Disabled / not configured

    # Run "populate_indicators()" only for new candle.
    process_only_new_candles = True

    # These values can be overridden in the config.
    use_exit_signal = True
    exit_profit_only = False
    ignore_roi_if_entry_signal = False

    # Number of candles the strategy requires before producing valid signals
    startup_candle_count: int = 30


    # Optional order type mapping.
    order_types = {
        'entry': 'limit',
        'exit': 'limit',
        'stoploss': 'market',
        'stoploss_on_exchange': False
    }

    # Optional order time in force.
    order_time_in_force = {
        'entry': 'gtc',
        'exit': 'gtc'
    }
    
    @property
    def plot_config(self):
        return {
            # Main plot indicators (Moving averages, ...)
            'main_plot': {
            },
            'subplots': {
            }
        }

    def informative_pairs(self):
        return []

    def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
        
        # Momentum Indicators
        # ------------------------------------

        # ADX
        dataframe['adx'] = ta.ADX(dataframe)

        # # Plus Directional Indicator / Movement
        # dataframe['plus_dm'] = ta.PLUS_DM(dataframe)
        # dataframe['plus_di'] = ta.PLUS_DI(dataframe)

        # # Minus Directional Indicator / Movement
        # dataframe['minus_dm'] = ta.MINUS_DM(dataframe)
        # dataframe['minus_di'] = ta.MINUS_DI(dataframe)

        # # Aroon, Aroon Oscillator
        # aroon = ta.AROON(dataframe)
        # dataframe['aroonup'] = aroon['aroonup']
        # dataframe['aroondown'] = aroon['aroondown']
        # dataframe['aroonosc'] = ta.AROONOSC(dataframe)

        # # Awesome Oscillator
        # dataframe['ao'] = qtpylib.awesome_oscillator(dataframe)

        # # Keltner Channel
        # keltner = qtpylib.keltner_channel(dataframe)
        # dataframe["kc_upperband"] = keltner["upper"]
        # dataframe["kc_lowerband"] = keltner["lower"]
        # dataframe["kc_middleband"] = keltner["mid"]
        # dataframe["kc_percent"] = (
        #     (dataframe["close"] - dataframe["kc_lowerband"]) /
        #     (dataframe["kc_upperband"] - dataframe["kc_lowerband"])
        # )
        # dataframe["kc_width"] = (
        #     (dataframe["kc_upperband"] - dataframe["kc_lowerband"]) / dataframe["kc_middleband"]
        # )

        # # Ultimate Oscillator
        # dataframe['uo'] = ta.ULTOSC(dataframe)

        # # Commodity Channel Index: values [Oversold:-100, Overbought:100]
        # dataframe['cci'] = ta.CCI(dataframe)

        # RSI
        dataframe['rsi'] = ta.RSI(dataframe)

        # # Inverse Fisher transform on RSI: values [-1.0, 1.0] (https://goo.gl/2JGGoy)
        # rsi = 0.1 * (dataframe['rsi'] - 50)
        # dataframe['fisher_rsi'] = (np.exp(2 * rsi) - 1) / (np.exp(2 * rsi) + 1)

        # # Inverse Fisher transform on RSI normalized: values [0.0, 100.0] (https://goo.gl/2JGGoy)
        # dataframe['fisher_rsi_norma'] = 50 * (dataframe['fisher_rsi'] + 1)

        # # Stochastic Slow
        # stoch = ta.STOCH(dataframe)
        # dataframe['slowd'] = stoch['slowd']
        # dataframe['slowk'] = stoch['slowk']

        # Stochastic Fast
        stoch_fast = ta.STOCHF(dataframe)
        dataframe['fastd'] = stoch_fast['fastd']
        dataframe['fastk'] = stoch_fast['fastk']

        # # Stochastic RSI
        # Please read https://github.com/freqtrade/freqtrade/issues/2961 before using this.
        # STOCHRSI is NOT aligned with tradingview, which may result in non-expected results.
        # stoch_rsi = ta.STOCHRSI(dataframe)
        # dataframe['fastd_rsi'] = stoch_rsi['fastd']
        # dataframe['fastk_rsi'] = stoch_rsi['fastk']

