virgo-modules 0.0.72__py3-none-any.whl → 0.9.0__py3-none-any.whl

virgo_modules/src/ticketer_source.py

@@ -1,7 +1,7 @@
 import yfinance as yf
 import pandas as pd
 import numpy as np
-import json
+import gc
 
 import matplotlib.pyplot as plt
 import matplotlib.gridspec as gridspec
@@ -36,7 +36,6 @@ from hmmlearn.hmm import GaussianHMM
 
 from plotly.colors import DEFAULT_PLOTLY_COLORS
 
-from sklearn.base import BaseEstimator, TransformerMixin
 from sklearn.pipeline import Pipeline
 from feature_engine.imputation import MeanMedianImputer
 
@@ -48,88 +47,38 @@ from feature_engine.timeseries.forecasting import LagFeatures
 from feature_engine.imputation import MeanMedianImputer
 from feature_engine.discretisation import EqualWidthDiscretiser
 
+from sklearn.linear_model import HuberRegressor
+
 from .aws_utils import upload_file_to_aws
 
 import logging
 
-class InverseHyperbolicSine(BaseEstimator, TransformerMixin):
-    def __init__(self, features, prefix = ''):
-        self.features = features
-        self.prefix = prefix
+from virgo_modules.src.hmm_utils import trainer_hmm
+from virgo_modules.src.transformer_utils import signal_combiner, FeatureSelector
+from virgo_modules.src.transformer_utils import FeaturesEntropy, VirgoWinsorizerFeature # imported bcs some models read this module otherwise it crashed mlflow.load()
 
-    def fit(self, X, y=None):
-        return self
-    
-    def transform(self, X, y=None):
-        for feature in self.features:
-            X[f'{self.prefix}{feature}'] = np.arcsinh(X[feature])
-        return X
-
-class VirgoWinsorizerFeature(BaseEstimator, TransformerMixin):
-    def __init__(self, feature_configs):
-        self.feature_configs = feature_configs
-    def fit(self, X, y=None):
-        return self
-
-    def transform(self, X, y=None):
-        for feature in self.feature_configs:
-            lower = self.feature_configs[feature]['min']
-            upper = self.feature_configs[feature]['max']
-            X[feature] = np.where( lower > X[feature], lower, X[feature])
-            X[feature] = np.where( upper < X[feature], upper, X[feature])
-        return X
-
-class FeatureSelector(BaseEstimator, TransformerMixin):
-    def __init__(self, columns):
-        self.columns = columns
-
-    def fit(self, X, y=None):
-        return self
-
-    def transform(self, X, y=None):
-        return X[self.columns]
-    
-def sharpe_ratio(return_series):
-    N = 255 # Trading days in the year (change to 365 for crypto)
-    rf = 0.005 # Half a percent risk free rare
-    mean = return_series.mean() * N -rf
-    sigma = return_series.std() * np.sqrt(N)
-    sharpe = round(mean / sigma, 3)
-    return sharpe
-
-class signal_combiner(BaseEstimator, TransformerMixin):
-    def __init__(self, columns, drop = True, prefix_up = 'signal_up_', prefix_low = 'signal_low_'):
-        self.columns = columns
-        self.drop = drop
-        self.prefix_up = prefix_up
-        self.prefix_low = prefix_low
-        
-    def fit(self, X, y=None):
-        return self
-
-    def transform(self, X, y=None):
-        for column in self.columns:
-            X['CombSignal_'+column] = np.where(
-                X[self.prefix_up + column] == 1,
-                1, 
-                np.where(
-                    X[self.prefix_low + column] == 1,
-                    1,
-                    0
-                )
-            )
-            if self.drop:
-                X = X.drop(columns = [self.prefix_up + column, self.prefix_low + column])
-        return X
-    
 def data_processing_pipeline(features_base,features_to_drop = False, lag_dict = False, combine_signals = False, discretize_columns = False, correlation = 0.77):
-    
+
+    '''
+    create a scikit learn pipeline object using different configurations and feature engineering blocks with a given flow
+
+            Parameters:
+                    features_to_drop (list): list of features to drop
+                    lag_dict (dict): feature dictionary with configurations to apply lags
+                    combine_signals (list): list of columns/signals to combine
+                    discretize_columns (list): list of features to discretize, bins is fixed
+                    correlation (float): correaltion score threshold for feature selection
+
+            Returns:
+                    pipe (obj): pipeline object
+    '''
+
     lag_pipe_sec = [(f'lags_{key}', LagFeatures(variables = key, periods = lag_dict[key])) for key in lag_dict] if lag_dict else []
     drop_pipe = [('drop_features' , DropFeatures(features_to_drop=features_to_drop))] if features_to_drop else []
     merge = [('signal_combiner', signal_combiner(combine_signals))] if combine_signals else []
     discretize = [('discretize',EqualWidthDiscretiser(discretize_columns, bins = 20 ))] if discretize_columns else []
     drop_corr = [('drop_corr', DropCorrelatedFeatures(threshold=correlation))] if correlation else []
-    
+
     pipe = Pipeline(
         [('selector', FeatureSelector(features_base))] + \
         [('encoding',OneHotEncoder(top_categories=None, variables=['hmm_feature']))]  + \
@@ -142,64 +91,178 @@ def data_processing_pipeline(features_base,features_to_drop = False, lag_dict =
     )
     return pipe
 
-def states_relevance_score(data, default_benchmark_sd = 0.00003, t_threshold = 2):
-    ## legnths
-    cluster_lengths = data.groupby(['hmm_feature','chain_id'],as_index = False).agg(chain_lenght = ('hmm_chain_order','max'))
-    cluster_lengths = cluster_lengths.groupby('hmm_feature').agg(cluster_length_median = ('chain_lenght','median'))
-    ## means
-    def quantile2(x):
-        return x.quantile(0.25)
-    def quantile3(x):
-        return x.quantile(0.75)
-    
-    cluster_returns = data.groupby('hmm_feature').agg(
-        n_uniques = ('chain_id','nunique'),
-        n_obs = ('Date','count'),
-        cluster_ret_q25 = ('chain_return',quantile2),
-        cluster_ret_median = ('chain_return','median'),
-        cluster_ret_q75 = ('chain_return',quantile3),
-    )
-    cluster_returns =  cluster_returns.join(cluster_lengths, how = 'left')
-    cluster_returns['perc_dispute'] = np.where(
-        np.sign(cluster_returns['cluster_ret_q25']) != np.sign(cluster_returns['cluster_ret_q75']),
-        1,0
-    )
-    cluster_returns['iqr'] = cluster_returns.cluster_ret_q75 - cluster_returns.cluster_ret_q25
-    cluster_returns['perc_25'] = abs(cluster_returns.cluster_ret_q25)/cluster_returns['iqr']
-    cluster_returns['perc_75'] = abs(cluster_returns.cluster_ret_q75)/cluster_returns['iqr']
-    cluster_returns['min_perc'] = cluster_returns[['perc_25','perc_75']].min(axis = 1)
-    cluster_returns['min_overlap'] = np.where(cluster_returns['perc_dispute'] == 1,cluster_returns['min_perc'],0)
-    cluster_returns['abs_median'] = abs(cluster_returns['cluster_ret_median'])
-    cluster_returns = cluster_returns.drop(columns = ['perc_25','perc_75','min_perc'])
-    
-    ## relevance or importance
-    # naive aproach
-    cluster_returns['relevance'] =  cluster_returns['abs_median'] + ( 0.5 - cluster_returns['min_overlap'])
-    cluster_returns['t_calc'] = (cluster_returns['cluster_ret_median'] - 0)/(cluster_returns['iqr']/cluster_returns['n_obs'] + default_benchmark_sd/cluster_returns['n_obs'])**(1/2)
-    cluster_returns['abs_t_accpted'] = abs(cluster_returns['t_calc'])
-    cluster_returns['t_accpted'] = abs(cluster_returns['abs_t_accpted']) > t_threshold
-    
-    mean_relevance = cluster_returns['abs_t_accpted'].mean()
-    number_relevant_states = len(cluster_returns[cluster_returns.t_accpted == True])
-
-    return mean_relevance, cluster_returns, number_relevant_states
+class stock_eda_panel(object):
 
+    """
+    Class that initialy gets stock data then apply feature enginering, enrichment, analysis, plotting, model training etc.
+
+    Attributes
+    ----------
+    stock_code : str
+        symbol of the asset
+    n_days : str
+        number of days to extract data
+    data_window : str
+        large window to extract data. Large window is required o extract more data. e.g. '5y', '10y', '15'
+    df : pd.DataFrame
+        Pandas dataframe of the asset data with features
+    strategy_log: pd.DataFrame
+        Pandas dataframe that has the results of different tested strategies (result from strategy simulator hmm)
+    best_strategy: list
+        features of the best performing strategy (result from strategy simulator hmm)
+    top_10_strategy: dict
+        top 10 best performing strategies (result from strategy simulator hmm)
+    settings: dict
+        configuration dictionary of the features and other parameters
+
+    Methods
+    -------
+    augmented_dickey_fuller_statistics(time_series=pd.Series, label=str):
+        Perform dickey fuller or stationary test for a given time series
+        It will print p value of the features
+    get_data():
+        Get asset data performing some data normalization or formating (in case of dates)
+    plot_series_returns(roll_mean_lags1=int, roll_mean_lags2=int)
+        Display plot that time series with mean rolling windows and rolling standard deviations of daily closing prices
+    seasonal_plot():
+        Display time series split by year
+    plot_price_signal(feature=str, feature_2=str, opacity=float):
+        Display botton and roof signals over the closing prices
+    volatility_analysis(lags=int, trad_days=int, window_log_return=int, plot=boolean, save_features=boolean):
+        this method performs log return and volatilyty analysis of the closing prices
+    find_lag(feature=str, lag_list=list, column_target=str,posterior_lag=int, test_size=int):
+        displays correlation curves, using spearman and pearson correlation, of a given feature at different time lags with respecto to a given target
+    outlier_plot(zlim=float, plot=boolean, save_features=boolean):
+        perform outlier analysis of the log returns. It also permors normality test of returns
+    analysis_roll_mean_log_returns(lags=int, plot=boolean):
+        perform analysis of lags of the mean rolling log return
+    compute_clip_bands(feature_name=str,threshold=float):
+        compute outlier detection for a given signal, Note that this follows mean reversion procedure and feature has to be stationary. Also botton and roof resulting signals is attached to the dataframe
+    extract_sec_data(symbol=str, base_columns=list(str), rename_columns=dict):
+        extract new asset data and merge it to the main asset data
+    lag_log_return(lags=int, feature=str, feature_name=str):
+        compute log return given some lags
+    produce_log_volatility(trad_days=int, feature=str, feature_name=str):
+        compute volatility
+    signal_plotter(feature_name=str):
+        display analysis plot of a feature with high and low signals
+    log_features_standard(feature_name=str):
+        save resulting feature names in an standard structure
+    relative_spread_MA(ma1=int, ma2=int, threshold=float, plot=boolean, save_features=boolean):
+        perform relative moving average features, one for short term and another for long/mid term
+    pair_feature(pair_symbol=str, plot=boolean):
+        initialize pair feature data extraction and analysis
+    calculate_cointegration(series_1=pd.series, series_2=pd.series):
+        calculate cointegration score for two time series
+    bidirect_count_feature(rolling_window=int, threshold=float, plot=boolean, save_features=boolean):
+        perform negative and positive return counting in a given rolling time window
+    get_relative_range_feature(window=int, threshold=float, plot=boolean, save_features=boolean):
+        perform relative spread of opening and closing price
+    rsi_feature_improved(window=int, threshold=float, plot=boolean, save_features=boolean):
+        perform relative strength index
+    days_features_bands(window=int, threshold=float, plot=boolean, save_features=boolean):
+        compute mean returns for a given day of the week in a window scope per day
+    analysis_smooth_volume(window=int, threshold=float, plot=boolean, save_features=boolean):
+        compute feature of thrading volumes
+    roc_feature(window=int, threshold=float, plot=boolean, save_features=boolean):
+        perform price rate of change
+    stoch_feature(window=int, smooth1=int, smooth2=int, threshold=float, plot=boolean, save_features=boolean):
+        perform stochastic oscilator RSI feature
+    stochastic_feature(window=int, smooth=int, threshold=float, plot=boolean, save_features=boolean):
+        perform stochastic oscilator feature
+    william_feature(lbp=int, threshold=float, plot=boolean, save_features=boolean):
+        perfom fast stochastic oscilator or william indicator
+    vortex_feature(window=int, threshold=float, plot=boolean, save_features=boolean):
+        perform vortex oscilator
+    expected_return(trad_days:int, feature:str, feature_name:str):
+        perform expected log return based on inversed shift of historical data and applying
+    rolling_feature(feature: str, window:int, function:callable):
+        perform rolling (non expanding) window operation for a given feature
+    time_distance(feature_base:str,feature_window:str, result_feature_name:str, max_window:int):
+        perform distancce time to a given window feature
+    minmax_pricefeature(type_func=str, window=int, distance=bolean, save_features=boolean)
+        get relative price/ distance feature with respect to the min/max price in a given window
+    pair_index_feature(pair_symbol=str, feature_label=str, window=int, threshold=float, plot=boolean, save_features=boolean):
+        perform additional asset ROC feature, then a new feature is created in the main dataframe
+    produce_order_features(feature_name=str, save_features=boolean):
+        perform a feature that captures high and low values in an index. this is usefull to know duration/persistence of a signal
+    compute_last_signal (feature_name=str, save_features=boolean):
+        perform a feature that captures high and low values in an index. this is usefull to know duration/persistence of a signal
+    create_hmm_derived_features():
+        create features derived from hmm states features. Features are the index of the state, the duration of the state, chain raturn
+    cluster_hmm_analysis(n_clusters=int,features_hmm=list, test_data_size=int, seed=int, lag_returns_state=int, plot=boolean, save_features=boolean, model=obj):
+        create or use a hmm model
+    sharpe_ratio(return_series=pd.Series, n_trad_days=int, rf=float):
+        perform sharpe ratio of a given time series return
+    treat_signal_strategy(test_data=pd.DataFrame, strategy=list):
+        helper method that treats signals and converts signals to 1 or 0
+    stategy_simulator(features=list, hmm_feature=boolean):
+        execute strategy and get some performance metrics like sharpe ratio, return
+    viz_strategy(strategy):
+        display analysis plot of a given strategy
+    deep_dive_analysis_hmm(test_data_size=int, split=str):
+        display analysis plot hmm model
+    get_targets(steps=int):
+        produce regression target return taking future prices
+    get_categorical_targets(horizon=int, flor_loss=float, top_gain=float):
+        produce binary target return taking future prices. it produce two targets, one for high returns and another for low returns
+    get_configurations(test_data_size=int, val_data_size=int, model_type=str):
+        produce configuration dictionary that were saved in the feature generation methods if save_features was activated
+    """
 
