# source: https://raw.githubusercontent.com/jaredrsommer/freqtradestrategies/63cd46ea087ed8ed01f7d54d8cfcf1dbed166b04/tsp0chickenV2.py
# pragma pylint: disable=W0105, C0103, C0114, C0115, C0116, C0301, C0302, C0303, C0325, C0411, C0413,  W1203, W291

"""
####################################################################################
EQUAL  - base class for 'simple' time series prediction
             Handles most of the logic for time series prediction. Subclasses should
             override the model-related functions

             Note that I use gain rather than price because it is a normalised value, and works better with prediction
             algorithms. I use the actual (future) gain to train a base model, which is then further refined for each
             individual pair.
             The model is created if it does not exist, and is trained on all available data before being saved.
             Models are saved in user_data/strategies/GYN/models/<class>/<class>.sav, where <class> is the name
             of the current class (Shifted if running this directly, or the name of the subclass).
             If the model already exits, then it is just loaded and used.
             So, it makes sense to do initial training over a long period of time to create the base model.
             If training, then no backtesting or tuning for individual pairs is performed (way faster).
             If you want to retrain (e.g. you changed indicators), then delete the model and run the strategy over a
             long time period

####################################################################################
"""

import copy
import cProfile
import os
import pstats
from tqdm import tqdm
import sys
import traceback
from datetime import datetime
from functools import reduce
from pathlib import Path
from typing import Optional
import random
import logging
import warnings
import time
import joblib
import numpy as np
import pandas as pd
import pywt
from pandas import DataFrame, Series
import talib.abstract as ta
import finta
import technical.indicators as ftt
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.preprocessing import RobustScaler
from freqtrade import leverage
import freqtrade.vendor.qtpylib.indicators as qtpylib
from freqtrade.persistence import Trade
import freqtrade.vendor.qtpylib.indicators as qtpylib
from freqtrade.strategy import CategoricalParameter, DecimalParameter, IntParameter, IStrategy
from lightgbm import LGBMRegressor
from sklearn.linear_model import PassiveAggressiveRegressor, SGDRegressor
from xgboost import XGBRegressor

group_dir = str(Path(__file__).parent)
strat_dir = str(Path(__file__).parent.parent)
sys.path.append(strat_dir)
sys.path.append(group_dir)

import utils.custom_indicators as cta
import utils.Wavelets as Wavelets
import utils.Forecasters as Forecasters
from utils.DataframeUtils import DataframeUtils, ScalerType  # pylint: disable=E0401

logger = logging.getLogger(__name__)

# log.setLevel(logging.DEBUG)
warnings.simplefilter(action="ignore", category=pd.errors.PerformanceWarning)
warnings.simplefilter(action="ignore", category=FutureWarning)
warnings.simplefilter(action="ignore", category=UserWarning)

pd.options.mode.chained_assignment = None  # default='warn'

class Github_jaredrsommer_freqtradestrategies__tsp0chickenV2__20260509_225308(IStrategy):
    # Do *not* hyperopt for the roi and stoploss spaces

    plot_config = {
        "main_plot": {
            "close": {"color": "cornflowerblue"},
        },
        "subplots": {
            "Diff": {
                "predicted_gain": {"color": "purple"},
                "shifted_pred": {"color": "skyblue"},
                # "squeeze": {"color": "red"},
                "gain": {"color": "green"},
                "target_profit": {"color": "lightgreen"},
                "target_loss": {"color": "lightsalmon"},
                "buy_region": {"color": "darkseagreen"},
                "sell_region": {"color": "darksalmon"},
            },
        },
    }

    # ROI table:
    minimal_roi = {"0": 0.30, "30": 0.20, "60": 0.10, "80": 0.05,"100": 0.025}

    # Stoploss:
    stoploss = -0.20

    # Trailing stop:s
    trailing_stop = False
    trailing_stop_positive = None
    trailing_stop_positive_offset = 0.0
    trailing_only_offset_is_reached = False

    timeframe = "5m"
    use_custom_stoploss = False   
    can_short = False
    leverage_input = 15.0
    
    
    # if setting can-short to True, remember to update the config file:
    #   "trading_mode": "futures",
    #   "margin_mode": "isolated",

    # Recommended
    use_exit_signal = True
    exit_profit_only = False
    ignore_roi_if_entry_signal = True

    # Required
    startup_candle_count: int = 128  # must be power of 2
    process_only_new_candles = True
    custom_trade_info = {}  # pair-specific data
    curr_pair = ""

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

    # Strategy Specific Variable Storage

    ## Hyperopt Variables

    lookahead = 0

    df_coeffs: DataFrame = None
    coeff_table = None
    coeff_array = None
    gain_data = None
    merge_indicators = False  # set to False to not merge indicators into prediction data

    use_rolling = False  # True = rolling (slow but realistic), False = Jumping (much faster, less realistic)
    single_col_prediction = False  # True = use only gain. False = use all columns (better, but much slower)

    wavelet_type: Wavelets.WaveletType = Wavelets.WaveletType.DWT
    wavelet = None

    forecaster_type: Forecasters.ForecasterType = Forecasters.ForecasterType.PA
    # forecaster_type:Forecasters.ForecasterType = Forecasters.ForecasterType.SGD
    # forecaster_type:Forecasters.ForecasterType = Forecasters.ForecasterType.SVR
    forecaster = None
    data = None
    dc = 0
    har = []
    wavelet_size = 64  # needed for consistently-sized transforms
    win_size = wavelet_size  # this can vary (64)

    train_min_len = wavelet_size  # longer = slower (64)
    train_len = min(128, wavelet_size * 4)  # longer = slower 256 (64*4)
    # scale_len = wavelet_size // 2 # no. recent candles to use when scaling
    scale_len = min(8, wavelet_size // 2)  # no. recent candles to use when scaling (64/2)
    win_size = min(32, wavelet_size)  # (64)
    model_window = wavelet_size  # longer = slower (64)

    # setting this to some form of harmonic ptp
    profit_nstd = 1.0
    loss_nstd = 1.0

    training_data = None
    training_labels = None
    training_mode = False  # do not set manually
    supports_incremental_training = True
    model_per_pair = False
    combine_models = True
    model_trained = False
    new_model = False
    detrend_data = False
    scale_results = False
    norm_data = False  # changing this requires new models
    dataframeUtils = None
    scaler = RobustScaler()
    model = None
    base_forecaster = None

    curr_dataframe: DataFrame = None

    target_profit = 0.0
    target_loss = 0.0

    # hyperparams

    # Buy hyperspace params:
    buy_params = {
        "cexit_min_profit_th": 0.3,
        "cexit_profit_nstd": 1.9,
        "enable_bb_check": False,
        "enable_guard_metric": True,
        "enable_squeeze": False,
        "entry_bb_factor": 0.82,
        "entry_bb_width": 0.025,
        "entry_guard_metric": -0.3,
    }

    # Sell hyperspace params:
    sell_params = {
        "cexit_loss_nstd": 1.8,
        "cexit_metric_overbought": 0.76,
        "cexit_metric_take_profit": 0.94,
        "cexit_min_loss_th": -0.3,
        "enable_exit_signal": False,
        "exit_bb_factor": 1.01,
        "exit_guard_metric": 0.7,
    }

    # FFT
    upper_limit = IntParameter(40, 80, default=40, space="buy", load=True, optimize=True)
    dc_x = DecimalParameter(low=3.0, high=5.0, default=4.5, decimals=1 ,space='buy', optimize=True, load=True)

    # Entry
    # the following flags apply to both entry and exit
    enable_guard_metric = CategoricalParameter([True, False], default=True, space="buy", load=True, optimize=False)
    enable_bb_check = CategoricalParameter([True, False], default=True, space="buy", load=True, optimize=True)
    enable_squeeze = CategoricalParameter([True, False], default=True, space="buy", load=True, optimize=True)
    entry_guard_metric = DecimalParameter(-0.8, 0.0, default=-0.2, decimals=1, space="buy", load=True, optimize=True)
    entry_bb_width = DecimalParameter(0.020, 0.100, default=0.02, decimals=3, space="buy", load=True, optimize=True)
    entry_bb_factor = DecimalParameter(0.70, 1.20, default=1.1, decimals=2, space="buy", load=True, optimize=True)

