from datetime import timedelta
from functools import reduce
import freqtrade.vendor.qtpylib.indicators as qtpylib
import talib.abstract as ta
from freqtrade.strategy import IStrategy
from pandas import DataFrame

def to_minutes(**timdelta_kwargs):
    return int(timedelta(**timdelta_kwargs).total_seconds() / 60)

class Saturn5(IStrategy):
    INTERFACE_VERSION = 3
    timeframe = '15m'
    # Stoploss
    stoploss = -0.2
    startup_candle_count: int = 480
    trailing_stop = False
    use_custom_stoploss = False
    use_exit_signal = False
    # signal controls
    entry_signal_1 = True
    entry_signal_2 = True
    entry_signal_3 = True
    # ROI table:
    minimal_roi = {'0': 0.05}
    # Indicator values:
    # Signal 1
    s1_ema_xs = 3
    s1_ema_sm = 5
    s1_ema_md = 10
    s1_ema_xl = 50
    s1_ema_xxl = 240
    # Signal 2
    s2_ema_input = 50
    s2_ema_offset_input = -1
    s2_bb_sma_length = 49
    s2_bb_std_dev_length = 64
    s2_bb_lower_offset = 3
    s2_fib_sma_len = 50
    s2_fib_atr_len = 14
    s2_fib_lower_value = 4.236

    @property
    def protections(self):
        # Don't enter a trade right after exiting a trade.
        # Stop trading if max-drawdown is reached.
        # Considering all pairs that have a minimum of 20 trades
        # If max-drawdown is > 20% this will activate
        # Stop trading if a certain amount of stoploss occurred within a certain time window.
        # Considering all pairs that have a minimum of 4 trades
        # Looks at all pairs
        # Lock pairs with low profits
        # Considering all pairs that have a minimum of 2 trades
        # If profit < 2% this will activate for a pair
        # Lock pairs with low profits
        # Considering all pairs that have a minimum of 4 trades
        # If profit < 1% this will activate for a pair
        return [{'method': 'CooldownPeriod', 'stop_duration': to_minutes(hours=1, minutes=15)}, {'method': 'MaxDrawdown', 'lookback_period': to_minutes(hours=12), 'trade_limit': 20, 'stop_duration': to_minutes(hours=1), 'max_allowed_drawdown': 0.2}, {'method': 'StoplossGuard', 'lookback_period': to_minutes(hours=6), 'trade_limit': 4, 'stop_duration': to_minutes(minutes=30), 'only_per_pair': False}, {'method': 'LowProfitPairs', 'lookback_period': to_minutes(hours=1, minutes=30), 'trade_limit': 2, 'stop_duration': to_minutes(hours=15), 'required_profit': 0.02}, {'method': 'LowProfitPairs', 'lookback_period': to_minutes(hours=6), 'trade_limit': 4, 'stop_duration': to_minutes(minutes=30), 'required_profit': 0.01}]

    def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
        # Adding EMA's into the dataframe
        dataframe['s1_ema_xs'] = ta.EMA(dataframe, timeperiod=self.s1_ema_xs)
        dataframe['s1_ema_sm'] = ta.EMA(dataframe, timeperiod=self.s1_ema_sm)
        dataframe['s1_ema_md'] = ta.EMA(dataframe, timeperiod=self.s1_ema_md)
        dataframe['s1_ema_xl'] = ta.EMA(dataframe, timeperiod=self.s1_ema_xl)
        dataframe['s1_ema_xxl'] = ta.EMA(dataframe, timeperiod=self.s1_ema_xxl)
        s2_ema_value = ta.EMA(dataframe, timeperiod=self.s2_ema_input)
        s2_ema_xxl_value = ta.EMA(dataframe, timeperiod=200)
        dataframe['s2_ema'] = s2_ema_value - s2_ema_value * self.s2_ema_offset_input
        dataframe['s2_ema_xxl_off'] = s2_ema_xxl_value - s2_ema_xxl_value * self.s2_fib_lower_value
        dataframe['s2_ema_xxl'] = ta.EMA(dataframe, timeperiod=200)
        s2_bb_sma_value = ta.SMA(dataframe, timeperiod=self.s2_bb_sma_length)
        s2_bb_std_dev_value = ta.STDDEV(dataframe, self.s2_bb_std_dev_length)
        dataframe['s2_bb_std_dev_value'] = s2_bb_std_dev_value
        dataframe['s2_bb_lower_band'] = s2_bb_sma_value - s2_bb_std_dev_value * self.s2_bb_lower_offset
        s2_fib_atr_value = ta.ATR(dataframe, timeframe=self.s2_fib_atr_len)
        s2_fib_sma_value = ta.SMA(dataframe, timeperiod=self.s2_fib_sma_len)
        dataframe['s2_fib_lower_band'] = s2_fib_sma_value - s2_fib_atr_value * self.s2_fib_lower_value
        s3_bollinger = qtpylib.bollinger_bands(qtpylib.typical_price(dataframe), window=20, stds=3)
        dataframe['s3_bb_lowerband'] = s3_bollinger['lower']
        # Volume weighted MACD
        dataframe['fastMA'] = ta.EMA(dataframe['volume'] * dataframe['close'], 12) / ta.EMA(dataframe['volume'], 12)
        dataframe['slowMA'] = ta.EMA(dataframe['volume'] * dataframe['close'], 26) / ta.EMA(dataframe['volume'], 26)
        dataframe['vwmacd'] = dataframe['fastMA'] - dataframe['slowMA']
        dataframe['signal'] = ta.EMA(dataframe['vwmacd'], 9)
        dataframe['hist'] = dataframe['vwmacd'] - dataframe['signal']
        return dataframe

    def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
        # basic entry methods to keep the strategy simple
        if self.entry_signal_1:
            conditions = [dataframe['vwmacd'] < dataframe['signal'], dataframe['low'] < dataframe['s1_ema_xxl'], dataframe['close'] < dataframe['s1_ema_xxl'], qtpylib.crossed_above(dataframe['s1_ema_sm'], dataframe['s1_ema_md']), dataframe['s1_ema_xs'] < dataframe['s1_ema_xl'], dataframe['volume'] > 0]
            dataframe.loc[reduce(lambda x, y: x & y, conditions), ['enter_long', 'enter_tag']] = (1, 'entry_signal_1')
        if self.entry_signal_2:
            conditions = [qtpylib.crossed_above(dataframe['s2_fib_lower_band'], dataframe['s2_bb_lower_band']), dataframe['close'] < dataframe['s2_ema'], dataframe['volume'] > 0]
            dataframe.loc[reduce(lambda x, y: x & y, conditions), ['enter_long', 'enter_tag']] = (1, 'entry_signal_2')
        if self.entry_signal_3:
            conditions = [dataframe['low'] < dataframe['s3_bb_lowerband'], dataframe['low'] > dataframe['s1_ema_xxl'], dataframe['volume'] > 0]
            dataframe.loc[reduce(lambda x, y: x & y, conditions), ['enter_long', 'enter_tag']] = (1, 'entry_signal_3')
        if not all([self.entry_signal_1, self.entry_signal_2, self.entry_signal_3]):
            dataframe.loc[(), 'enter_long'] = 0
        return dataframe

    def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
        # This is essentailly ignored as we're using strict ROI / Stoploss / TTP sale scenarios
        dataframe.loc[(), 'exit_long'] = 0
        return dataframe