-class stock_eda_panel(object):
-    
     def __init__(self, stock_code, n_days, data_window = '5y'):
+
+        """
+        Initialize object
+
+        Parameters
+        ----------
+        stock_code (str): symbol of the asset
+        n_days (str): number of days to extract data
+        data_window (str): large window to extract data. Large window is required o extract more data. e.g. '5y', '10y', '15'
+
+        Returns
+        -------
+        None
+        """
+
         self.stock_code = stock_code
         self.n_days = n_days
         self.today = datetime.date.today()
         self.features = list()
         self.signals = list()
         self.data_window = data_window
-        
+
     def augmented_dickey_fuller_statistics(self,time_series, label):
+        """
+        Perform dickey fuller or stationary test for a given time series
+        It will print p value of the features
+
+        Parameters
+        ----------
+        time_series (pd.Series): pandas series of the time series
+        label (pd.Series): feature name
+
+        Returns
+        -------
+        None
+        """
         result = adfuller(time_series.dropna().values)
         print('p-value: {} for the series {}'.format(round(result[1],6), label))
-        
+
     def get_data(self):
+        """
+        Get asset data performing some data normalization or formating (in case of dates)
+
+        Parameters
+        ----------
+        None
+
+        Returns
+        -------
+        None
+        """
+
         begin_date = self.today - relativedelta(days = self.n_days)
         begin_date_str = begin_date.strftime('%Y-%m-%d')
 
@@ -210,7 +273,7 @@ class stock_eda_panel(object):
         df.reset_index(inplace=True)
         df['Date'] = pd.to_datetime(df['Date'], format='mixed',utc=True).dt.date
         df['Date'] = pd.to_datetime(df['Date'])
-        
+
         df = df[df.Date >= begin_date_str ]
         self.settings_general = {
             'n_days':self.n_days,
@@ -219,44 +282,56 @@ class stock_eda_panel(object):
             'execution_date': self.today.strftime('%Y-%m-%d')
         }
         self.df = df
-        
+
         ### cleaning volume
         ### volume clearning
         self.df['Volume'] = np.where(self.df['Volume'] <= 10, np.nan, self.df['Volume'])
         self.df['Volume'] = self.df['Volume'].fillna(method='bfill')
-        
+
         ## filling
-        
+
         base_columns_unit_test = ['Open','High','Low','Close','Volume']
         self.df[base_columns_unit_test] = self.df[base_columns_unit_test].fillna(method='ffill')
-        
+
         ## cleaning nulls
-        
+
         xs = self.df[base_columns_unit_test].isnull().sum()/self.df[base_columns_unit_test].count()
         reject_columns = list(xs[xs > 0.5].index.values)
-        
+
         if len(reject_columns) > 0:
             logging.warning("the following columns have many nulls and are drop: {}".format(reject_columns))
             self.df = self.df.drop(columns = reject_columns)
-        
-        
+
     def plot_series_returns(self,roll_mean_lags1,roll_mean_lags2):
-        
+
+        """
+        Display plot that time series with mean rolling windows and rolling standard deviations of daily closing prices
+
+        Parameters
+        ----------
+        roll_mean_lags1 (int): short term window
+        roll_mean_lags2 (int): mid/long term window
+
+        Returns
+        -------
+        None
+        """
+
         df = self.df
         begin_date = self.today - relativedelta(days = self.n_days)
         begin_date_str = begin_date.strftime('%Y-%m-%d')
-        
+
          ### getting rolling mean
         df["Close_roll_mean"] = (
             df.sort_values("Date")["Close"]
             .transform(lambda x: x.rolling(roll_mean_lags1, min_periods=1).mean())
         )
-        
+
         df["Close_roll_mean_2"] = (
             df.sort_values("Date")["Close"]
             .transform(lambda x: x.rolling(roll_mean_lags2, min_periods=1).mean())
         )
-        
+
         ### getting rolling stdv
         df["Close_roll_std"] = (
             df.sort_values("Date")["Close"]
@@ -273,7 +348,7 @@ class stock_eda_panel(object):
                            ))
 
         fig.add_trace(go.Scatter(x=df['Date'], y=df.Close, marker_color = 'blue', name='Price'),row=1, col=1)
-        
+
         fig.add_trace(go.Scatter(x=df['Date'], y=df.Close_roll_mean, marker_color = 'black', name='roll mean' ),row=1, col=1)
         fig.add_trace(go.Scatter(x=df['Date'], y=df.Close_roll_mean_2, marker_color = 'grey', name='roll mean 2' ),row=1, col=1)
         fig.add_trace(go.Scatter(x=df['Date'], y=df.lower, marker_color = 'pink',legendgroup='bound', name='bound' ),row=1, col=1)
@@ -281,8 +356,21 @@ class stock_eda_panel(object):
 
         fig.update_layout(height=500, width=1200, title_text=f"stock {self.stock_code} vizualization")
         fig.show()
-        
+
     def seasonal_plot(self):
+
+        """
+        Display time series split by year
+
+        Parameters
+        ----------
+        None
+
+        Returns
+        -------
+        None
+        """
+
         df = self.df
         years = list(df['Date'].dt.year.unique())
         years.sort()
@@ -302,10 +390,24 @@ class stock_eda_panel(object):
 
         fig.update_layout(height=500, width=1400, title_text=f"stock {self.stock_code} seasonal vizualization")
         fig.show()
-        
+
     def plot_price_signal(self, feature, feature_2 = '', opacity = 0.3):
-    
-        signal_up_list = [f'signal_up_{feature}', f'signal_up_{feature_2}']  
+
+        """
+        Display botton and roof signals over the closing prices
+
+        Parameters
+        ----------
+        feature (str): name of the main feature to plot
+        feature_2 (str): name of the alternative feature to plot
+        opacity (float): opacity degree of the signals points
+
+        Returns
+        -------
+        None
+        """
+
+        signal_up_list = [f'signal_up_{feature}', f'signal_up_{feature_2}']
         signal_low_list = [f'signal_low_{feature}', f'signal_low_{feature_2}']
         norm_list = [f'norm_{feature}', f'z_{feature}', feature]
 
@@ -315,14 +417,14 @@ class stock_eda_panel(object):
             if norm_feat in self.df.columns:
                 fig.add_trace(go.Scatter(x=self.df['Date'], y=self.df[norm_feat],legendgroup="up", mode='lines',name = norm_feat, marker_color = 'blue'),col = 1, row = 1)
                 break
-        
-        
+
+
         fig.add_trace(go.Scatter(x=self.df['Date'], y=self.df['Close'], mode='lines',name = 'history', marker_color = 'grey'),col = 1, row = 2)
-        
+
         if feature == 'MA_spread':
             fig.add_trace(go.Scatter(x=self.df['Date'], y=self.df[self.ma1_column],legendgroup="ma", mode='lines',name = self.ma1_column, marker_color = 'black'),col = 1, row = 2)
             fig.add_trace(go.Scatter(x=self.df['Date'], y=self.df[self.ma2_column],legendgroup="ma", mode='lines',name = self.ma2_column, marker_color = 'grey'),col = 1, row = 2)
-        
+
         for norm_feat in norm_list:
             if norm_feat in self.df.columns:
                 fig.add_trace(go.Scatter(x=self.df['Date'], y=np.where(self.df[norm_feat] > 0, self.df['Close'], np.nan),legendgroup="up", mode='markers',name = 'up', marker_color = 'green',opacity = opacity),col = 1, row = 2)
@@ -338,8 +440,25 @@ class stock_eda_panel(object):
 
         fig.update_layout(height=900, width=1200)
         fig.show()
-    
+
     def volatility_analysis(self, lags, trad_days, window_log_return, plot = False, save_features = False):
+
+        """
+        this method performs log return and volatilyty analysis of the closing prices
+
+        Parameters
+        ----------
+        lags (int): number of lags to apply to the closing prices
+        trad_days (int): number of trading days to anualize returns or volatility
+        window_log_return (int): window for rolling returns
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+
+        Returns
+        -------
+        None
+        """
+
         df = self.df
         df['log_return'] = np.log(df.Close/df.Close.shift(lags))
         df['sqr_log_return'] = np.square(df.log_return)
@@ -349,13 +468,13 @@ class stock_eda_panel(object):
                 df.sort_values("Date")["log_return"]
                 .transform(lambda x: x.rolling(window_log_return, min_periods=1).mean())
             )
-        
+
         if save_features:
             self.features.append('volatility_log_return')
             self.features.append('roll_mean_log_return')
             self.features.append('log_return')
             self.settings_volatility = {'lags':lags, 'trad_days':trad_days, 'window_log_return':window_log_return}
-            
+
         if plot:
             fig = make_subplots(rows=3, cols=1,vertical_spacing = 0.02,shared_xaxes=True,
                             specs=[
@@ -395,10 +514,25 @@ class stock_eda_panel(object):
 
             self.augmented_dickey_fuller_statistics(df['log_return'], 'log_return')
             self.augmented_dickey_fuller_statistics(df['roll_mean_log_return'], 'roll_mean_log_return')
-            
-            
+
     def find_lag(self, feature, lag_list, column_target = 'log_return',posterior_lag = 4, test_size = 350):
 
+        """
+        displays correlation curves, using spearman and pearson correlation, of a given feature at different time lags with respecto to a given target
+
+        Parameters
+        ----------
+        feature (str): feature name to apply lags
+        lag_list (list): list of lags, each lag as integer
+        column_target (str): target to get correlation, e.g return or mean reaturn
+        posterior_lag (int): for the target, posterior window shift to calculate a window return
+        test_size (int): data size of the test data. The remaining is going to be used as training data. This parameters is ment to avoid overfiting and leackage
+
+        Returns
+        -------
+        None
+        """
+
         results = dict()
         df = self.df.iloc[:-test_size,:][['Date','Close','roll_mean_log_return','log_return',feature]].sort_values('Date').copy()
         for i,lag in enumerate(lag_list):
@@ -413,7 +547,7 @@ class stock_eda_panel(object):
                 'lag':lag,
                 'pearsonr_log_return':r_log[0],
                 'spearman_log_return': sp_log[0],
-            } 
+            }
         del df
         results_df = pd.DataFrame(results).T
 
@@ -426,10 +560,23 @@ class stock_eda_panel(object):
         plt.legend()
         plt.axhline(y=0, color='grey', linestyle='--')
         plt.show()
-    
-    
+
     def outlier_plot(self, zlim, plot = False, save_features = False):
-        
+
+        """
+        perform outlier analysis of the log returns. It also permors normality test of returns
+
+        Parameters
+        ----------
+        zlim (float): alpha or z thrsholds for normalized returns
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+
+        Returns
+        -------
+        None
+        """
+
         mean = self.df.log_return.mean()
         std = self.df.log_return.std()
         self.df['z_log_return'] = (self.df.log_return - mean)/std
@@ -440,7 +587,7 @@ class stock_eda_panel(object):
         self.df['up_outlier'] = zlim*self.df['z_std_log_return'] + mean_
         self.df['low_outlier'] = -zlim*self.df['z_std_log_return'] + mean_
 
-        self.df['signal_low_osutlier'] = np.where( (self.df['z_log_return'] < self.df['low_outlier'] ), 1, 0)
+        self.df['signal_low_outlier'] = np.where( (self.df['z_log_return'] < self.df['low_outlier'] ), 1, 0)
         self.df['signal_up_outlier'] = np.where( (self.df['z_log_return'] > self.df['up_outlier'] ), 1, 0)
         if save_features:
             self.signals.append('signal_low_outlier')
@@ -451,7 +598,7 @@ class stock_eda_panel(object):
             sigma = self.df['z_log_return'].std()
             x = np.linspace(self.df['z_log_return'].min(),self.df['z_log_return'].max(), 15000)
             y = stats.norm.pdf(x, loc = mu, scale = sigma)
-            
+
             fig, axs = plt.subplots(2, 1,figsize=(15,8))
 
             axs[0].hist(self.df['z_log_return'],density = True,bins = 100 , label = 'Returns distribution')
@@ -460,7 +607,7 @@ class stock_eda_panel(object):
             axs[0].axvline(l2, color='green', linestyle='--')
             axs[0].axvline(-l2, color='green', linestyle='--')
             axs[0].plot(x,y, linewidth = 3, color = 'r', label = 'Normal Dist Curve')
-            
+
             axs[1].plot(self.df['Date'],self.df['z_log_return'])
             axs[1].plot(self.df['Date'],self.df['low_outlier'], linestyle='--')
             axs[1].plot(self.df['Date'],self.df['up_outlier'], linestyle='--')
@@ -469,18 +616,31 @@ class stock_eda_panel(object):
             plt.show()
 
             z_stat, p_stat = stats.normaltest(self.df['z_log_return'].dropna())
-            p_stat = round(p_stat, 7) 
+            p_stat = round(p_stat, 7)
             print('---------------------- returns normality tests ----------------------------')
             if p_stat < 0.05:
                 print(f'pvalue: {p_stat} then, returns do not follow a normal distribution')
             else:
                 print(f'pvalue: {p_stat} then, returns follow a normal distribution')
-    
+
     def analysis_roll_mean_log_returns(self, lags, plot = False):
 
+        """
+        perform analysis of lags of the mean rolling log return
+
+        Parameters
+        ----------
+        lags (int): lags to apply to the roll log return
+        plot (boolean): True to display plot
+
+        Returns
+        -------
+        None
+        """
+
         self.df['lag'] = self.df.roll_mean_log_return.shift(lags)
         self.df['Diff'] = self.df['roll_mean_log_return'] - self.df['lag']
-        
+
         if plot:
 
             fig, axs = plt.subplots(1, 3,figsize=(19,4))
@@ -493,7 +653,20 @@ class stock_eda_panel(object):
             plt.show()
 
     def compute_clip_bands(self,feature_name,threshold):
-    
+
+        """
+        compute outlier detection for a given signal, Note that this follows mean reversion procedure and feature has to be stationary. Also botton and roof resulting signals is attached to the dataframe
+
+        Parameters
+        ----------
+        feature_name (str): feature name
+        threshold (float): alpha or z thrsholds for normalized returns
+
+        Returns
+        -------
+        None
+        """
+
         self.df[f'norm_{feature_name}'] =  (self.df[feature_name] - self.df[feature_name].mean())/self.df[feature_name].std()
         mean_ = self.df[f'norm_{feature_name}'].mean()
 
@@ -506,84 +679,140 @@ class stock_eda_panel(object):
         self.df[f'signal_low_{feature_name}'] = np.where( (self.df[f'norm_{feature_name}'] < self.df[f'lower_{feature_name}'] ), 1, 0)
         self.df[f'signal_up_{feature_name}'] = np.where( (self.df[f'norm_{feature_name}'] > self.df[f'upper_{feature_name}'] ), 1, 0)
 