    # Exit
    # use exit signal? If disabled, just rely on the custom exit checks (or stoploss) to get out
    enable_exit_signal = CategoricalParameter([True, False], default=True, space="sell", load=True, optimize=False)
    exit_guard_metric = DecimalParameter(0.0, 0.8, default=0.0, decimals=1, space="sell", load=True, optimize=True)
    exit_bb_factor = DecimalParameter(0.70, 1.20, default=0.8, decimals=2, space="sell", load=True, optimize=True)

    # Custom Exit
    # No. Standard Deviations of profit/loss for target, and lower limit
    cexit_min_profit_th = DecimalParameter(0.0, 1.5, default=0.7, decimals=1, space="buy", load=True, optimize=True)
    cexit_profit_nstd = DecimalParameter(0.0, 3.0, default=0.9, decimals=1, space="buy", load=True, optimize=True)

    cexit_min_loss_th = DecimalParameter(-1.5, -0.0, default=-0.4, decimals=1, space="sell", load=True, optimize=True)
    cexit_loss_nstd = DecimalParameter(0.0, 3.0, default=0.7, decimals=1, space="sell", load=True, optimize=True)

    # Guard metric sell limits - used to bail out when in profit
    cexit_metric_overbought = DecimalParameter(0.55, 0.99, default=0.96, decimals=2, space="sell", load=True, optimize=True)
    cexit_metric_take_profit = DecimalParameter(0.55, 0.99, default=0.76, decimals=2, space="sell", load=True, optimize=True)

    ###################################
    
    @property
    def protections(self):
        return [
            {
                "method": "CooldownPeriod",
                "stop_duration_candles": 5
            },
            {
                "method": "MaxDrawdown",
                "lookback_period_candles": 48,
                "trade_limit": 20,
                "stop_duration_candles": 4,
                "max_allowed_drawdown": 0.2
            },
            {
                "method": "StoplossGuard",
                "lookback_period_candles": 24,
                "trade_limit": 4,
                "stop_duration_candles": 2,
                "only_per_pair": False
            },
            {
                "method": "LowProfitPairs",
                "lookback_period_candles": 6,
                "trade_limit": 2,
                "stop_duration_candles": 60,
                "required_profit": 0.02
            },
            {
                "method": "LowProfitPairs",
                "lookback_period_candles": 24,
                "trade_limit": 4,
                "stop_duration_candles": 2,
                "required_profit": 0.01
            }
        ]
        
    
    def leverage(self, pair: str, current_time: datetime, current_rate: float,
                 proposed_leverage: float, max_leverage: float, entry_tag: Optional[str], side: str,
                 **kwargs) -> float:
        if self.leverage_input > max_leverage:
            return max_leverage

        return self.leverage_input
        
    def bot_start(self, **kwargs) -> None:
        if self.dataframeUtils is None:
            self.dataframeUtils = DataframeUtils()
            self.dataframeUtils.set_scaler_type(ScalerType.Robust)

        if self.wavelet is None:
            self.wavelet = Wavelets.make_wavelet(self.wavelet_type)

        if self.forecaster is None:
            self.forecaster = Forecasters.make_forecaster(self.forecaster_type)
            self.forecaster.set_detrend(self.detrend_data)

        if (not self.forecaster.supports_multiple_columns()) and (not self.single_col_prediction):
            logger.info("    ****")
            logger.info(f"    **** ERROR: forecaster ({self.forecaster_type.name}) does not support multiple indicators")
            logger.info("    ****")

        if not self.forecaster.supports_retrain():
            logger.info("    ****")
            logger.info(f"    **** WARNING: forecaster ({self.forecaster_type.name}) does not support retrainings")
            logger.info("    ****")

        # Just enable variables here, we overwrite them later
        self.win_size = self.wavelet_size  # this can vary
        self.train_min_len = self.wavelet_size  # longer = slower
        self.train_len = min(128, self.wavelet_size * 4)  # longer = slower
        # scale_len = wavelet_size // 2 # no. recent candles to use when scaling
        self.scale_len = min(16, self.wavelet_size // 2)  # no. recent candles to use when scaling
        self.win_size = min(32, self.wavelet_size)
        self.model_window = self.wavelet_size  # longer = slower

        logger.info("")
        logger.info(f"    wavelet_type:    {self.wavelet_type.name} ({self.wavelet_size})")
        logger.info(f"    win_size:        {self.win_size}")
        logger.info(f"    forecaster_type: {self.forecaster.get_name()}")
        logger.info(f"    detrend_data:    {self.forecaster.detrend_data}")
        logger.info("")
        return


    # update saved data based on current pairlist
    def update_pairlist_data(self):
        # this only makes sense in 'live' modes
        if self.dp.runmode.value in ("backtest", "plot", "hyperopt"):
            return

        # current pairlist
        curr_pairlist = np.array(self.dp.current_whitelist())

        # pairlist from previous calls
        saved_pairlist = np.array(list(self.custom_trade_info.keys()))

        # get the pairs that are no longer in the list
        removed_pairs = np.setdiff1d(saved_pairlist, curr_pairlist)
        added_pairs = np.setdiff1d(curr_pairlist, saved_pairlist)

        if len(removed_pairs) > 0:
            logger.info("    Pairlist changed:")
            logger.info(f"    old pairs: {saved_pairlist}")
            logger.info(f"    new pairs: {curr_pairlist}")
            logger.info(f"    pairs removed: {removed_pairs}")
            logger.info(f"    pairs added: {added_pairs}")

            for pair in removed_pairs:
                logger.info(f"    Removing historical data for: {pair}")
                del self.custom_trade_info[pair]

    ###################################
    """
    Indicator Definitions
    """

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

        # Ch1cK3N t4c0s 09/2/2024 - Use FFT to find the optimal wavelet size for each pair.
        # Use Hurst Dominant Cycle and Harmonics to generate envelopes and calculate wavelet 
        # average movement for trailing stoploss.

        # NOTE: if you change the indicators, you need to regenerate the model

        start_time = time.time()
        pair = metadata['pair']
        dataframe = dataframe.copy()
        dataframe['zero'] = 0
        dataframe['ha_close'] = (dataframe['high'] + dataframe['low'] + dataframe['close']) / 3
        dataframe['ha_open'] = dataframe['ha_close'].shift(1)

        dataframe['ha_high'] = dataframe[['high', 'ha_open', 'ha_close']].max(axis=1)
        dataframe['ha_low'] = dataframe[['low', 'ha_open', 'ha_close']].min(axis=1)

        if self.dp.runmode.value in ('dry_run'):
            window_size = 200 #self.window_size.value  # Adjust this value as appropriate
        else:
            window_size = None

        if len(dataframe) < 200: #self.window_size.value:
            raise ValueError(f"Insufficient data points for FFT: {len(dataframe)}. Need at least {window_size} data points.")


        # Perform FFT to identify cycles with a rolling window
        freq, power = perform_fft(dataframe['ha_close'], window_size=window_size)

        if len(freq) == 0 or len(power) == 0:
            raise ValueError("FFT resulted in zero or invalid frequencies. Check the data or the FFT implementation.")

        # Filter out the zero-frequency component and limit the frequency to below 500
        positive_mask = (freq > 0) & (1 / freq < self.upper_limit.value)
        positive_freqs = freq[positive_mask]
        positive_power = power[positive_mask]

        # Convert frequencies to periods
        cycle_periods = 1 / positive_freqs

        # Set a threshold to filter out insignificant cycles based on power
        power_threshold = 0.01 * np.max(positive_power)
        significant_indices = positive_power > power_threshold
        significant_periods = cycle_periods[significant_indices]
        significant_power = positive_power[significant_indices]

        # Identify the dominant cycle
        dominant_freq_index = np.argmax(significant_power)
        dominant_freq = positive_freqs[dominant_freq_index]
        cycle_period = int(np.abs(1 / dominant_freq)) if dominant_freq != 0 else np.inf

        if cycle_period == np.inf:
            raise ValueError("No dominant frequency found. Check the data or the method used.")