        # MACD
        macd = ta.MACD(dataframe)
        dataframe['macd'] = macd['macd']
        dataframe['macdsignal'] = macd['macdsignal']
        dataframe['macdhist'] = macd['macdhist']

        # MFI
        dataframe['mfi'] = ta.MFI(dataframe)

        # # ROC
        # dataframe['roc'] = ta.ROC(dataframe)

        # Overlap Studies
        # ------------------------------------

        # Bollinger Bands
        bollinger = qtpylib.bollinger_bands(qtpylib.typical_price(dataframe), window=20, stds=2)
        dataframe['bb_lowerband'] = bollinger['lower']
        dataframe['bb_middleband'] = bollinger['mid']
        dataframe['bb_upperband'] = bollinger['upper']
        dataframe["bb_percent"] = (
            (dataframe["close"] - dataframe["bb_lowerband"]) /
            (dataframe["bb_upperband"] - dataframe["bb_lowerband"])
        )
        dataframe["bb_width"] = (
            (dataframe["bb_upperband"] - dataframe["bb_lowerband"]) / dataframe["bb_middleband"]
        )

        # Bollinger Bands - Weighted (EMA based instead of SMA)
        # weighted_bollinger = qtpylib.weighted_bollinger_bands(
        #     qtpylib.typical_price(dataframe), window=20, stds=2
        # )
        # dataframe["wbb_upperband"] = weighted_bollinger["upper"]
        # dataframe["wbb_lowerband"] = weighted_bollinger["lower"]
        # dataframe["wbb_middleband"] = weighted_bollinger["mid"]
        # dataframe["wbb_percent"] = (
        #     (dataframe["close"] - dataframe["wbb_lowerband"]) /
        #     (dataframe["wbb_upperband"] - dataframe["wbb_lowerband"])
        # )
        # dataframe["wbb_width"] = (
        #     (dataframe["wbb_upperband"] - dataframe["wbb_lowerband"]) / dataframe["wbb_middleband"]
        # )

        # # EMA - Exponential Moving Average
        # dataframe['ema3'] = ta.EMA(dataframe, timeperiod=3)
        # dataframe['ema5'] = ta.EMA(dataframe, timeperiod=5)
        # dataframe['ema10'] = ta.EMA(dataframe, timeperiod=10)
        # dataframe['ema21'] = ta.EMA(dataframe, timeperiod=21)
        # dataframe['ema50'] = ta.EMA(dataframe, timeperiod=50)
        # dataframe['ema100'] = ta.EMA(dataframe, timeperiod=100)

        # # SMA - Simple Moving Average
        # dataframe['sma3'] = ta.SMA(dataframe, timeperiod=3)
        # dataframe['sma5'] = ta.SMA(dataframe, timeperiod=5)
        # dataframe['sma10'] = ta.SMA(dataframe, timeperiod=10)
        # dataframe['sma21'] = ta.SMA(dataframe, timeperiod=21)
        # dataframe['sma50'] = ta.SMA(dataframe, timeperiod=50)
        # dataframe['sma100'] = ta.SMA(dataframe, timeperiod=100)

        # Parabolic SAR
        dataframe['sar'] = ta.SAR(dataframe)

        # TEMA - Triple Exponential Moving Average
        dataframe['tema'] = ta.TEMA(dataframe, timeperiod=9)

        # Cycle Indicator
        # ------------------------------------
        # Hilbert Transform Indicator - SineWave
        hilbert = ta.HT_SINE(dataframe)
        dataframe['htsine'] = hilbert['sine']
        dataframe['htleadsine'] = hilbert['leadsine']

        # Pattern Recognition - Bullish candlestick patterns
        # ------------------------------------
        # # Hammer: values [0, 100]
        # dataframe['CDLHAMMER'] = ta.CDLHAMMER(dataframe)
        # # Inverted Hammer: values [0, 100]
        # dataframe['CDLINVERTEDHAMMER'] = ta.CDLINVERTEDHAMMER(dataframe)
        # # Dragonfly Doji: values [0, 100]
        # dataframe['CDLDRAGONFLYDOJI'] = ta.CDLDRAGONFLYDOJI(dataframe)
        # # Piercing Line: values [0, 100]
        # dataframe['CDLPIERCING'] = ta.CDLPIERCING(dataframe) # values [0, 100]
        # # Morningstar: values [0, 100]
        # dataframe['CDLMORNINGSTAR'] = ta.CDLMORNINGSTAR(dataframe) # values [0, 100]
        # # Three White Soldiers: values [0, 100]
        # dataframe['CDL3WHITESOLDIERS'] = ta.CDL3WHITESOLDIERS(dataframe) # values [0, 100]