+    def extract_sec_data(self, symbol, base_columns, rename_columns=None):
+        """
+        extract new asset data and merge it to the main asset data
+
+        Parameters
+        ----------
+        symbol (str): symbol to extract data
+        base_columns (list): list of columns to persist
+        rename_columns (dict): map of the new column names using pd.DataFrame.rename()
+
+        Returns
+        -------
+        None
+        """
+        begin_date = self.today - relativedelta(days = self.n_days)
+        begin_date_str = begin_date.strftime('%Y-%m-%d')
+        
+        stock = yf.Ticker(symbol)
+        df = stock.history(period=self.data_window)
+        df = df.sort_values('Date')
+        df.reset_index(inplace=True)
+        df['Date'] = pd.to_datetime(df['Date'], format='mixed',utc=True).dt.date
+        df['Date'] = pd.to_datetime(df['Date'])
+        df = df[df.Date >= begin_date_str ]
+        df = df[base_columns]
+        if rename_columns:
+            df = df.rename(columns=rename_columns)
+        right_df = df.copy()
+        
+        dates_vector = self.df.Date.to_frame()
+        right_df = dates_vector.merge(right_df, on ='Date',how = 'left')
+        right_df = right_df.fillna(method = 'bfill')
+        right_df = right_df.fillna(method = 'ffill')
+
+        self.df = self.df.merge(right_df, on ='Date',how = 'left')
+        self.df = self.df.sort_values("Date")
+        del right_df
+        gc.collect()
+
+    def lag_log_return(self, lags, feature, feature_name=False):
+        """
+        compute log return given some lags
+
+        Parameters
+        ----------
+        lags (int): lag to apply log return
+        feature (str): feature to apply log return
+        feature_name (str): rename resuling name
+
+        Returns
+        -------
+        None
+        """
+
+        feature_name = feature_name if feature_name else f"{feature}_log_return"
+        self.df[feature_name] = np.log(self.df[feature]/self.df[feature].shift(lags))
+    
+    def produce_log_volatility(self, trad_days, feature, feature_name=False):
+        """
+        compute log return given some lags
+
+        Parameters
+        ----------
+        trad_days (int): window function to calculate standard deviation
+        feature (str): feature to apply computation
+        feature_name (str): resulting feature name
+
+        Returns
+        -------
+        None
+        """
+        feature_name = feature_name if feature_name else f"{feature}_log_return_{trad_days}"
+        self.df[feature_name] = self.df.sort_values("Date")[feature].rolling(window = trad_days).std()*np.sqrt(252)
+
     def signal_plotter(self, feature_name):
+
+        """
+        display analysis plot of a feature with high and low signals
+
+        Parameters
+        ----------
+        feature_name (str): feature name
+
+        Returns
+        -------
+        None
+        """
+
         fig, axs = plt.subplots(1, 3,figsize=(17,5))
-        
+
         axs[0].plot(self.df[f'upper_{feature_name}'],color = 'grey', linestyle='--')
         axs[0].plot(self.df[f'lower_{feature_name}'],color = 'grey', linestyle='--')
         axs[0].plot(self.df[f'norm_{feature_name}'])
-        
+
         plot_acf(self.df[feature_name].dropna(),lags=25,ax = axs[1])
         axs[1].set_title(f'acf {feature_name}')
-        
+
         plot_pacf(self.df[feature_name].dropna(),lags=25,ax = axs[2])
         axs[2].set_title(f'pacf {feature_name}')
-        
+
         fig.show()
 
     def log_features_standard(self, feature_name):
+        """
+        save resulting feature names in an standard structure
+
+        Parameters
+        ----------
+        feature_name (str): feature name
+
+        Returns
+        -------
+        None
+        """
         self.features.append(feature_name)
         self.signals.append(f'signal_up_{feature_name}')
         self.signals.append(f'signal_low_{feature_name}')
-    
-    #######################
-    #### to be deprecated ####
-    def spread_MA(self, ma1, ma2, limit = 1.95, plot = False, save_features = False):
-
-        self.df[f'MA_{ma1}'] = (self.df.sort_values("Date")["Close"].transform(lambda x: x.rolling(ma1, min_periods=1).mean()))
-        self.df[f'MA_{ma2}'] = (self.df.sort_values("Date")["Close"].transform(lambda x: x.rolling(ma2, min_periods=1).mean()))
-
-        self.ma1_column = f'MA_{ma1}'
-        self.ma2_column = f'MA_{ma2}'
-        self.df['MA_spread'] = self.df[f'MA_{ma1}'] - self.df[f'MA_{ma2}']
-
-        self.df['norm_MA_spread'] =  (self.df['MA_spread'] - self.df['MA_spread'].mean())/self.df['MA_spread'].std()
-        mean_ = self.df['norm_MA_spread'].mean()
-        self.df['rollstd_MA_spread'] = self.df.sort_values("Date")["norm_MA_spread"].rolling(50).std()
-
-        self.df['upper_MA_spread'] = limit*self.df['rollstd_MA_spread'] + mean_
-        self.df['lower_MA_spread'] = -limit*self.df['rollstd_MA_spread'] + mean_
-
-        self.df['signal_low_MA_spread'] = np.where( (self.df['norm_MA_spread'] < self.df['lower_MA_spread'] ), 1, 0)
-        self.df['signal_up_MA_spread'] = np.where( (self.df['norm_MA_spread'] > self.df['upper_MA_spread'] ), 1, 0)
-        
-        ### ploting purposes
-        self.df[f"Roll_mean_{ma1}"] = (
-            self.df.sort_values("Date")["Close"]
-            .transform(lambda x: x.rolling(ma1, min_periods=1).mean())
-        )
-        self.df[f"Roll_mean_{ma2}"] = (
-            self.df.sort_values("Date")["Close"]
-            .transform(lambda x: x.rolling(ma2, min_periods=1).mean())
-        )
-        
-
-        print('--------------------------------------------------------------------')
-        if save_features:
-            self.features.append('MA_spread')
-            self.signals.append('signal_low_MA_spread')
-            self.signals.append('signal_up_MA_spread')
-            self.settings_spread_ma = {'ma1':ma1, 'ma2':ma2, 'limit':limit}  
-            
-        if plot:
-
-            fig, axs = plt.subplots(1, 3,figsize=(21,4))
-
-            axs[0].plot(self.df['Date'],self.df['norm_MA_spread'])
-            axs[0].plot(self.df['Date'],self.df['upper_MA_spread'], linestyle='--')
-            axs[0].plot(self.df['Date'],self.df['lower_MA_spread'], linestyle='--')
-            axs[0].set_title('MA_spread series')
 
-            plot_acf(self.df['MA_spread'].dropna(),lags=25, ax=axs[1])
-            axs[1].set_title('acf MA_spread series')
-
-            plot_pacf(self.df['MA_spread'].dropna(),lags=25, ax=axs[2])
-            axs[2].set_title('acf MA_spread series')
-            plt.show()
-    ##################################################
-    
     def relative_spread_MA(self, ma1, ma2, threshold = 1.95, plot = False, save_features = False):
-    
+        """
+        perform relative moving average features, one for short term and another for long/mid term
+
+        Parameters
+        ----------
+        ma1 (int): short term moving average window
+        ma2 (int): long/mid term moving average window
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+
+        Returns
+        -------
+        None
+        """
         feature_name = 'rel_MA_spread'
 
         self.df[f'MA_{ma1}'] = (self.df.sort_values("Date")["Close"].transform(lambda x: x.rolling(ma1, min_periods=1).mean()))
@@ -605,16 +834,27 @@ class stock_eda_panel(object):
             .transform(lambda x: x.rolling(ma2, min_periods=1).mean())
         )
 
-        print('--------------------------------------------------------------------')
         if save_features:
             self.log_features_standard(feature_name)
-            self.settings_relative_spread_ma = {'ma1':ma1, 'ma2':ma2, 'threshold':threshold}  
+            self.settings_relative_spread_ma = {'ma1':ma1, 'ma2':ma2, 'threshold':threshold}
 
         if plot:
-
             self.signal_plotter(feature_name)
-    
+
     def pair_feature(self, pair_symbol, plot = False):
+        """
+        initialize pair feature data extraction and analysis
+
+        Parameters
+        ----------
+        pair_symbol (str): symbol of the pair asset to extract
+        plot (boolean): True to display plot
+        
+        Returns
+        -------
+        None
+        """
+
         self.pair_symbol = pair_symbol
         begin_date = self.today - relativedelta(days = self.n_days)
         begin_date_str = begin_date.strftime('%Y-%m-%d')
@@ -627,7 +867,7 @@ class stock_eda_panel(object):
         df['Date'] = pd.to_datetime(df['Date'])
         df = df[df.Date >= begin_date_str ]
         self.pair_df = df
-        
+
         #### converting the same index ####
         dates_vector = self.df.Date.to_frame()
         self.pair_df = dates_vector.merge(self.pair_df, on ='Date',how = 'left')
@@ -653,8 +893,40 @@ class stock_eda_panel(object):
             plt.plot(self.df['Date'],asset_2_values,label = asset_2)
             plt.legend()
             plt.show()
-    
+
+    def smooth_logrets_interaction_term(self, feature_interact_with, resulting_feature_name="persisted_clip_diff_smooths", rollmean_window = 5, ext_threhold=0.015, persist_days = 3, save_features=False):
+        """
+        create an interaction term that is going to compare the distance of asset wolling window mean and market rolling window mean.
+        then get the outliers or high values using abs and this value persist for some days
+        goal persist big differences of market and asset returns
+
+        feature_interact_with: name of the market return
+        rollmean_window: rolling window or smoothing number of days
+        ext_threhold: threshold
+        persist_days: number of days to persis the signal
+        """
+        self.df["smooth_log_return"] = self.df['log_return'].rolling(rollmean_window).mean().values
+        self.df["smooth_market_log_return"] = self.df[feature_interact_with].rolling(rollmean_window).mean().values
+        self.df["diff_smooths"] = self.df["smooth_market_log_return"]-self.df["smooth_log_return"]
+        self.df["clip_diff_smooths"] = np.where(np.abs(self.df["diff_smooths"]) > ext_threhold, self.df["diff_smooths"] , 0)
+        self.df[resulting_feature_name] = self.df['clip_diff_smooths'].rolling(persist_days).mean().values
+        self.df = self.df.drop(columns=["smooth_log_return","smooth_market_log_return","diff_smooths","clip_diff_smooths"])
+
     def calculate_cointegration(self,series_1, series_2):
+        """
+        calculate cointegration score for two time series
+
+        Parameters
+        ----------
+        series_1 (pd.series): time series
+        series_2 (pd.series): time series
+        
+        Returns
+        -------
+        coint_flag (boolean): 1 if the p_value cointegration_t are lower than 0.05 and critical value
+        hedge_value (float): beta from the regression model
+        """
+
         coint_flag = 0
         coint_res = coint(series_1, series_2)
         coint_t = coint_res[0]
@@ -666,9 +938,22 @@ class stock_eda_panel(object):
         coint_flag = 1 if p_value < 0.05 and coint_t < critical_value else 0
 
         return coint_flag, hedge_value
-    
-    def produce_pair_score_plot(self, window, z_threshold, plot = False, save_features = False):
 
+    def produce_pair_score_plot(self, window, z_threshold, plot = False, save_features = False):
+        """
+        display analysis of the pair feature and save results in case if needed
+
+        Parameters
+        ----------
+        window (int): window to apply to the rolling spread between pair and main asset
+        z_threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         spread_series = pd.Series(self.df.pair_spread)
         mean = spread_series.rolling(center = False, window = window).mean()
         std = spread_series.rolling(center = False, window = window).std()
@@ -677,11 +962,11 @@ class stock_eda_panel(object):
         self.df['pair_z_score'] = z_score
         self.df['signal_low_pair_z_score'] = np.where(self.df['pair_z_score'] < -z_threshold, 1, 0)
         self.df['signal_up_pair_z_score'] = np.where(self.df['pair_z_score'] > z_threshold, 1, 0)
-        
+
         if save_features:
             self.log_features_standard('pair_z_score')
-            self.settings_pair_feature = {'pair_symbol':self.pair_symbol,'window':window, 'z_threshold':z_threshold}  
-            
+            self.settings_pair_feature = {'pair_symbol':self.pair_symbol,'window':window, 'z_threshold':z_threshold}
+
         if plot:
             pvalue = round(adfuller(z_score.dropna().values)[1],4)
             print(f'p value of the rolling z-score is {pvalue}')
@@ -695,7 +980,7 @@ class stock_eda_panel(object):
             axs[0,0].axhline(y=0, color='blue', linestyle='-.')
             axs[0,0].plot(self.df.pair_z_score)
             axs[0,0].set_title('z score from the spread')
-            
+
             axs[0,1].plot(self.df['Date'],self.df['pair_spread'])
             axs[0,1].plot(self.df['Date'],np.where(self.df['signal_low_pair_z_score'] == 1, self.df['pair_spread'], np.nan),'o-r',color = 'red')
             axs[0,1].plot(self.df['Date'],np.where(self.df['signal_up_pair_z_score'] == 1, self.df['pair_spread'], np.nan),'o-r',color = 'green')
@@ -704,44 +989,27 @@ class stock_eda_panel(object):
 
             plot_acf(self.df['pair_z_score'].dropna(),lags=25, ax=axs[1,0])
             axs[1,0].set_title('acf pair_z_score')
-            
+
             plot_pacf(self.df['pair_z_score'].dropna(),lags=25, ax=axs[1,1])
             axs[1,1].set_title('pacf pair_z_score')
-            
-            plt.show()
-
-    #######################
-    #### to be deprecated ####
-    def get_count_feature(self, rolling_window, threshold, plot = False, save_features = False):
-
-        # negative countiing and rolling countingng
-        self.df['RetClose'] = self.df['Close'].pct_change()
-        self.df['roll_pos_counting'] = np.where(self.df['RetClose'].shift(1) > 0,1,0 )
-        self.df['roll_pos_counting'] = self.df['roll_pos_counting'].rolling(window = rolling_window).sum()
-
-        mean = self.df['roll_pos_counting'].mean()
-        std = self.df['roll_pos_counting'].std()
-        self.df['norm_counting'] =  (self.df['roll_pos_counting'] - mean )/std
 
-        self.df['signal_up_roll_pos_counting'] = np.where((self.df['norm_counting'] > threshold),1,0)
-        self.df['signal_low_roll_pos_counting'] = np.where((self.df['norm_counting'] < -threshold),1,0)
-        
-        if save_features:
-            self.features.append('roll_pos_counting')
-            self.signals.append('signal_up_roll_pos_counting')
-            self.signals.append('signal_low_roll_pos_counting')
-            self.settings_count_features = {'rolling_window':rolling_window, 'threshold':threshold}  
-            
-        if plot:
-            fig = plt.figure(figsize = (10,4))
-            plt.plot(self.df['Date'],self.df.norm_counting)
-            plt.axhline(y=threshold, color='grey', linestyle='--')
-            plt.axhline(y=-threshold, color='grey', linestyle='--')
             plt.show()
-    #######################
-    
+
     def bidirect_count_feature(self, rolling_window, threshold, plot = False, save_features = False):
-    
+        """
+        perform negative and positive return counting in a given rolling time window
+
+        Parameters
+        ----------
+        rolling_window (int): window to apply to positive and negative returns
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'bidirect_counting'
         # negative countiing and rolling countingng
         self.df['RetClose'] = self.df['Close'].pct_change()
@@ -757,7 +1025,7 @@ class stock_eda_panel(object):
 
         if save_features:
             self.log_features_standard(feature_name)
-            self.settings_bidirect_count_features = {'rolling_window':rolling_window, 'threshold':threshold}  
+            self.settings_bidirect_count_features = {'rolling_window':rolling_window, 'threshold':threshold}
 
         if plot:
             fig = plt.figure(figsize = (10,4))
@@ -766,47 +1034,21 @@ class stock_eda_panel(object):
             plt.plot(self.df['Date'],self.df[f'lower_{feature_name}'], linestyle='--')
             plt.show()
 