        # Calculate harmonics for the dominant cycle
        harmonics = [cycle_period / (i + 1) for i in range(1, 4)]
        dataframe['dc_EWM'] = dataframe['ha_close'].ewm(span=int(cycle_period)).mean()
        dataframe['dc_1/2'] = dataframe['ha_close'].ewm(span=int(harmonics[0])).mean()
        dataframe['dc_1/3'] = dataframe['ha_close'].ewm(span=int(harmonics[1])).mean()
        dataframe['dc_1/4'] = dataframe['ha_close'].ewm(span=int(harmonics[2])).mean()

        # Calculate the rolling difference (ΔY)
        dataframe['diff'] = dataframe['dc_EWM'].diff(periods=2)
        delta_x = 2
        
        # Calculate the slope
        dataframe['slope'] = dataframe['diff'] / delta_x
        
        # Convert slope to angle (in radians), then convert to degrees
        dataframe['angle'] = np.degrees(np.arctan(dataframe['slope']))

        # Fractional Brownian Motion (fBm)
        n = len(dataframe['dc_EWM'])
        h = 0.5
        t = 1
        dt = t / n
        fBm = np.zeros(n)
        for i in range(1, n):
            fBm[i] = fBm[i-1] + np.sqrt(dt) * np.random.normal(0, 1)
        dataframe['fBm'] = fBm * (dt ** h)

        # Fractional differentiation
        dataframe['frac_diff'] = np.gradient(dataframe['ha_close'])
        dataframe['fBm_mean'] = np.mean(dataframe['fBm']) + 2 * np.std(dataframe['fBm'])
        dataframe['frac_sma_dc'] = dataframe['frac_diff'].ewm(span=int(cycle_period)).mean()
        dataframe['signal_UP_dc'] = np.where(dataframe['frac_sma_dc'] > 0, dataframe['frac_sma_dc'], np.nan)
        dataframe['signal_DN_dc'] = np.where(dataframe['frac_sma_dc'] < 0, dataframe['frac_sma_dc'], np.nan)
        dataframe['signal_UP_dc'] = dataframe['signal_UP_dc'].ffill()
        dataframe['signal_DN_dc'] = dataframe['signal_DN_dc'].ffill()
        if self.dp.runmode.value in ('live', 'dry_run'):
            dataframe['signal_MEAN_UP_dc'] = dataframe['signal_UP_dc'].rolling(int(cycle_period)).mean() * self.dc_x.value
            dataframe['signal_MEAN_DN_dc'] = dataframe['signal_DN_dc'].rolling(int(cycle_period)).mean() * self.dc_x.value
        else:
            # this step is for hyperopting and backtesting to simulate the rolling window of the exchange in live mode.
            # this needs to be a value specific for YOUR exchange and/or timeframe.
            dataframe['signal_MEAN_UP_dc'] = dataframe['signal_UP_dc'].rolling(1700).mean() * self.dc_x.value
            dataframe['signal_MEAN_DN_dc'] = dataframe['signal_DN_dc'].rolling(1700).mean() * self.dc_x.value

        dataframe['frac_sma_0'] = dataframe['frac_diff'].ewm(span=int(harmonics[0])).mean()
        dataframe['frac_sma_1'] = dataframe['frac_diff'].ewm(span=int(harmonics[1])).mean()
        dataframe['frac_sma_2'] = dataframe['frac_diff'].ewm(span=int(harmonics[2])).mean()

        # Apply rolling window operation to the 'OHLC4' column
        rolling_windowc = dataframe['ha_close'].rolling(cycle_period) 
        rolling_windowh0 = dataframe['ha_close'].rolling(int(harmonics[0]))
        rolling_windowh1 = dataframe['ha_close'].rolling(int(harmonics[1])) 
        rolling_windowh2 = dataframe['ha_close'].rolling(int(harmonics[2])) 

        # Calculate the peak-to-peak value on the resulting rolling window data
        ptp_valuec = rolling_windowc.apply(lambda x: np.ptp(x))
        ptp_valueh0 = rolling_windowh0.apply(lambda x: np.ptp(x))
        ptp_valueh1 = rolling_windowh1.apply(lambda x: np.ptp(x))
        ptp_valueh2 = rolling_windowh2.apply(lambda x: np.ptp(x))

        # Assign the calculated peak-to-peak value to the DataFrame column
        dataframe['cycle_move'] = ptp_valuec / dataframe['ha_close']
        dataframe['h0_move'] = ptp_valueh0 / dataframe['ha_close']
        dataframe['h1_move'] = ptp_valueh1 / dataframe['ha_close']
        dataframe['h2_move'] = ptp_valueh2 / dataframe['ha_close']

        if self.dp.runmode.value in ('live', 'dry_run'):
            dataframe['cycle_move_mean'] = dataframe['cycle_move'].mean()        
            dataframe['h0_move_mean'] = dataframe['h0_move'].mean()
            dataframe['h1_move_mean'] = dataframe['h1_move'].mean() 
            dataframe['h2_move_mean'] = dataframe['h2_move'].mean()
            dataframe['move_mean_bottom'] = dataframe['h2_move_mean']/2
        else:
            # this step is for hyperopting and backtesting to simulate the rolling window of the exchange in live mode.
            # this needs to be a value specific for YOUR exchange and/or timeframe.
            dataframe['cycle_move_mean'] = dataframe['cycle_move'].rolling(1700).mean()        
            dataframe['h0_move_mean'] = dataframe['h0_move'].rolling(1700).mean()
            dataframe['h1_move_mean'] = dataframe['h1_move'].rolling(1700).mean() 
            dataframe['h2_move_mean'] = dataframe['h2_move'].rolling(1700).mean()
            dataframe['move_mean_bottom'] = dataframe['h2_move_mean']/2

        # Add envelopes for the dominant cycle
        dataframe['upper_envelope'] = dataframe['dc_EWM'] * (1 + dataframe['cycle_move_mean'])
        dataframe['lower_envelope'] = dataframe['dc_EWM'] * (1 - dataframe['cycle_move_mean'])
        dataframe['upper_envelope_h0'] = dataframe['dc_EWM'] * (1 + dataframe['h0_move_mean'])
        dataframe['lower_envelope_h0'] = dataframe['dc_EWM'] * (1 - dataframe['h0_move_mean'])
        dataframe['upper_envelope_h1'] = dataframe['dc_EWM'] * (1 + dataframe['h1_move_mean'])
        dataframe['lower_envelope_h1'] = dataframe['dc_EWM'] * (1 - dataframe['h1_move_mean'])
        dataframe['upper_envelope_h2'] = dataframe['dc_EWM'] * (1 + dataframe['h2_move_mean'])
        dataframe['lower_envelope_h2'] = dataframe['dc_EWM'] * (1 - dataframe['h2_move_mean'])

        # Base pair dataframe timeframe indicators
        curr_pair = metadata["pair"]
        window_size = int(cycle_period)
        
        # set global variables.
        self.curr_dataframe = dataframe
        self.curr_pair = curr_pair
        self.lookahead = int(harmonics[2])
        self.dc = int(cycle_period)
        self.har = harmonics
        self.wavelet_size = int(cycle_period) # needed for consistently-sized transforms
        self.win_size = self.wavelet_size  # this can vary (64)

        self.train_min_len = self.wavelet_size  # longer = slower (64)
        self.train_len = self.wavelet_size * 4  # longer = slower 256 (64*4)
        self.scale_len = int(harmonics[0])  # no. recent candles to use when scaling (64/2)
        self.win_size = self.wavelet_size  # (64)
        self.model_window = self.wavelet_size  # longer = slower (64)

        self.update_pairlist_data()

        # backward looking gain
        dataframe["gain"] = (
            100.0
            * (dataframe["ha_close"] - dataframe["ha_close"].shift(self.lookahead))
            / dataframe["ha_close"].shift(self.lookahead)
        )
        dataframe["gain"].fillna(0.0, inplace=True)
        dataframe["gain"] = dataframe["gain"].round(4)

        # need to save the gain data for later scaling
        self.gain_data = dataframe["gain"].to_numpy().copy()

        # target profit/loss thresholds
        dataframe["profit"] = dataframe["gain"].clip(lower=0.0)
        dataframe["loss"] = dataframe["gain"].clip(upper=0.0)

        dataframe = self.update_gain_targets(dataframe)