        # Pattern Recognition - Bearish candlestick patterns
        # ------------------------------------
        # # Hanging Man: values [0, 100]
        # dataframe['CDLHANGINGMAN'] = ta.CDLHANGINGMAN(dataframe)
        # # Shooting Star: values [0, 100]
        # dataframe['CDLSHOOTINGSTAR'] = ta.CDLSHOOTINGSTAR(dataframe)
        # # Gravestone Doji: values [0, 100]
        # dataframe['CDLGRAVESTONEDOJI'] = ta.CDLGRAVESTONEDOJI(dataframe)
        # # Dark Cloud Cover: values [0, 100]
        # dataframe['CDLDARKCLOUDCOVER'] = ta.CDLDARKCLOUDCOVER(dataframe)
        # # Evening Doji Star: values [0, 100]
        # dataframe['CDLEVENINGDOJISTAR'] = ta.CDLEVENINGDOJISTAR(dataframe)
        # # Evening Star: values [0, 100]
        # dataframe['CDLEVENINGSTAR'] = ta.CDLEVENINGSTAR(dataframe)

        # Pattern Recognition - Bullish/Bearish candlestick patterns
        # ------------------------------------
        # # Three Line Strike: values [0, -100, 100]
        # dataframe['CDL3LINESTRIKE'] = ta.CDL3LINESTRIKE(dataframe)
        # # Spinning Top: values [0, -100, 100]
        # dataframe['CDLSPINNINGTOP'] = ta.CDLSPINNINGTOP(dataframe) # values [0, -100, 100]
        # # Engulfing: values [0, -100, 100]
        # dataframe['CDLENGULFING'] = ta.CDLENGULFING(dataframe) # values [0, -100, 100]
        # # Harami: values [0, -100, 100]
        # dataframe['CDLHARAMI'] = ta.CDLHARAMI(dataframe) # values [0, -100, 100]
        # # Three Outside Up/Down: values [0, -100, 100]
        # dataframe['CDL3OUTSIDE'] = ta.CDL3OUTSIDE(dataframe) # values [0, -100, 100]
        # # Three Inside Up/Down: values [0, -100, 100]
        # dataframe['CDL3INSIDE'] = ta.CDL3INSIDE(dataframe) # values [0, -100, 100]

        # # Chart type
        # # ------------------------------------
        # # Heikin Ashi Strategy
        # heikinashi = qtpylib.heikinashi(dataframe)
        # dataframe['ha_open'] = heikinashi['open']
        # dataframe['ha_close'] = heikinashi['close']
        # dataframe['ha_high'] = heikinashi['high']
        # dataframe['ha_low'] = heikinashi['low']

        # Retrieve best bid and best ask from the orderbook
        # ------------------------------------
        """
        # first check if dataprovider is available
        if self.dp:
            if self.dp.runmode.value in ('live', 'dry_run'):
                ob = self.dp.orderbook(metadata['pair'], 1)
                dataframe['best_bid'] = ob['bids'][0][0]
                dataframe['best_ask'] = ob['asks'][0][0]
        """

        return dataframe

    def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
        
        dataframe['enter_short'] = 0
        dataframe['enter_long'] = 0

        for e in range(1, self.num_episode + 1):
            logging.info(f'\n------Episode: {e}/{self.num_episode}-------')

            self.agent.reset() # reset to initial balance and hyperparameters

            for i in range(0, dataframe.shape[0]):
                if(i > self.window_size):
                    state = generate_combined_state(i, self.window_size, dataframe.iloc[:, dataframe.columns.get_loc('close')])
                    action = np.argmax(self.agent.model.predict(state, verbose = 0)[0])
                    random_action = self.agent.act(state)

                    if action != np.argmax(actions): 
                        logging.info(f"\t\t'{self.action_dict[action]}' is an exploration.")
                    if action == 0: # hold
                        execution_result = hold(actions)
                    if action == 1: # buy
                        execution_result = buy(t)      
                    if action == 2: # sell
                        execution_result = sell(t)   