-    #######################
-    #### to be deprecated ####
-    def get_range_feature(self, window, up_threshold, low_threshold, plot = False, save_features = False):
-
-        self.df["Range"] = self.df["High"] / self.df["Low"] - 1
-        self.df['Avg_range'] = self.df['Range'].rolling(window = 5).mean()
-        self.df['dist_range'] = self.df['Range'] - self.df['Avg_range']
-        self.df['norm_dist_range'] = (self.df['dist_range'] - self.df['dist_range'].mean())/ self.df['dist_range'].std()
-
-        mean_ = self.df['norm_dist_range'].mean()
-        self.df[f'std_norm_dist_range'] = (self.df.sort_values("Date")["norm_dist_range"].transform(lambda x: x.rolling(window, min_periods=1).std()))
-
-        self.df['up_bound_norm_dist_range'] = up_threshold*self.df['std_norm_dist_range'] + mean_
-        self.df['low_bound_norm_dist_range'] = -low_threshold*self.df['std_norm_dist_range'] + mean_
-
-        self.df['signal_up_dist_range'] = np.where(self.df['norm_dist_range'] > self.df['up_bound_norm_dist_range'],1,0 )
-        self.df['signal_low_dist_range'] = np.where(self.df['norm_dist_range'] < self.df['low_bound_norm_dist_range'],1,0 )
-        
-        if save_features:
-            self.features.append('dist_range')
-            self.signals.append('signal_up_dist_range')
-            self.signals.append('signal_low_dist_range')
-            self.settings_price_range = {'window':window, 'up_threshold':up_threshold, 'low_threshold':low_threshold} 
-
-        if plot:
-            fig, axs = plt.subplots(2, 2,figsize=(17,11))
-
-            axs[0,0].plot(self.df['Range'])
-            axs[0,0].set_title('range')
-
-            axs[0,1].plot(self.df['Avg_range'])
-            axs[0,1].set_title('Avg_range')
-
-            axs[1,0].plot(self.df['up_bound_norm_dist_range'],color = 'grey', linestyle='--')
-            axs[1,0].plot(self.df['low_bound_norm_dist_range'],color = 'grey', linestyle='--')
-            axs[1,0].plot(self.df['norm_dist_range'])
-            axs[1,0].set_title('norm_dist_range')
-    #######################
-    
     def get_relative_range_feature(self, window, threshold, plot = False, save_features = False):
-    
+        """
+        perform relative spread of opening and closing price
+
+        Parameters
+        ----------
+        window (int): window to apply to the feature
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'CO_Range'
         self.df[feature_name] = self.df["Close"] / self.df["Open"]-1
         self.df[f'norm_{feature_name}'] = (self.df[feature_name] - self.df[feature_name].mean())/ self.df[feature_name].std()
@@ -822,7 +1064,7 @@ class stock_eda_panel(object):
 
         if save_features:
             self.log_features_standard(feature_name)
-            self.settings_relative_price_range = {'window':window, 'threshold':threshold} 
+            self.settings_relative_price_range = {'window':window, 'threshold':threshold}
 
         if plot:
             fig, axs = plt.subplots(1, 2,figsize=(14,5))
@@ -835,46 +1077,24 @@ class stock_eda_panel(object):
             axs[1].plot(self.df[f'norm_{feature_name}'])
             axs[1].set_title(f'norm_{feature_name}')
 
-    #######################
-    #### to be deprecated ####
-    def rsi_feature(self, window, lag_rsi_ret, threshold, plot = False, save_features = False):
-
-        rsi = RSIIndicator(close = self.df['Close'], window = window).rsi()
-        self.df['RSI'] = rsi 
-        self.df['RSI_ret'] = self.df['RSI']/self.df['RSI'].shift(lag_rsi_ret)
-
-        mean = self.df['RSI_ret'].mean()
-        std = self.df['RSI_ret'].std()
-        self.df['norm_RSI_ret'] = (self.df['RSI_ret']-mean)/std
-        self.df['signal_up_RSI_ret'] = np.where(self.df['norm_RSI_ret'] > threshold,1,0)
-        self.df['signal_low_RSI_ret'] = np.where(self.df['norm_RSI_ret'] < -threshold,1,0)
-
-        if save_features:
-            self.features.append('RSI_ret')
-            self.signals.append('signal_up_RSI_ret')
-            self.signals.append('signal_low_RSI_ret')
-            self.settings_rsi_feature= {'window':window, 'lag_rsi_ret':lag_rsi_ret, 'threshold':threshold}
-
-        if plot:
-            fig, axs = plt.subplots(1, 3,figsize=(17,5))
-
-            axs[0].plot(self.df.norm_RSI_ret)
-            axs[0].axhline(y=threshold, color='grey', linestyle='--')
-            axs[0].axhline(y=-threshold, color='grey', linestyle='--')
-
-            plot_acf(self.df['RSI_ret'].dropna(),lags=25,ax = axs[1])
-            axs[1].set_title('acf RSI_ret')
-
-            plot_pacf(self.df['RSI_ret'].dropna(),lags=25,ax = axs[2])
-            axs[2].set_title('pacf RSI_ret')
-
-            fig.show()
-    #######################
-    
     def rsi_feature_improved(self, window, threshold, plot = False, save_features = False):
+        """
+        perform relative strength index
+
+        Parameters
+        ----------
+        window (int): window to apply to the feature
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'RSI'
         rsi = RSIIndicator(close = self.df['Close'], window = window).rsi()
-        self.df[feature_name] = rsi 
+        self.df[feature_name] = rsi.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
         self.compute_clip_bands(feature_name,threshold)
 
         if save_features:
@@ -883,54 +1103,22 @@ class stock_eda_panel(object):
 
         if plot:
             self.signal_plotter(feature_name)
-            
-    #######################
-    #### to be deprecated ####
-    def days_features(self, window_day, limit, plot = False, save_features = False):
-
-        self.df['dow'] = self.df.Date.dt.dayofweek
-        self.df['dow'] = self.df['dow'].astype('str')
-
-        self.df['target_mean_input'] = (self.df.sort_values("Date").groupby('dow')['roll_mean_log_return'].transform(lambda x: x.rolling(window_day, min_periods=1).mean()))
-
-        mean = self.df['target_mean_input'].mean()
-        std = self.df['target_mean_input'].std()
-
-        self.df['norm_dow_input'] = (self.df['target_mean_input']-mean)/std
-        mean_ = self.df['norm_dow_input'].mean()
-        self.df['std_dow_input'] = self.df.sort_values("Date")["norm_dow_input"].rolling(50).std()
-
-        self.df['up_dow_input'] = limit*self.df['std_dow_input'] + mean_
-        self.df['low_dow_input'] = -limit*self.df['std_dow_input'] - mean_
-
-        self.df['signal_up_target_mean_input'] = np.where(self.df['norm_dow_input'] > self.df['up_dow_input'],1,0)
-        self.df['signal_low_target_mean_input'] = np.where(self.df['norm_dow_input'] < self.df['low_dow_input'],1,0)
 
-        if save_features:
-
-            self.features.append('target_mean_input')
-            self.signals.append('signal_up_target_mean_input')
-            self.signals.append('signal_low_target_mean_input')
-            self.settings_days_features = {'window_day':window_day, 'limit':limit}
-
-        if plot:
-            fig, axs = plt.subplots(1, 3,figsize=(17,5))
-
-            axs[0].plot(self.df['norm_dow_input']) 
-            axs[0].plot(self.df['up_dow_input'], linestyle='--')
-            axs[0].plot(self.df['low_dow_input'], linestyle='--')
-
-            plot_acf(self.df['norm_dow_input'].dropna(),lags=25,ax = axs[1])
-            axs[1].set_title('acf day feature')
-
-            plot_pacf(self.df['norm_dow_input'].dropna(),lags=25,ax = axs[2])
-            axs[2].set_title('pacf day feature')
-
-            fig.show()
-    #######################
-    
     def days_features_bands(self, window, threshold, plot = False, save_features = False):
-
+        """
+        compute mean returns for a given day of the week in a window scope per day 
+
+        Parameters
+        ----------
+        window (int): window to apply to the feature
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         self.df['dow'] = self.df.Date.dt.dayofweek
         self.df['dow'] = self.df['dow'].astype('str')
 
@@ -947,65 +1135,22 @@ class stock_eda_panel(object):
 
         if plot:
             self.signal_plotter(feature_name)
-            
-    #######################
-    #### to be deprecated ####
-    def analysis_volume(self,lag_volume, threshold, window, plot = False, save_features = False):
-    
-        self.df['log_Volume'] = np.log(self.df['Volume'])
-        self.df['ret_log_Volume'] = self.df['log_Volume'].pct_change(lag_volume)
-
-        self.df['norm_ret_log_Volume'] = (self.df['ret_log_Volume'] - self.df['ret_log_Volume'].mean())/ self.df['ret_log_Volume'].std()
-        mean_ = self.df['norm_ret_log_Volume'].mean()
-        self.df[f'std_norm_ret_log_Volume'] = (self.df.sort_values("Date")["norm_ret_log_Volume"].transform(lambda x: x.rolling(window, min_periods=1).std()))
 
-        self.df['up_bound_ret_log_Volume'] = threshold*self.df['std_norm_ret_log_Volume'] + mean_
-        self.df['low_bound_ret_log_Volume'] = -threshold*self.df['std_norm_ret_log_Volume'] + mean_
-
-        self.df['signal_up_ret_log_Volume'] = np.where(self.df['norm_ret_log_Volume'] > self.df['up_bound_ret_log_Volume'],1,0 )
-        self.df['signal_low_ret_log_Volume'] = np.where(self.df['norm_ret_log_Volume'] < self.df['low_bound_ret_log_Volume'],1,0 )
-
-        if save_features:
-            self.features.append('ret_log_Volume')
-            self.signals.append('signal_up_ret_log_Volume')
-            self.signals.append('signal_low_ret_log_Volume')
-            self.settings_volume_feature= {'lag_volume':lag_volume, 'threshold':threshold, 'window':window}
-        if plot:
-            fig, axs = plt.subplots(3, 2,figsize=(11,13))
-            axs[0,0].plot(self.df.Date, self.df.Volume)
-            axs[0,0].set_title('Volume')
-            axs[0,1].plot(self.df.Date, self.df.log_Volume)
-            axs[0,1].set_title('log Volume')
-
-            plot_acf(self.df['log_Volume'].dropna(),lags=25, ax = axs[1,0])
-            axs[1,0].set_title('acf log_Volume')
-            plot_pacf(self.df['log_Volume'].dropna(),lags=25, ax = axs[1,1])
-            axs[1,1].set_title('pacf log_Volume')
-
-            plot_acf(self.df['ret_log_Volume'].dropna(),lags=25, ax = axs[2,0])
-            axs[2,0].set_title('acf ret_log_Volume')
-            plot_pacf(self.df['ret_log_Volume'].dropna(),lags=25, ax = axs[2,1])
-            axs[2,1].set_title('pacf ret_log_Volume')
-
-            plt.show()
-
-            print('--------------------------------------------------------------')
-
-            fig, axs = plt.subplots(1, 2,figsize=(10,4))
-
-            axs[0].plot(self.df.Date, self.df.norm_ret_log_Volume)
-            axs[0].plot(self.df.Date, self.df.up_bound_ret_log_Volume)
-            axs[0].plot(self.df.Date, self.df.low_bound_ret_log_Volume)
-            axs[0].set_title('norm_ret_log_Volume')
-
-            axs[1].plot(self.df.Date, self.df.std_norm_ret_log_Volume)
-            axs[1].set_title('std_norm_ret_log_Volume')
-
-            plt.show()
-    #######################
-    
     def analysis_smooth_volume(self, window, threshold, plot = False, save_features = False):
-    
+        """
+        compute feature of thrading volumes
+
+        Parameters
+        ----------
+        window (int): window to apply to the feature
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'smooth_Volume'
         self.df[feature_name] = np.log(self.df['Volume'])
         # self.df[feature_name] = self.df['log_Volume'].rolling(window).mean()
@@ -1039,7 +1184,7 @@ class stock_eda_panel(object):
 
             fig, axs = plt.subplots(1,2,figsize=(10,4))
 
-            axs[0].plot(self.df[f'{feature_name}']) 
+            axs[0].plot(self.df[f'{feature_name}'])
             axs[0].set_title(f'{feature_name}')
 
             axs[1].plot(self.df[f'z_{feature_name}'], linestyle='--')
@@ -1048,9 +1193,23 @@ class stock_eda_panel(object):
             plt.show()
 
     def roc_feature(self, window, threshold, plot = False, save_features = False):
+        """
+        perform price rate of change
+
+        Parameters
+        ----------
+        window (int): window to apply to the feature
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'ROC'
         roc = ROCIndicator(close = self.df['Close'], window = window).roc()
-        self.df[feature_name] = roc 
+        self.df[feature_name] = roc.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
         self.compute_clip_bands(feature_name,threshold)
 
         if save_features:
@@ -1058,11 +1217,27 @@ class stock_eda_panel(object):
             self.settings_roc_feature = {'window':window, 'threshold':threshold}
         if plot:
             self.signal_plotter(feature_name)
-    
+
     def stoch_feature(self, window, smooth1, smooth2, threshold, plot = False, save_features = False):
+        """
+        perform stochastic oscilator RSI feature
+
+        Parameters
+        ----------
+        window (int): window to apply to the feature
+        smooth1 (int): smoothing parameter 1
+        smooth2 (int): smoothing parameter 2
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'STOCH'
         stoch = StochRSIIndicator(close = self.df['Close'], window = window, smooth1=smooth1, smooth2=smooth2).stochrsi()
-        self.df[feature_name] = stoch
+        self.df[feature_name] = stoch.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
         self.compute_clip_bands(feature_name,threshold)
 
         if save_features:
@@ -1072,9 +1247,24 @@ class stock_eda_panel(object):
             self.signal_plotter(feature_name)
 
     def stochastic_feature(self, window, smooth, threshold, plot = False, save_features = False):
+        """
+        perform stochastic oscilator feature
+
+        Parameters
+        ----------
+        window (int): window to apply to the feature
+        smooth (int): smoothing parameter 
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'STOCHOSC'
         stochast = StochasticOscillator(close = self.df['Close'], high = self.df['High'], low = self.df['Low'], window = window,smooth_window=smooth).stoch()
-        self.df[feature_name] = stochast
+        self.df[feature_name] = stochast.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
         self.compute_clip_bands(feature_name,threshold)
 
         if save_features:
@@ -1084,9 +1274,23 @@ class stock_eda_panel(object):
             self.signal_plotter(feature_name)
 
     def william_feature(self, lbp, threshold, plot = False, save_features = False):
+        """
+        perfom fast stochastic oscilator or william indicator
+
+        Parameters
+        ----------
+        lbp (int): look back parameter
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'WILL'
-        will = WilliamsRIndicator(close = self.df['Close'], high = self.df['High'], low = self.df['Low'], lbp = lbp).williams_r() 
-        self.df[feature_name] = will
+        will = WilliamsRIndicator(close = self.df['Close'], high = self.df['High'], low = self.df['Low'], lbp = lbp).williams_r()
+        self.df[feature_name] = will.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
         self.compute_clip_bands(feature_name,threshold)
 
         if save_features:
@@ -1096,9 +1300,23 @@ class stock_eda_panel(object):
             self.signal_plotter(feature_name)
 
     def vortex_feature(self, window, threshold, plot = False, save_features = False):
+        """
+        perform vortex oscilator
+
+        Parameters
+        ----------
+        window (int): window to apply to the feature
+        threshold (float): alpha or z thrsholds for the normalized feature
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         feature_name = 'VORTEX'
         vortex = VortexIndicator(close = self.df['Close'], high = self.df['High'], low = self.df['Low'], window = window).vortex_indicator_diff()
-        self.df[feature_name] = vortex
+        self.df[feature_name] = vortex.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
         self.compute_clip_bands(feature_name,threshold)
 
         if save_features:
@@ -1107,11 +1325,160 @@ class stock_eda_panel(object):
         if plot:
             self.signal_plotter(feature_name)
 