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

        # RSI
        dataframe["rsi"] = ta.RSI(dataframe, timeperiod=window_size)

        # init prediction column
        dataframe["predicted_gain"] = 0.0

        # RMI: https://www.tradingview.com/script/kwIt9OgQ-Relative-Momentum-Index/
        dataframe["rmi"] = cta.RMI(dataframe, length=window_size, mom=5)

        # scaled version for use as guard metric
        dataframe["srmi"] = 2.0 * (dataframe["rmi"] - 50.0) / 100.0

        # guard metric must be in range [-1,+1], with -ve values indicating oversold and +ve values overbought
        dataframe["guard_metric"] = dataframe["srmi"]

        # create and init the model, if first time (dataframe has to be populated first)
        if self.model is None:
            # print("    Loading model")
            self.load_model(np.shape(dataframe))

        # add the predictions
        # print("    Making predictions...")
        dataframe = self.add_predictions(dataframe)

        dataframe.fillna(0.0, inplace=True)
        gain = dataframe['gain'].iloc[-1]
        profit = dataframe['profit'].iloc[-1]
        loss = dataframe['loss'].iloc[-1]

        logger.info(f'{pair} - Gain: {gain} | Profit: {profit:.2f} | Loss: {loss:.2f}')        
        logger.info(f'{pair} - DC: {cycle_period:.2f} | 1/2: {harmonics[0]:.2f} | 1/3: {harmonics[1]:.2f} | 1/4: {harmonics[2]:.2f}')
        end_time = time.time()
        logger.info(f"Indicators done for {pair} in {end_time - start_time:.2f} secs")

        return dataframe

    def update_gain_targets(self, dataframe):
        # win_size = max(self.lookahead, 6)
        win_size = self.scale_len
        self.profit_nstd = float(self.cexit_profit_nstd.value)
        self.loss_nstd = float(self.cexit_loss_nstd.value)

        dataframe["target_profit"] = (
            dataframe["profit"].rolling(window=win_size).mean()
            + self.profit_nstd * dataframe["profit"].rolling(window=win_size).std()
        )

        dataframe["target_loss"] = dataframe["loss"].rolling(window=win_size).mean() - self.loss_nstd * abs(
            dataframe["loss"].rolling(window=win_size).std()
        )

        dataframe["target_profit"] = dataframe["target_profit"].clip(lower=float(self.cexit_min_profit_th.value))
        dataframe["target_loss"] = dataframe["target_loss"].clip(upper=float(self.cexit_min_loss_th.value))

        dataframe["target_profit"] = np.nan_to_num(dataframe["target_profit"])
        dataframe["target_loss"] = np.nan_to_num(dataframe["target_loss"])

        dataframe["local_mean"] = dataframe["close"].rolling(window=win_size).mean()
        dataframe["local_min"] = dataframe["close"].rolling(window=win_size).min()
        dataframe["local_max"] = dataframe["close"].rolling(window=win_size).max()

        return dataframe

    def smooth(self, y, window):
        box = np.ones(window) / window
        y_smooth = np.convolve(y, box, mode="same")
        # Hack: constrain to 3 decimal places (should be elsewhere, but convenient here)
        y_smooth = np.round(y_smooth, decimals=3)
        return np.nan_to_num(y_smooth)

    # -----------------------

    # look ahead to get future gain. Do *not* put this into the main dataframe!
    def get_future_gain(self, dataframe):
        df = self.convert_dataframe(dataframe)
        future_gain = df["gain"].shift(-self.lookahead).to_numpy()
        future_gain[-self.lookahead :] = 0.0
        future_gain = np.round(future_gain, decimals=3)
        future_gain = np.nan_to_num(future_gain)

        # future_gain = dataframe['gain'].shift(-self.lookahead).to_numpy()
        # return self.smooth(future_gain, 8)
        return future_gain

    # -------------
    # Normalisation

    array_scaler = RobustScaler()

    def update_scaler(self, data):
        if not self.array_scaler:
            self.array_scaler = RobustScaler()

        self.array_scaler.fit(data.reshape(-1, 1))

    def norm_array(self, a):
        return self.array_scaler.transform(a.reshape(-1, 1))

    def denorm_array(self, a):
        return self.array_scaler.inverse_transform(a.reshape(-1, 1)).squeeze()

    # scales array data, based on array target
    def scale_array(self, target, data):
        # detrend the input arrays
        t = np.arange(0, len(target))
        t_poly = np.polyfit(t, target, 1)
        t_line = np.polyval(t_poly, target)
        x = target - t_line

        t = np.arange(0, len(data))
        d_poly = np.polyfit(t, data, 1)
        d_line = np.polyval(d_poly, data)
        y = data - d_line

        # scale untrended data
        self.update_scaler(x)
        y_scaled = self.denorm_array(y)

        # retrend
        y_scaled = y_scaled + d_line

        return y_scaled


    def convert_dataframe(self, dataframe: DataFrame) -> DataFrame:
        df = dataframe.copy()

        # convert date column so that it can be scaled.
        if "date" in df.columns:
            dates = pd.to_datetime(df["date"], utc=True)
            df["date"] = dates.astype("int64")

        df.fillna(0.0, inplace=True)

        df.set_index("date")
        df.reindex()

        # print(f'    norm_data:{self.norm_data}')
        if self.norm_data:
            # scale the dataframe
            self.scaler.fit(df)
            df = pd.DataFrame(self.scaler.transform(df), columns=df.columns)

        return df

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

    # Model-related funcs. Override in subclass to use a different type of model

    def get_model_path(self, pair):
        category = self.__class__.__name__
        root_dir = group_dir + "/models/" + category
        model_name = category
        if self.model_per_pair and (len(pair) > 0):
            model_name = model_name + "_" + pair.split("/")[0]
        path = root_dir + "/" + model_name + ".sav"
        return path

    def load_model(self, df_shape):
        model_path = self.get_model_path("")

        # load from file or create new model
        if os.path.exists(model_path):
            # use joblib to reload model state
            print("    loading from: ", model_path)
            self.model = joblib.load(model_path)
            self.model_trained = True
            self.new_model = False
            self.training_mode = False
            # set the model in the forecaster
            self.forecaster.set_model(self.model)
        else:
            self.model = self.forecaster.get_model()
            self.model_trained = False
            self.new_model = True
            self.training_mode = True

        # sklearn family of regressors sometimes support starting with an existing model (warm_start),
        # or incremental training (partial_fit())
        if hasattr(self.model, "warm_start"):
            self.model.warm_start = True
            self.supports_incremental_training = True  # override default

        if hasattr(self.model, "partial_fit"):
            self.supports_incremental_training = True  # override default

        # if self.model is None:
        #     print("***    ERR: model was not created properly ***")

        return

    # -------------

    def save_model(self):
        # save trained model

        model_path = self.get_model_path("")

        # create directory if it doesn't already exist
        save_dir = os.path.dirname(model_path)
        if not os.path.exists(save_dir):
            os.makedirs(save_dir)

        # extract underlying model from forecaster
        model = self.forecaster.get_model()

        # use joblib to save model state
        if model_path is not None:
            logger.info(f"Saving to:  {model_path}")
        joblib.dump(model, model_path)

        return

    # -------------

    # train the model. Override if not an sklearn-compatible algorithm
    # set save_model=False if you don't want to save the model (needed for ML algorithms)
    def train_model(self, forecaster: Forecasters.base_forecaster, data: np.array, results: np.array, save_model):
        if forecaster is None:
            logger.info("***ERR: no forecaster ***")
            return

        x = np.nan_to_num(data)
        y = np.nan_to_num(results)

        forecaster.train(x, y, incremental=True)