################

            for t in range(1, trading_period + 1):

                reward = 0
                next_state = generate_combined_state(t, window_size, stock_prices, agent.balance, len(agent.inventory))

                actions = agent.model.predict(state)[0]
                action = agent.act(state)
                
                if action != np.argmax(actions): logging.info(f"\t\t'{action_dict[action]}' is an exploration.")
                if action == 0: # hold
                    execution_result = hold(actions)
                if action == 1: # buy
                    execution_result = buy(t)      
                if action == 2: # sell
                    execution_result = sell(t)        
                
                # check execution result
                if execution_result is None:
                    reward -= average_interest_rate() * agent.balance  # missing opportunity
                else:
                    if isinstance(execution_result, tuple): # if execution_result is 'Hold'
                        actions = execution_result[1]
                        execution_result = execution_result[0]
                    logging.info(execution_result) 

                # calculate reward
                current_portfolio_value = len(agent.inventory) * stock_prices[t] + agent.balance
                unrealized_profit = current_portfolio_value - agent.initial_portfolio_value
                reward += unrealized_profit

                agent.portfolio_values.append(current_portfolio_value)
                #return daily P/L rates
                agent.return_rates.append((current_portfolio_value - previous_portfolio_value) / previous_portfolio_value)

                done = True if t == trading_period else False
                agent.remember(state, actions, reward, next_state, done)

                # update state
                state = next_state

                # experience replay
                if len(agent.memory) > agent.buffer_size:
                    num_experience_replay += 1
                    loss = agent.experience_replay()
                    logging.info('Episode: {}\tLoss: {:.2f}\tAction: {}\tReward: {:.2f}\tBalance: {:.2f}\tNumber of Stocks: {}'.format(e, loss, action_dict[action], reward, agent.balance, len(agent.inventory)))

                if done:
                    portfolio_return = evaluate_portfolio_performance(agent, logging)
                    returns_across_episodes.append(portfolio_return)

            # save models every episode
            agent.model.save('saved_models/DQN_ep' + str(e) + '.h5')
            logging.info('model saved')
            plot_portfolio_performance_comparison(stock_name, agent)
            #print("Training Complete")
       

        for i in range(0, dataframe.shape[0]):
            # logger.info(dataframe.iloc[:, dataframe.columns.get_loc('close')])
            if(i > self.window_size):
                state = generate_combined_state(i, self.window_size, dataframe.iloc[:, dataframe.columns.get_loc('close')])

                action = np.argmax(self.agent.model.predict(state, verbose = 0)[0])
                
                if action == 1: #Buy
                    #print("Buy Taken")
                    dataframe.iloc[i, dataframe.columns.get_loc('enter_long')] = 1
                if action == 2: #Sell
                    pass
                    # #print("Sell Taken")
                    # dataframe.iloc[i, dataframe.columns.get_loc('enter_short')] = 1

        return dataframe

    def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
        """
        Based on TA indicators, populates the exit signal for the given dataframe
        :param dataframe: DataFrame
        :param metadata: Additional information, like the currently traded pair
        :return: DataFrame with exit columns populated
        """
        dataframe.loc[
            (
                (qtpylib.crossed_above(dataframe['rsi'], self.sell_rsi.value)) &  # Signal: RSI crosses above sell_rsi
                (dataframe['tema'] > dataframe['bb_middleband']) &  # Guard: tema above BB middle
                (dataframe['tema'] < dataframe['tema'].shift(1)) &  # Guard: tema is falling
                (dataframe['volume'] > 0)  # Make sure Volume is not 0
            ),
            'exit_long'] = 1
        # Uncomment to use shorts (Only used in futures/margin mode. Check the documentation for more info)
        """
        dataframe.loc[
            (
                (qtpylib.crossed_above(dataframe['rsi'], self.buy_rsi.value)) &  # Signal: RSI crosses above buy_rsi
                (dataframe['tema'] <= dataframe['bb_middleband']) &  # Guard: tema below BB middle
                (dataframe['tema'] > dataframe['tema'].shift(1)) &  # Guard: tema is raising
                (dataframe['volume'] > 0)  # Make sure Volume is not 0
            ),
            'exit_short'] = 1
        """
        return dataframe
    