-    def pair_index_feature(self, pair_symbol, feature_label, window, threshold, plot = False, save_features = False):
+    def minmax_pricefeature(self, type_func, window, distance = False, plot = False, save_features = False):
+        """
+        perform relative price/distance with respect to the min/max price in a given time scope
+
+        Parameters
+        ----------
+        type_func (str): either min or max
+        window (int): window scope
+        distance (boolean): if true, get distance feature else relative feature
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
+        if type_func == 'min':
+            self.df['Price_ref'] = self.df[['Open','High', 'Low','Close']].min(axis = 1)
+        elif type_func == 'max':
+            self.df['Price_ref'] = self.df[['Open','High', 'Low','Close']].max(axis = 1)
+
+        init_shape = self.df.shape[0]
+        df_date = self.df[['Date','Price_ref']].rename(columns = {'Date':'Date_ref'}).copy()
+        
+        self.df = self.df.rename(columns = {'Price_ref':'Price_to_use'})
+        
+        if type_func == 'min':
+            self.df[f'window_price'] = (self.df.sort_values("Date")["Price_to_use"].transform(lambda x: x.rolling(window, min_periods=1).min()))
+        elif type_func == 'max':
+            self.df[f'window_price'] = (self.df.sort_values("Date")["Price_to_use"].transform(lambda x: x.rolling(window, min_periods=1).max()))
+        
+        
+        self.df = self.df.merge(df_date, left_on = 'window_price', right_on = 'Price_ref', how = 'left')
+        self.df['date_span'] = self.df['Date'] - self.df['Date_ref']
+        
+        self.df['RN'] = self.df.sort_values(['date_span'], ascending=False).groupby(['Date']).cumcount() + 1
+        self.df = self.df[self.df['RN'] == 1]
+
+        if distance:
+            self.df[f'{type_func}_distance_to_price'] = pd.to_numeric(self.df['date_span'].dt.days, downcast='integer')
+
+        if not distance:
+            if type_func == 'min':
+                self.df[f'{type_func}_relprice'] = self.df['Price_to_use']/self.df['window_price']-1
+            
+            if type_func == 'max':
+                self.df[f'{type_func}_relprice'] = self.df['window_price']/self.df['Price_to_use']-1
+        
+        self.df = self.df.drop(columns = ['RN', 'date_span', 'Price_to_use', 'window_price', 'Date_ref','Price_ref'])
+        
+        end_shape = self.df.shape[0]
+
+        if init_shape != end_shape:
+            raise Exception("shapes are not the same")
+
+        if save_features:
+            if distance:
+                self.features.append(f'{type_func}_distance_to_price')
+                name_attr = f'{type_func}_distance'
+            if not distance:
+                self.features.append(f'{type_func}_relprice')
+                name_attr = f'{type_func}_relprice'
+                
+            setattr(self,f'settings_{name_attr}_pricefeature' , {'type_func': type_func, 'window': window, 'distance': distance})
+    
+    def expected_return(self, trad_days, feature, feature_name=False):
+        """
+        perform expected log return based on inversed shift of historical data and applying
+
+        Parameters
+        ----------
+        trad_days (int): window or differenciation
+        feature (int): feature to apply expected log return
+        feature_name (str): resulting feature name
+        
+        Returns
+        -------
+        None
+        """
+        feature_name = feature_name if feature_name else f"{feature}_log_return_{trad_days}"
+        tmp_names = list()
+        for ind in range(1,trad_days+1):
+            tmp_name = f"expected_{ind}"
+            self.df[tmp_name] = self.df[feature].shift(-ind)/self.df[feature]-1
+            tmp_names.append(tmp_name)
+        self.df[feature_name] = self.df[tmp_names].max(axis=1)
+        self.df = self.df.drop(columns = tmp_names)
+    
+    def rolling_feature(self, feature, window, function):
+        """
+        perform rolling (non expanding) window operation for a given feature
+
+        Parameters
+        ----------
+        feature (int): feature to apply window operation
+        window (int): window size
+        function (str): window function e.g MIN, MAX, AVG
+        
+        Returns
+        -------
+        None
+        """
+        feature_name = f"{feature}_{window}_{function}"
+        self.df[feature_name] = getattr(self.df.sort_values("Date")[feature].rolling(window), function)()
+
+    def time_distance(self, feature_base,feature_window, result_feature_name, max_window=None):
+        """
+        perform distancce time to a given window feature
+
+        Parameters
+        ----------
+        feature_base (str): name of the underlaying feature
+        feature_window (str): name of the window feature
+        result_feature_name (str): resulting feature name
+        max_window (int): apply a top value to the time to distance feature
+        
+        Returns
+        -------
+        None
+        """
+        self.df["Date_pivot"] = np.nan
+        self.df["Date_pivot"] = self.df["Date_pivot"].case_when([
+            (self.df[feature_base] == self.df[feature_window], self.df["Date"]), 
+
+        ])
+        self.df["Date_pivot"] = self.df.sort_values("Date")["Date_pivot"].fillna(method="ffill")
+        self.df[result_feature_name] = self.df["Date"] - self.df["Date_pivot"]
+        self.df[result_feature_name] = self.df[result_feature_name].dt.days
+        if max_window:
+            self.df[result_feature_name] = self.df[result_feature_name].clip(0,max_window)
+        self.df = self.df.drop(columns = ["Date_pivot"])
+
+    def pair_index_feature(self, pair_symbol, feature_label,threshold, window = None,ta_method='ROC',param_set=False,plot = False, save_features = False):
+        """
+        perform additional asset ROC feature, then a new feature is created in the main dataframe
+
+        Parameters
+        ----------
+        pair_symbol (str): symbol of the asset to extract the data
+        feature_label (str): name of the resulting feature
+        window (int): window to apply to the feature as default (this parameter is going to be deprecated)
+        threshold (float): alpha or z thrsholds for the normalized feature
+        param_set (dict): parameter set in case ta_method is other than ROC
+        ta_method (str): method to use, available RSI, ROC, VORTEX, STOCH 
+        plot (boolean): True to display plot
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         self.pair_index = pair_symbol
         begin_date = self.today - relativedelta(days = self.n_days)
         begin_date_str = begin_date.strftime('%Y-%m-%d')
-        
+
         if feature_label in self.df.columns:
             self.df = self.df.drop(columns = [feature_label])
 
@@ -1123,14 +1490,27 @@ class stock_eda_panel(object):
         df['Date'] = pd.to_datetime(df['Date'])
         df = df[df.Date >= begin_date_str ]
         self.pair_index_df = df
-        
+
         #### converting the same index ####
         dates_vector = self.df.Date.to_frame()
         self.pair_index_df = dates_vector.merge(self.pair_index_df, on ='Date',how = 'left')
         self.pair_index_df = self.pair_index_df.fillna(method = 'bfill')
         self.pair_index_df = self.pair_index_df.fillna(method = 'ffill')
-        
-        self.pair_index_df[feature_label] = ROCIndicator(close = self.pair_index_df['Close'], window = window).roc()
+
+        if ta_method == 'ROC':
+            window = window if window else param_set.get('window')
+            roc = ROCIndicator(close = self.pair_index_df['Close'], window = window).roc()
+            self.pair_index_df[feature_label] = roc.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
+        elif ta_method == 'RSI':
+            rsi = RSIIndicator(close = self.pair_index_df['Close'], **param_set).rsi()
+            self.pair_index_df[feature_label] = rsi.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
+        elif ta_method == 'VORTEX':
+            vortex = VortexIndicator(close = self.pair_index_df['Close'], high = self.pair_index_df['High'], low = self.pair_index_df['Low'], **param_set).vortex_indicator_diff()
+            self.pair_index_df[feature_label] = vortex.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
+        elif ta_method == 'STOCH':
+            stoch = StochRSIIndicator(close = self.pair_index_df['Close'], **param_set).stochrsi()
+            self.pair_index_df[feature_label] = stoch.replace([np.inf, -np.inf], 0).fillna(method = 'ffill')
+
         df_to_merge = self.pair_index_df[['Date',feature_label]]
         self.df = self.df.merge(df_to_merge, on ='Date',how = 'left')
 
@@ -1140,7 +1520,7 @@ class stock_eda_panel(object):
         if save_features:
             self.log_features_standard(feature_label)
             parameters = {feature_label:{'pair_symbol':pair_symbol, 'feature_label':feature_label, 'window':window,'threshold':threshold}}
-            try: 
+            try:
                 len(self.settings_pair_index_feature)
                 print('existing')
                 self.settings_pair_index_feature.append(parameters)
@@ -1153,10 +1533,21 @@ class stock_eda_panel(object):
             self.signal_plotter(feature_label)
 
     def produce_order_features(self, feature_name, save_features = False):
+        """
+        perform a feature that captures high and low values in an index. this is usefull to know duration/persistence of a signal
 
+        Parameters
+        ----------
+        feature_name (str): name of the feature
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
         signal_feature_name = f'discrete_signal_{feature_name}'
         order_feature_name = f'order_signal_{feature_name}'
-        
+
         self.df[signal_feature_name] = np.where(
             self.df[f'signal_up_{feature_name}'] == 1,1,
             np.where(
@@ -1173,14 +1564,107 @@ class stock_eda_panel(object):
         self.df[order_feature_name] = self.df.groupby('chain_id')["Date"].rank(method="first", ascending=True)
         self.df[order_feature_name] = self.df[order_feature_name]*self.df[signal_feature_name]
         self.df = self.df.drop(columns = [f'lag_{signal_feature_name}', 'breack', "chain_id"])
-        
+
         ## saving features
         if save_features:
             self.signals.append(signal_feature_name)
             self.signals.append(order_feature_name)
-            
+    
+    def get_order_feature_nosignal(self,feature_name, save_features=False):
+        """
+        perform a feature that captures number of steps after the end of a signal
+
+        Parameters
+        ----------
+        feature_name (str): name of the feature
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
+        order_feature_name = f'order_signal_{feature_name}'
+        ns_order_feature_name = f'ns_order_{feature_name}'
+        self.df = self.df.sort_values('Date')
+        self.df['lag_'] = self.df[order_feature_name].shift(1)
+        self.df['flag'] = np.where((self.df[order_feature_name] == 0) & (self.df['lag_']!=0),1,np.nan)
+        self.df = self.df.drop(columns=['lag_'])
+        self.df['order_'] = self.df.sort_values('Date').groupby(['flag']).cumcount() + 1
+        self.df['order_'] = self.df['order_'].fillna(method='ffill')
+        self.df['order_'] = np.where(self.df[order_feature_name]==0,self.df['order_'],0)
+        self.df = self.df.drop(columns=['flag'])
+        self.df['order_'] = self.df.sort_values('Date').groupby(['order_']).cumcount() + 1
+        norm_list = [f'norm_{feature_name}', f'z_{feature_name}', feature_name]
+        for norm_feature in norm_list:
+            try:
+                self.df['order_'] = np.sign(self.df[norm_feature])*self.df['order_']
+                break
+            except:
+                pass
+        self.df['order_'] = np.where(self.df[order_feature_name]==0,self.df['order_'],0)
+        self.df = self.df.rename(columns={'order_':ns_order_feature_name})
+        if save_features:
+            self.signals.append(ns_order_feature_name)
+
+    def compute_last_signal(self,feature, save_features = False):
+        """
+        perform two new features when signal is observed, one for the last duration of the previous chain, second for the last duration of the same sign signal
+
+        Parameters
+        ----------
+        feature_name (str): name of the feature
+        save_features (boolean): True to save feature configuration and feature names
+        
+        Returns
+        -------
+        None
+        """
+        def create_last_signal(df, feature, prefix, type ='0'):
+            if type == '0':
+                condition = df[f'order_signal_{feature}'] != 0
+            elif type == '+':
+                condition = df[f'order_signal_{feature}'] > 0
+            elif type == '-':
+                condition = df[f'order_signal_{feature}'] < 0
+            df[f'last_maxorder_{feature}'] = np.where(condition, df[f'order_signal_{feature}'],np.nan)
+            df['tmp_chain_index'] = df[f'last_maxorder_{feature}'].shift(-1)
+            df['last'] = np.where((df[f'last_maxorder_{feature}'] != 0) & (df['tmp_chain_index'].isna()),df[f'last_maxorder_{feature}'], np.nan )
+            df['last'] = df['last'].shift(1)
+            df[f'last_maxorder_{feature}'] = df['last'].fillna(method = 'ffill')
+            df = df.drop(columns = ['tmp_chain_index','last'])
+            df[f'last_maxorder_{feature}'] = np.where(df[f'order_signal_{feature}'] != 0,df[f'last_maxorder_{feature}'],np.nan)
+            df[f'last_maxorder_{feature}'] = df[f'last_maxorder_{feature}'].fillna(0)
+            df = df.rename(columns = {f'last_maxorder_{feature}':f'{prefix}_{feature}'})
+            return df
+        prefix0, prefix1, prefix2 = 'ldur', 'pos', 'neg'
+        self.df = create_last_signal(self.df, feature, prefix0, type ='0')
+        self.df = create_last_signal(self.df, feature, prefix1, type ='+')
+        self.df = create_last_signal(self.df, feature, prefix2, type ='-')
+
+        self.df[f'sldur_{feature}'] = np.where(
+            self.df[f'order_signal_{feature}'] > 0, self.df[f'{prefix1}_{feature}'],
+            np.where(
+                self.df[f'order_signal_{feature}'] < 0, self.df[f'{prefix2}_{feature}'],
+                0
+            )
+        )
+        self.df = self.df.drop(columns = [f'{prefix1}_{feature}',f'{prefix2}_{feature}'])
+        if save_features:
+                self.signals.append(f'sldur_{feature}')
+                self.signals.append(f'ldur_{feature}')
+
     def create_hmm_derived_features(self, lag_returns):
+        """
+        create features derived from hmm states features. Features are the index of the state, the duration of the state, chain raturn
+
+        Parameters
+        ----------
+        lag_returns (int): lag paramter (not used)
         