        # print(f'   train_model() data:{np.shape(data)} results:{np.shape(results)}')

        return

    # -------------

    # initial training of the model
    def init_model(self, dataframe: DataFrame):
        # if model is not yet trained, or this is a new model and we want to combine across pairs, then train
        if (not self.model_trained) or (self.new_model and self.combine_models):
            df = dataframe

            future_gain_data = self.get_future_gain(df)
            data = self.get_data(df)

            if self.single_col_prediction:
                training_data = dataframe["gain"].to_numpy()
                # training_data = self.smooth(training_data, 2)
                training_data = training_data.reshape(-1, 1)
            else:
                training_data = data.copy()
            training_data = training_data[: -self.lookahead - 1]
            training_labels = future_gain_data[: -self.lookahead - 1].copy()

            if not self.model_trained:
                logger.info(f"Initial training ({self.curr_pair})")
            else:
                logger.info(f"Incremental training ({self.curr_pair})")

            if self.forecaster.supports_retrain:
                # loop through data and train on self.wavelet_length amounts of data
                start = 0
                end = self.train_len - 1
                num_buffs = int((np.shape(training_data)[0]) / self.train_len)
                for i in tqdm(range(num_buffs), desc="Training Progress"):
                    logger.info(f'Training start:{start} end:{end} self.train_len:{self.train_len} Epoch {i}:{num_buffs}')
                    self.train_model(self.forecaster, training_data[start:end], training_labels[start:end], True)

                    start = start + self.train_len
                    end = end + self.train_len

            else:
                self.train_model(self.forecaster, training_data, training_labels, True)

            self.model_trained = True

            if self.new_model:
                self.save_model()

        # print(f'    model_trained:{self.model_trained} new_model:{self.new_model}  combine_models:{self.combine_models}')

        return

    # -------------

    # set the data for this straegy. Override if necessary
    def get_data(self, dataframe):
        # default is to just normalise the dataframe and convert to numpy array
        self.curr_dataframe = dataframe
        df = dataframe.copy()
        gain = df["gain"].to_numpy()
        # gain = self.smooth(gain, 2)
        df["gain"] = gain
        self.data = np.array(self.convert_dataframe(df))
        return self.data

    # -------------

    # generate predictions for an np array (intended to be overriden if needed)
    def predict_data(self, forecaster: Forecasters.base_forecaster, data):
        x = np.nan_to_num(data)

        preds = forecaster.forecast(x, self.lookahead)

        # print(f'    data:{np.shape(data)} preds:{np.shape(preds)}')

        # # smooth predictions to try and avoid drastic changes
        # preds = self.smooth(preds, 2)

        # scale the results to generally match the input characteristics
        if self.scale_results:
            preds = self.scale_array(data[-8:], preds)

        preds = np.clip(preds, -3.0, 3.0)
        return preds

    # -------------

    # single prediction (for use in rolling calculation)
    def predict(self, gain, dataframe) -> float:
        # Get the start and end index labels of the series
        start = gain.index[0]
        end = gain.index[-1]

        # Get the integer positions of the labels in the dataframe index
        start_row = dataframe.index.get_loc(start)
        end_row = dataframe.index.get_loc(end) + 1  # need to add the 1, don't know why!

        # if end_row < (self.wavelet_size + self.lookahead):
        if start_row < (self.wavelet_size + self.lookahead):  # need buffer for training
            return 0.0

        # train on previous data
        train_end = start_row - self.lookahead - 1
        train_start = max(0, train_end - self.train_len)
        scale_start = max(0, end - self.scale_len)

        if (not self.training_mode) and (self.supports_incremental_training):
            train_data = self.training_data[train_start : start - 1].copy()
            train_results = self.training_labels[train_start : start - 1].copy()
            # pair_forecaster = copy.deepcopy(self.forecaster)  # reset to avoid over-training
            pair_forecaster = self.forecaster
            self.train_model(pair_forecaster, train_data, train_results, False)
        else:
            pair_forecaster = self.forecaster

        # predict for current window
        dslice = self.training_data[start:end].copy()
        self.gain_data = np.array(dataframe["gain"].iloc[scale_start:end])  # needed for scaling
        y_pred = self.predict_data(pair_forecaster, dslice)

        return y_pred[-1]

    # -------------

    # alternate rolling prediction approach. The pandas rolling mechanism seems to have issues for some reason
    def rolling_predict(self, gain, window_size):
        win_size = window_size

        x = np.nan_to_num(np.array(gain))
        preds = np.zeros(len(x), dtype=float)
        nrows = np.shape(self.training_data)[0]

        start = 0
        end = start + win_size
        scale_start = max(0, end - self.scale_len)

        # train_end = max(0, start  - self.lookahead - 1)
        # train_end = max(0, start  - 1)
        # train_end = max(0, start - self.lookahead - 1)
        train_end = min(end - self.lookahead - 1, nrows - self.lookahead - 2)  # potential lookahead problem
        train_start = max(0, train_end - self.train_len)

        # get the forecaster for this pair
        if self.custom_trade_info[self.curr_pair]["forecaster"] is None:
            # make a deep copy so that we don't override the baseline model
            pair_forecaster = copy.deepcopy(self.forecaster)
            self.custom_trade_info[self.curr_pair]["forecaster"] = pair_forecaster
        else:
            pair_forecaster = self.custom_trade_info[self.curr_pair]["forecaster"]

        # loop through each row
        while end <= len(x):
            if start < (self.wavelet_size + self.lookahead):  # need buffer for training
                preds[end - 1] = 0.0
            else:
                # (re-)train the model on prior data and get predictions

                if (not self.training_mode) and (self.supports_incremental_training):
                    train_data = self.training_data[train_start:train_end].copy()
                    train_results = self.training_labels[train_start:train_end].copy()
                    # pair_forecaster = copy.deepcopy(self.forecaster)  # reset to avoid over-training
                    self.train_model(pair_forecaster, train_data, train_results, False)
                    logger.info(f'start:{start} end:{end} train_start:{train_start} train_end:{train_end}')

                # rebuild data up to end of current window
                dslice = self.training_data[start:end].copy()
                self.gain_data = x[scale_start:end]  # needed for scaling
                forecast = self.predict_data(pair_forecaster, dslice)

                logger.info(f'Forecast:{forecast}')
                preds[end - 1] = forecast[-1]

            # move the window to the next segment
            end = end + 1
            start = start + 1
            # train_end = start - self.lookahead - 1
            # train_end = start - 1
            train_end = min(end - self.lookahead - 1, nrows - self.lookahead - 2)  # potential lookahead problem
            train_start = max(0, train_end - self.train_len)

        # save the updated/trained forecaster
        self.custom_trade_info[self.curr_pair]["forecaster"] = pair_forecaster

        return preds

    # ----------
    # add predictions in a jumping fashion. This is a compromise - the rolling version is very slow
    # Note: you probably need to manually tune the parameters, since there is some limited lookahead here
    def add_jumping_predictions(self, dataframe: DataFrame) -> DataFrame:
        df = dataframe

        # roll through the close data and predict for each step
        nrows = np.shape(df)[0]

        # set up training data
        future_gain_data = self.get_future_gain(df)
        data = self.get_data(dataframe)

        self.training_data = data.copy()
        self.training_labels = np.zeros(np.shape(future_gain_data), dtype=float)
        self.training_labels = future_gain_data.copy()

        # initialise the prediction array, using the close data
        pred_array = np.zeros(np.shape(future_gain_data), dtype=float)

        win_size = self.model_window

        # loop until we get to/past the end of the buffer
        # start = win_size
        start = self.lookahead + self.train_len
        end = start + win_size - 1
        # train_end = max(0, start - self.lookahead - 1)
        # train_end = max(0, end - self.lookahead - 1)
        train_size = self.train_len
        # train_start = max(0, train_end - train_size)
        scale_start = max(0, end - self.scale_len)

        train_end = min(end - self.lookahead - 1, nrows - self.lookahead - 2)  # potential lookahead problem
        train_start = max(0, train_end - self.train_len)

        # get the forecaster for this pair
        if self.custom_trade_info[self.curr_pair]["forecaster"] is None:
            # make a deep copy so that we don't override the baseline model
            pair_forecaster = copy.deepcopy(self.forecaster)
        else:
            pair_forecaster = self.custom_trade_info[self.curr_pair]["forecaster"]

        # loop through the rows
        while end < nrows:
            # extract the data and coefficients from the current window

            # (re-)train the model on prior data and get predictions

            if (not self.training_mode) and (self.supports_incremental_training):
                train_data = self.training_data[train_start:train_end].copy()
                train_results = self.training_labels[train_start:train_end].copy()
                pair_forecaster = copy.deepcopy(self.forecaster)  # reset to avoid over-training
                self.train_model(pair_forecaster, train_data, train_results, False)
                # logger.info(f'train_data: {np.shape(train_data)}')
                # logger.info(f'train_results: {np.shape(train_results)}')

            # rebuild data up to end of current window
            dslice = self.training_data[start:end].copy()
            self.gain_data = np.array(dataframe["gain"].iloc[scale_start:end])  # needed for scaling
            preds = self.predict_data(pair_forecaster, dslice)

            # print(f'dslice: {np.shape(dslice)}')
            # print(f'preds: {np.shape(preds)}')

            # copy the predictions for this window into the main predictions array
            pred_array[start:end] = preds.copy()

            # move the window to the next segment
            end = end + win_size
            start = start + win_size
            train_end = end - self.lookahead - 1
            train_start = max(0, train_end - train_size)

        # make sure the last section gets processed (the loop above may not exactly fit the data)
        # Note that we cannot use the last section for training because we don't have forward looking data

        # predict for last window
        dslice = self.training_data[-win_size:]
        # preds = self.forecaster.predict(dslice)
        slen = win_size
        self.gain_data = np.array(dataframe["gain"].iloc[-slen:])  # needed for scaling
        preds = self.predict_data(pair_forecaster, dslice)
        pred_array[-len(preds) :] = preds.copy()

        dataframe["predicted_gain"] = pred_array.copy()

        # save the updated/trained forecaster
        self.custom_trade_info[self.curr_pair]["forecaster"] = pair_forecaster

        return dataframe

    # -------------

    def add_rolling_predictions(self, dataframe: DataFrame) -> DataFrame:
        try:
            # set up training data
            future_gain_data = self.get_future_gain(dataframe)
            data = self.get_data(dataframe)

            if self.single_col_prediction:
                self.training_data = dataframe["gain"].to_numpy()
                # self.training_data = self.smooth(self.training_data, 1)
                self.training_data = self.training_data.reshape(-1, 1)
            else:
                self.training_data = data.copy()

            self.training_labels = np.zeros(np.shape(future_gain_data), dtype=float)
            self.training_labels = future_gain_data.copy()

            # dataframe['predicted_gain'] = dataframe['gain'].rolling(window=self.model_window).apply(self.predict, args=(dataframe,))
            dataframe["predicted_gain"] = self.rolling_predict(dataframe["gain"], self.model_window)

            # dataframe['predicted_gain'] = self.smooth(dataframe['predicted_gain'], 2)

        except Exception as e:
            print("*** Exception in add_rolling_predictions()")
            print(e)  # prints the error message
            print(traceback.format_exc())  # prints the full traceback

        return dataframe

    # -------------

    # add the latest prediction, and update training periodically
    def add_latest_prediction(self, dataframe: DataFrame) -> DataFrame:
        df = dataframe

        try:
            # set up training data
            # TODO: see if we can do this incrementally instead of rebuilding every time, or just use portion of data
            future_gain_data = self.get_future_gain(df)
            data = self.get_data(dataframe)

            plen = len(self.custom_trade_info[self.curr_pair]["predictions"])
            dlen = len(dataframe["gain"])
            clen = min(plen, dlen)

            # self.training_data = data[-clen:].copy()
            # self.training_labels = future_gain_data[-clen:].copy()
            self.training_data = data
            self.training_labels = future_gain_data

            pred_array = np.zeros(clen, dtype=float)

            logger.info(f"[predictions]:{np.shape(self.custom_trade_info[self.curr_pair]['predictions'])}  pred_array:{np.shape(pred_array)}")

            # copy previous predictions and shift down by 1
            pred_array[-clen:] = self.custom_trade_info[self.curr_pair]["predictions"][-clen:].copy()
            pred_array = np.roll(pred_array, -1)
            pred_array[-1] = 0.0

            # train on previous data
            # train_end = clen - self.model_window - self.lookahead
            train_end = np.shape(self.training_data)[0] - self.lookahead - 2
            train_start = max(0, train_end - self.train_len)

            # cannot use last portion because we are looking ahead
            tslice = self.training_data[train_start:train_end]
            lslice = self.training_labels[train_start:train_end]

            # get the forecaster for this pair
            if self.custom_trade_info[self.curr_pair]["forecaster"] is None:
                # make a deep copy so that we don't override the baseline model
                pair_forecaster = copy.deepcopy(self.forecaster)
                # forecaster should already be there, so print warning
                print(f"*** WARNING: No pre-existing forecaster. Creating from model")
            else:
                pair_forecaster = self.custom_trade_info[self.curr_pair]["forecaster"]

            # update forecaster and get predictions

            self.train_model(pair_forecaster, tslice, lslice, False)

            slen = min(clen, self.scale_len)
            self.gain_data = np.array(dataframe["gain"].iloc[-slen:])  # needed for scaling
            preds = self.predict_data(pair_forecaster, self.training_data[-self.model_window :])

            # self.forecaster = copy.deepcopy(base_forecaster) # restore original model

            # only replace last prediction (i.e. don't overwrite the historical predictions)
            pred_array[-1] = preds[-1]

            dataframe["predicted_gain"] = 0.0
            dataframe["predicted_gain"][-clen:] = pred_array[-clen:].copy()
            self.custom_trade_info[self.curr_pair]["predictions"][-clen:] = pred_array[-clen:].copy()

            # save the updated/trained forecaster
            self.custom_trade_info[self.curr_pair]["forecaster"] = pair_forecaster

            """"""
            # Debug: logger.info info if in buy or sell region (nothing otherwise)
            pg = preds[-1]
            if pg <= dataframe["target_loss"].iloc[-1]:
                logger.info(f"Downwards predict {pg:6.2f}% loss for:   {self.curr_pair}")
            elif pg >= dataframe["target_profit"].iloc[-1]:
                logger.info(f"Upwards predict {pg:6.2f}% profit for: {self.curr_pair}")

            """"""

        except Exception as e:
            print("*** Exception in add_latest_prediction()")
            print(e)  # prints the error message
            print(traceback.format_exc())  # prints the full traceback

        return dataframe

    # -------------

    # add predictions to dataframe['predicted_gain']
    def add_predictions(self, dataframe: DataFrame) -> DataFrame:
        # print(f"    {self.curr_pair} adding predictions")

        run_profiler = False

        if run_profiler:
            prof = cProfile.Profile()
            prof.enable()

        self.scaler = RobustScaler()  # reset scaler each time

        self.init_model(dataframe)

        if self.curr_pair not in self.custom_trade_info:
            self.custom_trade_info[self.curr_pair] = {
                "forecaster": None,
                "initialised": False,
                "predictions": None,
                "curr_prediction": 0.0,
                "curr_target": 0.0,
            }

        if self.training_mode:
            logger.info(f"Training mode. Skipping backtest for {self.curr_pair}")
            dataframe["predicted_gain"] = 0.0
        else:
            """
            if not self.custom_trade_info[self.curr_pair]["initialised"]:
                logger.info(f"    backtesting {self.curr_pair}")
                if self.use_rolling:
                    dataframe = self.add_rolling_predictions(dataframe)
                else:
                    dataframe = self.add_jumping_predictions(dataframe)

                self.custom_trade_info[self.curr_pair]["initialised"] = True
                self.custom_trade_info[self.curr_pair]["predictions"] = dataframe["predicted_gain"].copy()
            else:
                # logger.info(f'    updating latest prediction for: {self.curr_pair}')
                dataframe = self.add_latest_prediction(dataframe)

                # save latest prediction and threshold for later use (where dataframe is not available)
                self.custom_trade_info[self.curr_pair]["curr_prediction"] = dataframe["predicted_gain"].iloc[-1]
                self.custom_trade_info[self.curr_pair]["curr_target"] = dataframe["target_profit"].iloc[-1]

            """

            logger.info(f"Backtesting {self.curr_pair}")
            if self.use_rolling:
                dataframe = self.add_rolling_predictions(dataframe)
            else:
                dataframe = self.add_jumping_predictions(dataframe)

            # predictions can spike, so constrain range
            dataframe["predicted_gain"] = dataframe["predicted_gain"].clip(lower=-3.0, upper=3.0)

            # save target rate for later use
            dataframe["curr_target"] = dataframe["close"] * (1.0 + dataframe["predicted_gain"] / 100.0)
            # TODO: really should set target to value predicted at previous buy signal