+        Returns
+        -------
+        None
+        """
         self.df = self.df.sort_values('Date')
         ## indexing chains
         self.df['lag_hmm_feature'] = self.df['hmm_feature'].shift(1)
@@ -1189,31 +1673,44 @@ class stock_eda_panel(object):
         self.df["chain_id"] = np.where(self.df['breack'] == 1,self.df["chain_id"],np.nan)
         self.df["chain_id"] = self.df["chain_id"].fillna(method='ffill')
         self.df["hmm_chain_order"] = self.df.groupby('chain_id')["Date"].rank(method="first", ascending=True)
-        
-        ### returns using the first element in a chain
-        self.df['first'] = np.where(self.df['hmm_chain_order'] == 1, self.df['Close'], np.nan)
-        self.df['first'] = self.df.sort_values('Date')['first'].fillna(method='ffill')
-        self.df['chain_return'] = (self.df['Close']/self.df['first'] -1) * 100
 
-        self.df = self.df.drop(columns = ['breack','first'])
+        ### returns using the windowsseeds
+        self.df['lag_chain_close'] = self.df.sort_values(by=["Date"]).groupby(['chain_id'])['Close'].shift(lag_returns)
+        self.df['chain_return'] = (self.df['Close']/self.df['lag_chain_close'] -1) * 100
+        self.df = self.df.drop(columns = ['breack'])
 
-    def cluster_hmm_analysis(self, n_clusters,features_hmm, test_data_size, seed, lag_returns_state=7, plot = False, save_features = False, model = False):
+    def cluster_hmm_analysis(self, n_clusters,features_hmm, test_data_size, seed, lag_returns_state=7, corr_threshold = 0.75, plot = False, save_features = False, model = False):
+        """
+        create or use a hmm model
+
+        Parameters
+        ----------
+        n_clusters (int): number of clusters or states to calculate
+        features_hmm (list): features to be considered in hmm model when training
+        test_data_size (int): size of the test data. Note that the remaining is going to be used as training data
+        seed (int): seed for the model inizialization
+        lag_returns_state (int) : lags for returns of the state
+        corr_threshold (float): correlation threshold for initial feature selection
+        plot (boolean): True to display hmm states analysis
+        save_features (boolean): True to save features and configurations
+        model (obj): if provided, no model will be trainend and the provided model will be used to get hmm features
+        
+        Returns
+        -------
+        None
+        """
         if not model:
-            
+
             df_new = self.df
-            pipeline_hmm = Pipeline([
-                ('selector', FeatureSelector(columns=features_hmm)),
-                ('fillna', MeanMedianImputer(imputation_method='median',variables=features_hmm)),
-                ('hmm',GaussianHMM(n_components =  n_clusters, covariance_type = 'full', random_state = seed))
-                ])
             data_train = df_new.iloc[:-test_data_size,:]
             data_test = df_new.iloc[-test_data_size:,:]
 
-            pipeline_hmm.fit(data_train)
-
+            th = trainer_hmm(data_train, features_hmm, n_clusters=n_clusters,corr_thrshold=corr_threshold, seed = seed)
+            th.train()
+            pipeline_hmm = th.hmm_model
             self.model_hmm = pipeline_hmm
             self.test_data_hmm = data_test
-            
+
             ### first feature: the hidden state
             self.df['hmm_feature'] = self.model_hmm.predict(self.df)
             self.create_hmm_derived_features(lag_returns = lag_returns_state)
@@ -1230,15 +1727,15 @@ class stock_eda_panel(object):
             hidden_states = pipeline_hmm.predict(data_test)
             data_test['HMM'] = hidden_states
             data_test['HMM_state'] =  data_test['HMM'].map(map_)
-            
+
         if model:
             self.df['hmm_feature'] = model.predict(self.df)
             self.create_hmm_derived_features(lag_returns = lag_returns_state)
-            
+
         if save_features:
             self.features.append('hmm_feature')
             self.features.append('hmm_chain_order')
-            self.settings_hmm = {'n_clusters':n_clusters,'features_hmm':features_hmm, 'test_data_size':test_data_size, 'seed':seed,'lag_returns_state':lag_returns_state }
+            self.settings_hmm = {'n_clusters':n_clusters,'features_hmm':features_hmm, 'test_data_size':test_data_size, 'seed':seed,'lag_returns_state':lag_returns_state, 'corr_threshold':corr_threshold }
 
         if plot:
 
@@ -1263,14 +1760,38 @@ class stock_eda_panel(object):
             fig.show()
 
     def sharpe_ratio(self, return_series, n_trad_days = 255, rf = 0.01):
+        """
+        perform sharpe ratio of a given time series return
+
+        Parameters
+        ----------
+        return_series (pd.series): time series of the returns
+        n_trad_days (int): trading days to anualize returns
+        rf (float): anual free risk rate
+        
+        Returns
+        -------
+        sharpe_ratio (float): sharpe ratio 
+        """
         nsqrt = np.sqrt(n_trad_days)
         mean = return_series.mean() * n_trad_days
         sigma = return_series.std() * nsqrt
         sharpe_ratio = round((mean-rf)/sigma,2)
         return sharpe_ratio
-    
+
     def treat_signal_strategy(self,test_data, strategy):
-    
+        """
+        helper method that treats signals and converts signals to 1 or 0
+
+        Parameters
+        ----------
+        test_data (pd.DataFrame): test data
+        strategy (list): features to get the strategy
+        
+        Returns
+        -------
+        test_data (pd.DataFrame): test data with extra columns that are the strategy (main_signal)
+        """
         hmm_states_list = [x for x in strategy if 'hmm_state_' in x]
         other_features = [x for x in strategy if x not in hmm_states_list]
 
@@ -1299,10 +1820,21 @@ class stock_eda_panel(object):
         elif len(hmm_states_list) == 0 and len(other_features) > 0:
             test_data['main_signal'] = np.where((test_data['features_signal'] == 1) & (test_data['hmm_signal'] == 0),1,0)
 
-        return test_data  
+        return test_data
 
     def stategy_simulator(self, features, hmm_feature = True):
+        """
+        execute strategy and get some performance metrics like sharpe ratio, return. This method creates some new attributes
 
+        Parameters
+        ----------
+        features (list): list of features to be tested as strategies
+        hmm_feature (boolean): include hmm feature
+        
+        Returns
+        -------
+        None
+        """
         columns_ = ['Date', 'Close','Open'] + features + ['HMM']
         states = list(self.df.hmm_feature.unique())
         states.sort()
@@ -1372,8 +1904,19 @@ class stock_eda_panel(object):
         self.strategy_log = df_returns_log
         self.best_strategy =  df_returns_log.iloc[0,:].strategy
         self.top_10_strategy = list(df_returns_log.iloc[0:10,:].strategy.values)
-        
+
     def viz_strategy(self, strategy):
+        """
+        display analysis plot of a given strategy
+
+        Parameters
+        ----------
+        strategy (list): list of features of the strategy
+        
+        Returns
+        -------
+        None
+        """
         test_data = self.test_data_strategy
 
         test_data = self.treat_signal_strategy(test_data, strategy)
@@ -1406,62 +1949,26 @@ class stock_eda_panel(object):
         plt.legend()
         plt.show()
 
-    ### deprecated ############################
-    def create_strategy(self, favourable_states):
-
-        test_data = self.test_data_hmm
-        # add MA signal
-        test_data.loc[test_data[self.ma1_column] > test_data[self.ma2_column], 'MA_signal'] = 1
-        test_data.loc[test_data[self.ma1_column] <= test_data[self.ma2_column], 'MA_signal'] = 0
-
-        # add hnn signal
-
-        test_data['HMM_signal'] =  np.where(test_data['HMM'].isin(favourable_states),1,0)
-
-        ## combined signals
-        test_data['main_signal'] = 0
-        test_data.loc[(test_data['MA_signal'] == 1) & (test_data['HMM_signal'] == 1), 'main_signal'] = 1
-        test_data['main_signal'] = test_data['main_signal'].shift(1)
-
-        ## benchmark return
-        test_data['lrets_bench'] = np.log(test_data['Close']/test_data['Close'].shift(1))
-        test_data['bench_prod'] = test_data['lrets_bench'].cumsum()
-        test_data['bench_prod_exp'] = np.exp(test_data['bench_prod']) - 1
-
-        ## strategy return
-        # test_data['lrets_strat'] = np.log(test_data['Open'].shift(-1)/test_data['Open']) * test_data['main_signal']
-        test_data['lrets_strat'] = np.log(test_data['Close'].shift(-1)/test_data['Close']) * test_data['main_signal']
-        test_data['lrets_prod'] = test_data['lrets_strat'].cumsum()
-        test_data['strat_prod_exp'] = np.exp(test_data['lrets_prod']) - 1
-        test_data.dropna(inplace = True)
-
-        bench_rets = round(test_data['bench_prod_exp'].values[-1]*100,1)
-        strat_rets = round(test_data['strat_prod_exp'].values[-1]*100,1)
-
-        bench_sharpe = self.sharpe_ratio(test_data['bench_prod_exp'].values)
-        strat_sharpe = self.sharpe_ratio(test_data['strat_prod_exp'].values)
-
-        print(f'returns benchmark {bench_rets}%')
-        print(f'returns strategy {strat_rets}%')
-        print('-----------------------------')
-        print(f'sharpe benchmark {bench_sharpe}')
-        print(f'sharpe strategy {strat_sharpe}')
-
-        fig = plt.figure(figsize = (10,4))
-        plt.plot(test_data['bench_prod_exp'])
-        plt.plot(test_data['strat_prod_exp'])
-        self.settings_hmm_states = {'favourable_states':favourable_states}
-    ################################################
-    
     def deep_dive_analysis_hmm(self, test_data_size, split = 'train'):
-    
+        """
+        display analysis plot hmm model
+
+        Parameters
+        ----------
+        test_data_size (int): test data size, the remaining is the train data
+        split (str): options (train or test). Split type to assess
+        
+        Returns
+        -------
+        None
+        """
         if split == 'train':
             df = self.df.iloc[:-test_data_size,:]
         elif split == 'test':
             df = self.df.iloc[-test_data_size:,:]
 
         ## returns plot
-        fig = px.box(df.sort_values('hmm_feature'), y = 'chain_return',x = 'hmm_feature', color = 'hmm_feature', 
+        fig = px.box(df.sort_values('hmm_feature'), y = 'chain_return',x = 'hmm_feature', color = 'hmm_feature',
                     height=400, width=1000, title = 'returns chain hmm feature')
         fig.add_shape(type='line',x0=-0.5,y0=0,x1=max(df.hmm_feature)+0.5,y1=0,line=dict(color='grey',width=1),xref='x',yref='y')
         fig.show()
@@ -1490,6 +1997,17 @@ class stock_eda_panel(object):
         del df
 
     def get_targets(self, steps):
+        """
+        produce regression target return taking future prices
+
+        Parameters
+        ----------
+        steps (int): number of lags and steps for future returns
+        
+        Returns
+        -------
+        None
+        """
         self.targets = list()
         self.target = list()
         columns = list()
@@ -1501,9 +2019,23 @@ class stock_eda_panel(object):
         self.df[f'mean_target'] = self.df[columns].mean(axis=1)
         self.target.append(f'mean_target')
         self.settings_target_lasts = {'steps':steps, 'type':'regression'}
-        
-    def get_categorical_targets(self, horizon, flor_loss, top_gain):
-    
+
+    def get_categorical_targets(self, horizon, flor_loss, top_gain, min_pos=1 , min_negs=1):
+        """
+        produce binary target return taking future prices. it produce two targets, one for high returns and another for low returns
+
+        Parameters
+        ----------
+        horizon (int): number of lags and steps for future returns
+        flor_loss (float): min loss return
+        top_gain (float): max gain return
+        min_pos (int): minimun number of positives to count in a window for target_up
+        min_negs (int): minimun number of negatives to count in a window for target_down
+
+        Returns
+        -------
+        None
+        """
         self.target = list()
         self.targets = list()
         columns = list()
@@ -1516,7 +2048,7 @@ class stock_eda_panel(object):
             self.df[f'target_{i}'] = np.where(self.df[f'target_{i}'] >= top_gain,1,0)
             columns.append(f'target_{i}')
         self.df[f'target_up'] = self.df[columns].sum(axis=1)
-        self.df[f'target_up'] = np.where(self.df[f'target_up'] >=1,1,0 )
+        self.df[f'target_up'] = np.where(self.df[f'target_up'] >=min_pos,1,0 )
         self.df = self.df.drop(columns = columns)
 
         for i in range(1,horizon+1):
@@ -1526,7 +2058,7 @@ class stock_eda_panel(object):
             self.df[f'target_{i}'] = np.where(self.df[f'target_{i}'] <= flor_loss,1,0)
             columns.append(f'target_{i}')
         self.df[f'target_down'] = self.df[columns].sum(axis=1)
-        self.df[f'target_down'] = np.where(self.df[f'target_down'] >= 1,1,0 )
+        self.df[f'target_down'] = np.where(self.df[f'target_down'] >= min_negs,1,0 )
         self.df = self.df.drop(columns = columns)
 
         self.targets.append('target_up')
@@ -1535,7 +2067,19 @@ class stock_eda_panel(object):
         self.settings_target_lasts = {'horizon':horizon, 'flor_loss':flor_loss, 'top_gain':top_gain, 'type': 'classification'}
 