            # save latest prediction and threshold for later use (where dataframe is not available)
            curr_prediction = dataframe["predicted_gain"].iloc[-1]
            curr_target = dataframe["close"].iloc[-1] * (1.0 + curr_prediction / 100.0)
            self.custom_trade_info[self.curr_pair]["curr_prediction"] = curr_prediction
            self.custom_trade_info[self.curr_pair]["curr_target"] = curr_target

        # add shifted version, for debug only
        dataframe["shifted_pred"] = dataframe["predicted_gain"].shift(self.lookahead)

        if run_profiler:
            prof.disable()
            # print profiling output
            stats = pstats.Stats(prof).strip_dirs().sort_stats("cumtime")
            stats.print_stats(20)  # top 20 rows

        return dataframe

    ###################################
    
    def populate_entry_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
        conditions_long = []
        conditions_short = []
        guard_conditions_long = []
        guard_conditions_short = []

        dataframe.loc[:, "enter_tag"] = ""
        dataframe["enter_long"] = 0
        dataframe["buy_region"] = 0
        dataframe["enter_short"] = 0
        dataframe["sell_region"] = 0

        if self.training_mode:
            return dataframe

        # Atualizar os alvos de ganho aqui para que possamos usar parâmetros do hyperopt
        dataframe = self.update_gain_targets(dataframe)

        # Verificar volume de negociação
        conditions_long.append(dataframe["volume"] > 1.0)
        conditions_short.append(dataframe["volume"] > 1.0)

        # Métrica de proteção (guard metric)
        if self.enable_guard_metric.value:
            # Guard metric em região de sobrevenda
            guard_conditions_long.append(dataframe["guard_metric"] < self.entry_guard_metric.value)
            guard_conditions_short.append(dataframe["guard_metric"] > self.entry_guard_metric.value)

        # Verificação de Bollinger Bands
        if self.enable_bb_check.value:
            lower_limit = dataframe["bb_middleband"] - self.entry_bb_factor.value * (dataframe["bb_middleband"] - dataframe["bb_lowerband"])
            upper_limit = dataframe["bb_middleband"] + self.entry_bb_factor.value * (dataframe["bb_upperband"] - dataframe["bb_middleband"])

            dataframe["bullish"] = np.where((dataframe["close"] <= lower_limit), 1, 0)
            dataframe["bearish"] = np.where((dataframe["close"] >= upper_limit), 1, 0)

            guard_conditions_long.append(dataframe["bullish"] > 0)
            guard_conditions_short.append(dataframe["bearish"] > 0)

        if self.enable_squeeze.value:
            if not ("squeeze_long" in dataframe.columns):
                dataframe["squeeze_long"] = np.where((dataframe["bb_width"] >= self.entry_bb_width.value), 1, 0)
            if not ("squeeze_short" in dataframe.columns):
                dataframe["squeeze_short"] = np.where((dataframe["bb_width"] >= self.entry_bb_width.value), 1, 0)

            guard_conditions_long.append(dataframe["squeeze_long"] > 0)
            guard_conditions_short.append(dataframe["squeeze_short"] > 0)

        # Adicionar coluna que combina condições de guard (para plotagem)
        if guard_conditions_long:
            dataframe.loc[reduce(lambda x, y: x & y, guard_conditions_long), "buy_region"] = 1
        if guard_conditions_short:
            dataframe.loc[reduce(lambda x, y: x & y, guard_conditions_short), "sell_region"] = -1

        # Condições para posições longas
        model_cond_long = (
            (dataframe["buy_region"] > 0) &
            (
                qtpylib.crossed_above(dataframe["predicted_gain"], dataframe["target_profit"]) |
                (
                    (dataframe["predicted_gain"] > dataframe["target_profit"]) &
                    (dataframe["predicted_gain"].shift() < dataframe["target_profit"].shift())
                )
            )
        )
        conditions_long.append(model_cond_long)

        # Condições para posições curtas
        model_cond_short = (
            (dataframe["sell_region"] < 0) &
            (
                qtpylib.crossed_below(dataframe["predicted_gain"], dataframe["target_loss"]) |
                (
                    (dataframe["predicted_gain"] < dataframe["target_loss"]) &
                    (dataframe["predicted_gain"].shift() > dataframe["target_loss"].shift())
                )
            )
        )
        conditions_short.append(model_cond_short)

        # Definir tags de entrada
        dataframe.loc[model_cond_long, "enter_tag"] += "model_long "
        dataframe.loc[model_cond_short, "enter_tag"] += "model_short"

        if conditions_long:
            dataframe.loc[reduce(lambda x, y: x & y, conditions_long), "enter_long"] = 1
        if conditions_short:
            dataframe.loc[reduce(lambda x, y: x & y, conditions_short), "enter_short"] = 1

        return dataframe
     

    
    def confirm_trade_entry(
        self,
        pair: str,
        order_type: str,
        amount: float,
        rate: float,
        time_in_force: str,
        current_time: datetime,
        entry_tag: Optional[str],
        side: str,
        **kwargs,
    ) -> bool:
        # this only makes sense in 'live' modes
        if self.dp.runmode.value in ("backtest", "plot", "hyperopt"):
            return True

        # in 'real' systems, there is often a delay between the signal and the trade
        # double-check that predicted gain is still above threshold

        if pair in self.custom_trade_info:
            curr_pred = self.custom_trade_info[pair]["curr_prediction"]

            # check rate against target

            curr_target = self.custom_trade_info[pair]["curr_target"]
            if rate >= curr_target:
                if self.dp.runmode.value not in ("backtest", "plot", "hyperopt"):
                    logger.info("")
                    logger.info(f"    *** {pair} Trade cancelled. Rate ({rate:.2f}) above target ({curr_target:.2f}) ")
                    logger.info("")
                return False

        # just debug
        if self.dp.runmode.value not in ("backtest", "plot", "hyperopt"):
            logger.info("")
            logger.info(
                f"    Trade Entry: {pair}, rate: {rate:.4f} Predicted gain: {curr_pred:.2f}% Target: {curr_target:.2f}"
            )
            logger.info("")

        return True
    
    ###################################

    """
    exit Signal
    """

    def populate_exit_trend(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
        conditions_long = []
        conditions_short = []
        guard_conditions_long = []
        guard_conditions_short = []
        dataframe.loc[:, "exit_tag"] = ""
        dataframe["exit_long"] = 0
        dataframe["sell_region"] = 0
        dataframe["exit_short"] = 0
        dataframe["sell_region"] = 0

        if self.training_mode or (not self.enable_exit_signal.value):
            return dataframe

        dataframe["sell_region"] = 0
    
        # Atualizar os alvos de ganho aqui para que possamos usar parâmetros do hyperopt
        dataframe = self.update_gain_targets(dataframe)

        # Verificar volume de negociação
        conditions_short.append(dataframe["volume"] > 1.0)

        if self.enable_guard_metric.value:
            # Guard metric em região de sobrecompra
            guard_conditions_short.append(dataframe["guard_metric"] > self.entry_guard_metric.value)

        if self.enable_bb_check.value:
            upper_limit = dataframe["bb_middleband"] + self.exit_bb_factor.value * (dataframe["bb_upperband"] - dataframe["bb_middleband"])

            dataframe["bearish"] = np.where((dataframe["close"] >= upper_limit), 1, 0)

            # Região bearish
            guard_conditions_short.append(dataframe["bearish"] < 0)

        if self.enable_squeeze.value:
            if not ("squeeze_short" in dataframe.columns):
                dataframe["squeeze_short"] = np.where((dataframe["bb_width"] >= self.entry_bb_width.value), 1, 0)

            guard_conditions_short.append(dataframe["squeeze_short"] > 0)

        if guard_conditions_short:
            dataframe.loc[reduce(lambda x, y: x & y, guard_conditions_short), "sell_region"] = -1

            # Triggers do modelo
            model_cond_short = (
                (dataframe["sell_region"] < 0) &
                ( 
                    qtpylib.crossed_below(dataframe["predicted_gain"], dataframe["target_loss"]) |
                    (
                        (dataframe["predicted_gain"] < dataframe["target_loss"]) &
                        (dataframe["predicted_gain"].shift() < dataframe["target_loss"].shift())
                    )
                )
            )
        else:
            model_cond_short = qtpylib.crossed_below(dataframe["predicted_gain"], dataframe["target_loss"])

        conditions_short.append(model_cond_short)
        dataframe.loc[model_cond_short, "exit_tag"] += "model_exit_long "

        if conditions_short:
            dataframe.loc[reduce(lambda x, y: x & y, conditions_short), "exit_long"] = 1

        # Condições para saídas de posições longas
        dataframe["buy_region"] = 0

        conditions_long.append(dataframe["volume"] > 1.0)

        if self.enable_guard_metric.value:
            guard_conditions_long.append(dataframe["guard_metric"] < self.entry_guard_metric.value)

        if self.enable_bb_check.value:
            lower_limit = dataframe["bb_middleband"] - self.exit_bb_factor.value * (dataframe["bb_middleband"] - dataframe["bb_lowerband"])

            dataframe["bullish"] = np.where((dataframe["close"] <= lower_limit), 1, 0)

            guard_conditions_long.append(dataframe["bullish"] > 0)

        if self.enable_squeeze.value:
            if not ("squeeze_long" in dataframe.columns):
                dataframe["squeeze_long"] = np.where((dataframe["bb_width"] >= self.entry_bb_width.value), 1, 0)

            guard_conditions_long.append(dataframe["squeeze_long"] > 0)

        if guard_conditions_long:
            dataframe.loc[reduce(lambda x, y: x & y, guard_conditions_long), "buy_region"] = 1

            model_cond_long = (
                (dataframe["buy_region"] > 0) &
                (
                    qtpylib.crossed_above(dataframe["predicted_gain"], dataframe["target_profit"]) |
                    (
                        (dataframe["predicted_gain"] > dataframe["target_profit"]) &
                        (dataframe["predicted_gain"].shift() > dataframe["target_profit"].shift())
                    )
                )
            )
        else:
            model_cond_long = qtpylib.crossed_above(dataframe["predicted_gain"], dataframe["target_profit"])

        conditions_long.append(model_cond_long)
        dataframe.loc[model_cond_long, "exit_tag"] += "model_exit_short"

        if conditions_long:
            dataframe.loc[reduce(lambda x, y: x & y, conditions_long), "exit_long"] = 1

        return dataframe

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

    def confirm_trade_exit(
        self,
        pair: str,
        trade: Trade,
        order_type: str,
        amount: float,
        rate: float,
        time_in_force: str,
        exit_reason: str,
        current_time: datetime,
        **kwargs,
    ) -> bool:
        if self.dp.runmode.value not in ("backtest", "plot", "hyperopt"):
            logger.info("")
            logger.info(f"    Trade Exit: {pair}, rate: {rate:.4f)}")
            logger.info("")

        return True

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

    """
    Custom Stoploss
    """
    """
    def custom_stoploss(
        self,
        pair: str,
        trade: Trade,
        current_time: datetime,
        current_rate: float,
        current_profit: float,
        after_fill: bool,
        **kwargs,
    ) -> float:
        # Assuming trade.side correctly identifies 'long' or 'short' trades via your custom setup.
        trade_side = getattr(trade, 'side', 'long')  # Default to 'long' if 'side' is not set

        # Adjust stoploss based on the side of the trade
        if trade_side == 'long':
            # For long trades, use a trailing stop loss that becomes tighter as profit increases
            if current_profit > 0.05:  # If profit is over 5%
                # Calculate new stop loss as 2.5% below the current rate
                new_stop_loss = -(0.025 * current_rate)
                return max(new_stop_loss, -0.1)  # Ensure stop loss does not exceed 10%
            else:
                return max(self.stoploss, -0.1)  # Use default or -10% as stoploss
        elif trade_side == 'short':
            # For short trades, adjust the stoploss inversely
            if current_profit > 0.05:  # If profit is over 5%
                # Calculate new stop loss as 2.5% above the current rate
                new_stop_loss = 0.025 * current_rate
                return min(new_stop_loss, 0.1)  # Ensure stop loss does not exceed 10%
            else:
                return min(self.stoploss, 0.1)  # Use default or 10% as stoploss

        return self.stoploss

    """
    ###################################

    """
    Custom Exit
    (Note that this runs even if use_custom_stoploss is False)
    """

    # simplified version of custom exit

def custom_exit(self, pair: str, trade: Trade, current_time: "datetime", current_rate: float, current_profit: float, **kwargs):
    dataframe, _ = self.dp.get_analyzed_dataframe(pair=pair, timeframe=self.timeframe)
    last_candle = dataframe.iloc[-1].squeeze()

    if not self.use_custom_stoploss:
        return None

    # Check volume as a basic filter
    if last_candle["volume"] <= 1.0:
        logger.info(f"Exiting {pair} due to insufficient volume: {last_candle['volume']}")
        return None

    # Define the action based on the type of trade: long or short
    if trade.is_short:
        return self.handle_short_exit(pair, trade, current_time, current_rate, current_profit, last_candle)
    else:
        return self.handle_long_exit(pair, trade, current_time, current_rate, current_profit, last_candle)

def handle_short_exit(self, pair, trade, current_time, current_rate, current_profit, last_candle):
    # Handle exit conditions for short trades
    if current_profit > 0.0:
        if last_candle["guard_metric"] <= self.cexit_metric_overbought.value:
            logger.info(f"Covering short for {pair} as overbought guard metric reached: {last_candle['guard_metric']}")
            return "cover_overbought"

        if current_profit > 0.005:
            if last_candle["guard_metric"] <= self.cexit_metric_take_profit.value:
                logger.info(f"Taking profit for short on {pair} as guard metric reached: {last_candle['guard_metric']}")
                return "cover_take_profit"

        if last_candle["predicted_loss"] >= last_candle["target_profit"]:
            logger.info(f"Predicting rise for {pair}, covering short.")
            return "predict_rise"

        if last_candle["exit_short"] > 0:
            logger.info(f"Exit signal triggered for short on {pair}")
            return "exit_signal_short"

    return self.check_common_exit_conditions(trade, current_time, current_profit)

def handle_long_exit(self, pair, trade, current_time, current_rate, current_profit, last_candle):
    # Handle exit conditions for long trades
    if current_profit > 0.0:
        if last_candle["guard_metric"] >= self.cexit_metric_overbought.value:
            logger.info(f"Selling {pair} as overbought guard metric reached: {last_candle['guard_metric']}")
            return "metric_overbought"

        if current_profit > 0.005:
            if last_candle["guard_metric"] >= self.cexit_metric_take_profit.value:
                logger.info(f"Taking profit on {pair} as take profit guard metric reached: {last_candle['guard_metric']}")
                return "take_profit"

        if last_candle["predicted_gain"] <= last_candle["target_loss"]:
            logger.info(f"Predicting drop for {pair}, selling for potential market drop.")
            return "predict_drop"

        if last_candle["exit_long"] > 0:
            logger.info(f"Exit signal triggered for long on {pair}")
            return "exit_signal"

    return self.check_common_exit_conditions(trade, current_time, current_profit)

def check_common_exit_conditions(self, trade, current_time, current_profit):
    time_delta = current_time - trade.open_date_utc
    num_hours = time_delta.total_seconds() / 3600
    num_days = time_delta.days

    # Check common time-based exit conditions
    if num_hours >= 12 and current_profit > 0.001:
        logger.info(f"Unclogging position held over 12 hours with profit.")
        return "unclog_12h"

    if num_days >= 1 and current_profit > 0:
        logger.info(f"Unclogging position held over 1 day with profit.")
        return "unclog_1d"

    if num_days >= 7:
        logger.info(f"Unclogging position held over 7 days, at a loss or profit.")
        return "unclog_7d"

    return None

def perform_fft(price_data, window_size=None):
    if window_size is not None:
        # Apply rolling window to smooth the data
        price_data = price_data.rolling(window=window_size, center=True).mean().dropna()

    normalized_data = (price_data - np.mean(price_data)) / np.std(price_data)
    n = len(normalized_data)
    fft_data = np.fft.fft(normalized_data)
    freq = np.fft.fftfreq(n)
    power = np.abs(fft_data) ** 2
    power[np.isinf(power)] = 0
    return freq, power