     def get_configurations(self,test_data_size =250, val_data_size = 250, model_type = False):
-        
+        """
+        produce configuration dictionary that were saved in the feature generation methods if save_features was activated
+
+        Parameters
+        ----------
+        test_data_size (int): test data size
+        val_data_size (int): validation data size 
+        model_type (str): model type, options: 'Forecaster','Classifier'
+
+        Returns
+        -------
+        None
+        """
         self.settings = {
             'features':list(set(self.features)),
             'signals' :list(set(self.signals)),
@@ -1547,19 +2091,21 @@ class stock_eda_panel(object):
                 'outlier': self.settings_outlier,
             }
         }
-        
+
         if model_type in ['Forecaster','Classifier']:
-            
+
             target_list = list(set(self.targets))
             target_list.sort()
             self.settings['model_type'] = model_type
             self.settings['target'] = list(set(self.target))
             self.settings['targets'] = target_list
-        
+
         ## for now this is hard coded
         feature_list = ['spread_ma','relative_spread_ma','pair_feature','count_features','bidirect_count_features','price_range','relative_price_range','rsi_feature',
                         'rsi_feature_v2', 'days_features','days_features_v2', 'volume_feature','smooth_volume', 'roc_feature', 'stoch_feature', 'stochastic_feature',
-                        'william_feature', 'vortex_feature', 'pair_index_feature','hmm']
+                        'william_feature', 'vortex_feature', 'pair_index_feature','hmm',
+                        'min_distance_pricefeature', 'min_relprice_pricefeature', 'max_distance_pricefeature','max_relprice_pricefeature'
+                        ]
 
         for feature in feature_list:
             try:
@@ -1570,7 +2116,7 @@ class stock_eda_panel(object):
             self.settings['settings']['target_lasts'] = self.settings_target_lasts
         except:
             pass
-        
+
         try:
             self.settings['settings']['strategies'] = {
                 'best_strategy':self.best_strategy,
@@ -1580,512 +2126,280 @@ class stock_eda_panel(object):
             pass
 
 class produce_model:
+    """
+    Class that produces a machine learning model in a scikit-learn pipeline wrapper.
+
+    Attributes
+    ----------
+    data  : pd.DataFrame
+        symbol of the asset
+    X_train : pd.DataFrame
+    y_train : pd.Series
+    X_test : pd.DataFrame
+    y_test : pd.Series
+    X_val : pd.DataFrame
+    y_val : pd.Series
+    pipeline : obj
+        trained pipeline that includes a ml model
+    features_to_model: list
+        features in end step of the pipeline
+
+    Methods
+    -------
+    preprocess(test_data_size=int, target=str, val_data_size=int):
+        prepare data, split train, test, validation data and X and Y
+    get_sample(x=pd.DataFrame, sample=int, max_=int):
+        sample data
+    """
     def __init__(self,data):
+        """
+        Initialize object
+
+        Parameters
+        ----------
+        data (pd.DataFrame): data
+
+        Returns
+        -------
+        None
+        """
         self.data = data.copy()
-    
+
     def preprocess(self, test_data_size, target, val_data_size = False):
-        
+        """
+        prepare data, split train, test, validation data and X and Y
+
+        Parameters
+        ----------
+        test_data_size (int): test data size
+        target (str): target column
+        val_data_size (int): validation data size
+
+        Returns
+        -------
+        None
+        """
         train_data, test_data = self.data.iloc[:-test_data_size,:].dropna() , self.data.iloc[-test_data_size:,:].dropna()
-        
+
         if val_data_size:
             train_data, val_data = train_data.iloc[:-val_data_size,:], train_data.iloc[-val_data_size:,:]
-            
+
         self.test_data = test_data
-        
+
         X_train, y_train = train_data.iloc[0:,1:], train_data[target]
         X_test, y_test = test_data.iloc[0:,1:], test_data[target]
         self.X_train = X_train
         self.y_train = y_train
         self.X_test = X_test
         self.y_test = y_test
-        
+
         if val_data_size:
             X_val, y_val = val_data.iloc[0:,1:], val_data[target]
             self.X_val = X_val
             self.y_val = y_val
-        
+
     def get_sample(self, x, sample, max_=900):
+        """
+        sample data
+
+        Parameters
+        ----------
+        x (pd.DataFrame): input data
+        sample (int): sample size
+        max_ (int): max sample
+
+        Returns
+        -------
+        sample (float): sample size
+        """
         length = len(x)
         if length > max_:
             return 1.0
         else:
             return sample
-    
+
     def train_model(self, pipe, model, cv_ = False):
+        """
+        train pipeline
+
+        Parameters
+        ----------
+        pipe (obj): pipeline object
+        model (obj): model object
+        cv_ (obj): cross validation procedure
+
+        Returns
+        -------
+        sample (float): sample size
+        """
         self.model = model
         self.pipe_transform = pipe
         self.pipeline = Pipeline([('pipe_transform',self.pipe_transform), ('model',self.model)])
-        self.features_to_model = self.pipe_transform.fit_transform(self.X_train).columns
         self.pipeline.fit(self.X_train, self.y_train)
-        
-        
-class hmm_feature_selector():
-    
-    def __init__(self, data, n_clusters, init_features_hmm, test_data_size, select_n_features, n_trials = 1,limit_search = False, default_benchmark_sd = 0.00003, t_threshold = 2):
-        self.data = data.copy()
-        self.n_clusters = n_clusters
-        self.init_features_hmm = init_features_hmm
-        self.test_data_size = test_data_size
-        self.select_n_features = select_n_features
-        self.n_trials = n_trials
-        self.limit_search= limit_search
-        self.default_benchmark_sd = default_benchmark_sd
-        self.t_threshold = t_threshold
-        
-    def split_data(self):
-        
-        self.data_train = self.data.iloc[:-self.test_data_size,:]
-        self.data_test = self.data.iloc[-self.test_data_size:,:]
-        
-    def train_model(self,features_hmm):
-        pipeline_hmm = Pipeline([
-                ('selector', FeatureSelector(columns=features_hmm)),
-                ('fillna', MeanMedianImputer(imputation_method='median',variables=features_hmm)),
-                ('hmm',GaussianHMM(n_components =  self.n_clusters, covariance_type = 'full'))
-                ])
-        
-        self.pipeline_hmm = pipeline_hmm.fit(self.data_train)
-        self.features_used_in_model = features_hmm
-        
-    def feature_list_generator(self):
-        
-        feature_combinations = set(list(combinations(self.init_features_hmm, self.select_n_features)))
-        feature_combinations = list(map(list, feature_combinations))
-        
-        self.feature_combinations = feature_combinations
-        
-    def get_error(self):
-        
-        self.data_train_ = self.data_train.copy()
-        
-        self.data_train_['hmm_feature'] = self.pipeline_hmm.predict(self.data_train_)
-        self.data_train_ = self.data_train_[['Date','hmm_feature','Close']].sort_values('Date')
-        
-        ## indexing chains
-        self.data_train_['lag_hmm_feature'] = self.data_train_['hmm_feature'].shift(1)
-        self.data_train_['breack'] = np.where(self.data_train_['lag_hmm_feature'] != self.data_train_['hmm_feature'],1,0)
-        self.data_train_["chain_id"] = self.data_train_.groupby("breack")["Date"].rank(method="first", ascending=True)
-        self.data_train_["chain_id"] = np.where(self.data_train_['breack'] == 1,self.data_train_["chain_id"],np.nan)
-        self.data_train_["chain_id"] = self.data_train_["chain_id"].fillna(method='ffill')
-        self.data_train_["hmm_chain_order"] = self.data_train_.groupby('chain_id')["Date"].rank(method="first", ascending=True)
-        
-        ### returns using the first element in a chain
-        self.data_train_['first'] = np.where(self.data_train_['hmm_chain_order'] == 1, self.data_train_['Close'], np.nan)
-        self.data_train_['first'] = self.data_train_.sort_values('Date')['first'].fillna(method='ffill')
-        self.data_train_['chain_return'] = (self.data_train_['Close']/self.data_train_['first'] -1) * 100
-        
-        self.data_train_ = self.data_train_.drop(columns = ['first'])
-        
-        mean_relevance, cluster_returns, number_relevant_states = states_relevance_score(self.data_train_)
-        self.mean_relevance = mean_relevance
-        
-    def execute_selector(self):
-        
-        self.split_data()
-        self.feature_list_generator()
-        maxi = -1
-        print(f'it is expected {len(self.feature_combinations)} combinations')
-        feature_results = dict()
-        
-        if self.limit_search:
-            print(f' taking just {self.limit_search} combinations')
-            maxi = self.limit_search
-            
-        for i,features_hmm in enumerate(self.feature_combinations[0:maxi]):
-    
-            feature_results[f'group_{i}'] = {
-                'features':list(features_hmm),
-                'relevances':list()
-            }
-            
-            for _ in range(self.n_trials):
-                try:
-                    self.train_model(features_hmm)
-                    self.get_error()
-                    feature_results[f'group_{i}']['relevances'].append(self.mean_relevance)
-                except:
-                    print('error')
-            feature_results[f'group_{i}']['mean relevance'] = np.mean(feature_results[f'group_{i}']['relevances'])
-        self.feature_results = feature_results
-        self.best_features = pd.DataFrame(self.feature_results).T.sort_values('mean relevance').iloc[-1,:].features
-        
-class signal_analyser_object:
-    
-    def __init__(self, data,symbol_name, show_plot = True, save_path = False, save_aws = False, aws_credentials = False, return_fig = False):
+        self.features_to_model = self.pipeline[:-1].transform(self.X_train).columns
+
+class analyse_index(stock_eda_panel):
+    """
+    class that is going to train hmm models to perform feature selection
+
+    Attributes
+    ----------
+    data_index : pd.DataFrame
+         name of the index
+    indexes: list
+        list of indexes
+    asset : str
+         name of the asset
+    n_obs : int
+         number of rows to extract
+    lag : int
+         lag to apply
+    data_window : str
+         5y 10y 15y
+    show_plot : bool
+         If True, show plots
+    save_path : str
+         local path for saving e.g r'C:/path/to/the/file/'
+    save_aws : str
+         remote key in s3 bucket path e.g. 'path/to/file/'
+    aws_credentials : dict
+         dict with the aws credentials
+    merger_df : pd.DataFrame
+        dataframe with the index and asset data
+    states_result = dict
+        betas and correlation score results
+
+    Methods
+    -------
+    process_data():
+        using stock_eda_panel, get data and merge data
+    plot_betas(sample_size=int, offset=int, subsample_ts=int):
+        display beta analysis plot
+    get_betas(subsample_ts=int)
+        get general beta and last sample beta, correlation score is included too
+    """
+    def __init__(self, index_data, asset, n_obs, lag, data_window = '5y', show_plot = False, save_path = False, save_aws = False, aws_credentials = False, return_fig = False):
         """
-        data: pandas df
-        symbol_name: str name of the asset
-        show_plot: bool
-        save_path: str local path for saving e.g r'C:/path/to/the/file/'
-        save_aws: str remote key in s3 bucket path e.g. 'path/to/file/'
-        aws_credentials: dict
-        return_fig: boolean return the image function as result
+        Initialize object
+
+        Parameters
+        ----------
+        index_data (pd.DataFrame or str): index data dataframe or index string
+        asset (str): name of the asset
+        n_obs (int): number of rows to extract
+        lag (int): lag to apply
+        data_window (str): 5y 10y 15y
+        show_plot (bool): If True, show plots
+        save_path (str): local path for saving e.g r'C:/path/to/the/file/'
+        save_aws (str): remote key in s3 bucket path e.g. 'path/to/file/'
+        aws_credentials (dict): dict with the aws credentials
+
+        Returns
+        -------
+        None
         """
-        self.data = data.copy()
-        self.ticket_name = symbol_name
-        self.show_plot = show_plot
-        self.save_path = save_path
-        self.save_aws = save_aws
-        self.aws_credentials = aws_credentials
-        self.return_fig = return_fig
-
-    def signal_analyser(self, test_size, feature_name, days_list, threshold = 0.05,verbose = False, signal_position = False):
-        data = self.data
-        self.feature_name = feature_name
-        up_signal, low_signal= f'signal_up_{feature_name}', f'signal_low_{feature_name}'
-        features_base = ['Date', up_signal, low_signal, 'Close']
-
-        df = data[features_base].sort_values('Date').iloc[0:-test_size,:]
-        returns_list = list()
-
-        for days in days_list:
-
-            feature_ = f'return_{days}d'
-            df[feature_] = (df['Close'].shift(-days)/df['Close']-1)*100
-            returns_list.append(feature_)
 
-        df['signal_type'] = np.where(
-            df[up_signal] == 1, 
-            'up', 
-            np.where(
-                df[low_signal] == 1, 
-                'down',
-                None
-            )
-        )
-        df = df[~df.signal_type.isna()]
-        # df['Date'] = df.index
-        df['lag_Date'] = df['Date'].shift(1)
-        df['span'] = (pd.to_datetime(df['Date']) - pd.to_datetime(df['lag_Date'])).dt.days - 1
-        df['break'] = np.where(df['span'] > 3, 1, 0)
-        df['break'] = np.where(df['span'].isna(), 1, df['break'])
-
-        df['chain_id'] = df.sort_values(['Date']).groupby(['break']).cumcount() + 1
-        df['chain_id'] = np.where(df['break'] == 1, df['chain_id'], np.nan )
-        df['chain_id'] = df['chain_id'].fillna(method = 'ffill')
-
-        df['internal_rn'] = df.sort_values(['Date']).groupby(['chain_id']).cumcount() + 1
-        df['inv_internal_rn'] = df.sort_values(['Date'],ascending = False).groupby(['chain_id']).cumcount() + 1
-
-        df['first_in_chain'] = np.where(df['internal_rn'] == 1, True, False)
-        df['last_in_chain'] = np.where(df['inv_internal_rn'] == 1, True, False)
-
-        df = df.drop(columns = ['break','span','lag_Date','inv_internal_rn']).sort_values('Date')
-        self.df_signal = df
         
-        n_signals_up = len(list(df[df.signal_type == 'up'].chain_id.unique()))
-        n_signals_down = len(list(df[df.signal_type == 'down'].chain_id.unique()))
-        p_scores = list()
-        medians_down = list()
-        validations = list()
-        if not signal_position: ### for now it is based on the last signal on a chain
-            df_melt = df[df.last_in_chain == True].melt(id_vars=['signal_type'], value_vars=returns_list, var_name='time', value_name='value')
-            df_melt = df_melt.dropna()
-
-        for evalx in returns_list:
-
-            sample1 = df_melt[(df_melt.time == evalx) & (df_melt.signal_type == 'up')].value.values
-            sample2 = df_melt[(df_melt.time == evalx) & (df_melt.signal_type == 'down')].value.values
-            pvalue = stats.ttest_ind(sample1, sample2).pvalue
-            median_down = np.median(sample2)
-            median_up = np.median(sample1) 
-            validations.append(median_up < 0)
-            validations.append(median_down > 0)
-            p_scores.append(pvalue)
-            medians_down.append(median_down)
-        self.df_melt = df_melt
-        null_ho_eval = threshold > np.mean(p_scores)
-        mean_median_return = np.median(medians_down)  ## end metric
-        median_signal_type_eval = validations.count(validations[0]) == len(validations)
-
-        if verbose:
-            print('number of signal up:',n_signals_up)
-            print('number of signal down:',n_signals_down)
-            print('reject ho: ', null_ho_eval)
-            print('mean median:', mean_median_return)
-            print('all validations: ', median_signal_type_eval)
-
-        # if median_signal_type_eval == True and null_ho_eval == True:
-        if null_ho_eval == True:
-            if verbose:
-                print('success evals')
-            self.mean_median_return = mean_median_return
+        if type(index_data) != str:
+            index_data['Date'] = pd.to_datetime(index_data['Date'])
+            self.index_data = index_data
+            self.indexes = [ x for x in list(index_data.columns) if x != 'Date']
         else:
-            self.mean_median_return = np.nan
-
-        df2 = df.copy()
-        df2 = df2[df2.last_in_chain == True]
-
-
-        df2['lagdate'] = df2.Date.shift(1)
-        df2['span'] = (pd.to_datetime(df2['Date']) - pd.to_datetime(df2['lagdate'])).dt.days
-
-        fig, axs = plt.subplots(1, 3, figsize = (15,5))
-
-        sns.boxplot(data=df2, y="span",ax = axs[0])
-        axs[0].set_title('span between last signals')
-        del df2
-        sns.boxplot(data=df[df.last_in_chain == True], y="internal_rn",ax = axs[1])
-        axs[1].set_title('signal duration distribution')
-        sns.boxplot(data=df_melt, x="time", y="value", hue="signal_type",ax = axs[2])
-        axs[2].axhline(y=0, color='grey', linestyle='--')
-        axs[2].set_title('signal type expected returns distribution at different time lapses')
-            
-        if self.show_plot:
-            plt.show()
+            self.indexes = [index_data]
             
-        if self.save_path:
-            result_plot_name = f'signals_strategy_distribution_{feature_name}.png'
-            fig.savefig(self.save_path+result_plot_name)
-            # pickle.dump(axs, open(self.save_path+result_plot_name, 'wb'))
-
-        if self.save_path and self.save_aws:
-            # upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = f'market_plots/{self.ticket_name}/'+result_plot_name, input_path = self.save_path+result_plot_name)
-            upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = self.save_aws + result_plot_name, input_path = self.save_path + result_plot_name, aws_credentials = self.aws_credentials)
-        if not self.show_plot:
-            plt.close()
-
-        del df
-
-        if self.return_fig:
-            return fig
-        
-    def create_backtest_signal(self,days_strategy, test_size, feature_name, high_exit = False, low_exit = False):
-        asset_1 = 'Close'
-        up_signal, low_signal= f'signal_up_{feature_name}', f'signal_low_{feature_name}'
-        df1 = self.data.iloc[-test_size:,:].copy()
-        df2 = df1.copy()
-        df2['signal_type'] = np.where(
-                    df2[up_signal] == 1, 
-                    'up', 
-                    np.where(
-                        df2[low_signal] == 1, 
-                        'down',
-                        None
-                    )
-                )
-        df2 = df2[~df2.signal_type.isna()]
-        # df2['Date_'] = df2.index
-        df2['lag_Date'] = df2['Date'].shift(1)
-        df2['span'] = (pd.to_datetime(df2['Date']) - pd.to_datetime(df2['lag_Date'])).dt.days - 1
-        df2['break'] = np.where(df2['span'] > 3, 1, 0)
-        df2['break'] = np.where(df2['span'].isna(), 1, df2['break'])
-    
-        df2['chain_id'] = df2.sort_values(['Date']).groupby(['break']).cumcount() + 1
-        df2['chain_id'] = np.where(df2['break'] == 1, df2['chain_id'], np.nan )
-        df2['chain_id'] = df2['chain_id'].fillna(method = 'ffill')
-    
-        df2['internal_rn'] = df2.sort_values(['Date']).groupby(['chain_id']).cumcount() + 1
-        df2['inv_internal_rn'] = df2.sort_values(['Date'],ascending = False).groupby(['chain_id']).cumcount() + 1
-    
-        df2['first_in_chain'] = np.where(df2['internal_rn'] == 1, True, False)
-        df2['last_in_chain'] = np.where(df2['inv_internal_rn'] == 1, True, False)
-    
-        df2 = df2.drop(columns = ['break','span','lag_Date','inv_internal_rn']).sort_values('Date')
-    
-        df2 = df2[(df2.last_in_chain == True) & (df2.signal_type == 'down')][['last_in_chain']]
-        dft = df1.merge(df2,how = 'left',left_index=True, right_index=True )
-    
-        dft['chain_id'] = dft.sort_values(['Date']).groupby(['last_in_chain']).cumcount() + 1
-        dft['chain_id'] = np.where(dft['last_in_chain'] == True, dft['chain_id'], np.nan )
-        dft['chain_id'] = dft['chain_id'].fillna(method = 'ffill')
-    
-        dft['internal_rn'] = dft.sort_values(['Date']).groupby(['chain_id']).cumcount() + 1
-        dft['flag'] = np.where(dft['internal_rn'] < days_strategy, 1,0)
-        
-        dft['lrets_bench'] = np.log(dft[asset_1]/dft[asset_1].shift(1))
-        dft['bench_prod'] = dft['lrets_bench'].cumsum()
-        dft['bench_prod_exp'] = np.exp(dft['bench_prod']) - 1
-        
-        if high_exit and low_exit:
-            dft['open_strat'] = np.where(dft.last_in_chain == True, dft.Open, np.nan)
-            dft['open_strat'] = dft['open_strat'].fillna(method = 'ffill')
-            dft['open_strat'] = np.where(dft.flag == 1, dft.open_strat, np.nan)
-            dft['high_strat_ret'] = (dft['High']/dft['open_strat']-1)*100
-            dft['low_strat_ret'] = (dft['Low']/dft['open_strat']-1)*100
-            dft['high_exit'] =  np.where(((dft['high_strat_ret'] >= high_exit) | (dft['internal_rn'] == days_strategy)), 1, np.nan)
-            dft['low_exit'] =  np.where((dft['low_strat_ret'] <= low_exit), -1, np.nan)
-            
-            dft["exit_type"] = dft[["high_exit", "low_exit"]].max(axis=1)
-            dft['exit_type'] = np.where(dft["exit_type"] == 1, 1, np.where(dft["exit_type"] == -1,-1,np.nan))
-            dft['exit'] = np.where(dft['exit_type'].isnull(), np.nan, 1)
-            dft['exit_order'] = dft.sort_values(['Date']).groupby(['chain_id','exit']).cumcount() + 1
-            dft['exit'] = np.where(dft['exit_order'] == 1, True, np.nan)
-            dft = dft.drop(columns = ['exit_order'])
-            ## if last signal is near
-            max_id = dft.chain_id.max()
-            dft['max_internal_rn'] = dft.sort_values(['Date']).groupby(['chain_id']).internal_rn.transform('max')
-            dft['exit'] = np.where((dft.chain_id == max_id) & (dft.max_internal_rn < days_strategy) & (dft.max_internal_rn == dft.internal_rn), 1, dft['exit'])
-            
-            dft['exit_step'] = np.where(dft.exit == 1, dft.internal_rn, np.nan)
-            dft['exit_step'] = dft.sort_values(['Date']).groupby(['chain_id']).exit_step.transform('max')
-            
-            dft['flag'] = np.where(dft.internal_rn <= dft.exit_step, 1, 0)
-            dft = dft.drop(columns = ['open_strat', 'high_strat_ret', 'low_strat_ret','exit_step', 'exit','exit_type','high_exit','low_exit', 'max_internal_rn'])
-        
-        dft['lrets_strat'] = np.log(dft[asset_1].shift(-1)/dft[asset_1]) * dft['flag']
-        dft['lrets_strat'] = np.where(dft['lrets_strat'].isna(),-0.0,dft['lrets_strat'])
-        dft['lrets_prod'] = dft['lrets_strat'].cumsum()
-        dft['strat_prod_exp'] = np.exp(dft['lrets_prod']) - 1
-    
-        bench_rets = round(dft['bench_prod_exp'].values[-1]*100,1)
-        strat_rets = round(dft['strat_prod_exp'].values[-1]*100,1)
-        
-        bench_sr = round(sharpe_ratio(dft.bench_prod_exp.dropna()),1)
-        strat_sr = round(sharpe_ratio(dft.strat_prod_exp.dropna()),1)
-        
-        message1 = f'{bench_rets}%'
-        message2 = f'{strat_rets}%'
-        
-        messages = {
-            'benchmark return:':message1,
-            'benchmark sharpe ratio:': bench_sr,
-            'strategy return:':message2,
-            'strategy sharpe ratio:': strat_sr,
-        }
-        if self.show_plot:
-            print('----------------------------')
-            print(messages)
-            print('----------------------------')
-            
-        fig = plt.figure(1)
-        plt.plot(dft.bench_prod_exp.values, label = 'benchmark')
-        plt.scatter(range(len(dft)),np.where(dft[low_signal] == 1,dft.bench_prod_exp.values,np.nan),color = 'red', label = 'signal')
-        plt.plot(dft.strat_prod_exp.values, label = 'strategy')
-        plt.legend()
-        plt.title('strategy and cumulative returns based on signal strategy')
-        if self.show_plot:
-            plt.plot()
-            
-        if self.save_path:
-            result_json_name = f'signals_strategy_return_{feature_name}.json'
-            result_plot_name = f'signals_strategy_return_{feature_name}.png'
-            
-            plt.savefig(self.save_path+result_plot_name)
-            # pickle.dump(fig, open(self.save_path+result_plot_name, 'wb'))
-            
-            with open(self.save_path+result_json_name, "w") as outfile: 
-                json.dump(messages, outfile)
-                
-        if self.save_path and self.save_aws:
-            # upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = f'market_plots/{self.ticket_name}/'+result_json_name ,input_path = self.save_path+result_json_name)
-            # upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = f'market_plots/{self.ticket_name}/'+result_plot_name,input_path = self.save_path+result_plot_name)
-            
-            upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = self.save_aws + result_json_name, input_path = self.save_path + result_json_name, aws_credentials = self.aws_credentials)
-            upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = self.save_aws + result_plot_name, input_path = self.save_path + result_plot_name, aws_credentials = self.aws_credentials)
-    
-        if not self.show_plot:
-            plt.close()
-            
-        del df1,df2,dft
-    
-        if self.return_fig:
-            return fig, messages
-        
-def execute_signal_analyser(test_data_size, feature_name, days_list, configuration, method, object_stock, signal_analyser_object, plot = False, backtest= False, exit_params = {}):
-     
-    method(**configuration)
-    signal_assess = signal_analyser_object(object_stock.df,object_stock.stock_code,show_plot = plot)
-    signal_assess.signal_analyser(test_size = test_data_size, feature_name = feature_name, days_list = days_list, threshold = 1)
-
-    if backtest:
-        print('-----------------------back test ---------------------------')
-        signal_assess.create_backtest_signal(backtest, test_data_size, feature_name, **exit_params )
-    
-    return signal_assess.mean_median_return
-
-def iterate_signal_analyser(test_data_size,feature_name, days_list, arguments_to_test, method, object_stock, signal_analyser_object, plot = True):
-
-    results = list()
-    for key in arguments_to_test.keys():
-        configuration = arguments_to_test.get(key)
-        mean_median_return = execute_signal_analyser(test_data_size, feature_name, days_list, configuration, method, object_stock, signal_analyser_object)
-        results.append(mean_median_return)
-    
-    df_result = pd.DataFrame({'keys':arguments_to_test.keys(),'results':results})
-    if plot:
-        plt.plot(df_result['keys'], df_result['results'])
-        plt.scatter(df_result['keys'], df_result['results'])
-        plt.title('simulation between configurations')
-        plt.ylabel('median expected return') 
-        plt.show()
-        
-    best_result = df_result.sort_values('results',ascending = False)['keys'].values[0]
-    return best_result
-    
-class analyse_index(stock_eda_panel):
-    def __init__(self, index, asset, n_obs, lag, data_window = '5y', show_plot = True, save_path = False, save_aws = False, aws_credentials = False):
-
-        """
-        data: pandas df
-        index: str name of the index
-        asset: str name of the asset
-        n_obs: int
-        lag: int
-        data_window: str eg 5y 10y 15y
-        show_plot: bool
-        save_path: str local path for saving e.g r'C:/path/to/the/file/'
-        save_aws: str remote key in s3 bucket path e.g. 'path/to/file/'
-        aws_credentials: dict
-        """
-         
-        self.index = index
+        self.index_data = index_data
         self.asset = asset
         self.n_obs = n_obs
         self.data_window = data_window
         self.lag = lag
-        
+
         self.show_plot = show_plot
+        self.return_fig = return_fig
         self.save_path = save_path
         self.save_aws = save_aws
-        
-    def process_data(self):
-        
-        index = stock_eda_panel(self.index, self.n_obs, self.data_window)
-        index.get_data()
-        index.df['shift'] = index.df.Close.shift(self.lag)
-        index.df['index_return'] = index.df.Close/index.df['shift'] - 1
 
-        asset =  stock_eda_panel(self.asset, self.n_obs, self.data_window)
+    def process_data(self):
+        """
+        using stock_eda_panel, get data and merge data
+    
+        Parameters
+        ----------
+        None
+    
+        Returns
+        -------
+        None
+        """
+        asset =  stock_eda_panel(self.asset, self.n_obs, data_window=self.data_window)
         asset.get_data()
-        asset.df['shift'] = asset.df.Close.shift(self.lag)
-        asset.df['asset_return'] = asset.df.Close/asset.df['shift'] - 1
+        df = asset.df[['Date','Close']]
         
-        df1 = index.df[['Date','index_return']]
-        df2 = asset.df[['Date','asset_return','Close']]
-        merger = df1.merge(df2, on = 'Date', how = 'inner')
-        merger.dropna(inplace = True)
-        self.merger_df = merger
-        
-    def plot_betas(self,sample_size, offset, subsample_ts =False):
-    
-        ### extracting data
+        if type(self.index_data) != str:
+            df_merge = df.merge(self.index_data, on = ['Date'], how = 'left').sort_values('Date')
+            
+        else:
+            indx =  stock_eda_panel(self.index_data, self.n_obs, data_window=self.data_window)
+            indx.get_data()
+            indx_df = indx.df[['Date','Close']].rename(columns = {'Close':self.index_data})
+            df_merge = df.merge(indx_df, on = ['Date'], how = 'left').sort_values('Date')
+            
+        for colx in ['Close'] + self.indexes:
+            df_merge[f'{colx}_pct'] = df_merge[colx]/df_merge[colx].shift(self.lag) - 1
+            
+        df_merge.dropna(inplace = True)
+        self.merger_df = df_merge.rename(columns = {'Close_pct': 'asset_return'})
 
-        self.process_data()
-        
-         ### ploting analysis
+    def plot_betas(self,sample_size, offset, subsample_ts =False, index = False):
+        """
+        display beta analysis plot
+
+        Parameters
+        ----------
+        sample_size (int): number of days or window size to calculate beta
+        offset (int): overlap between windows
+        subsample_ts (int): subsample size of data 
+
+        Returns
+        -------
+        None
+        """
+        if (type(self.index_data) == str) & (index != False):
+            raise Exception("No need of index argument")
+        else:
+            index = self.indexes[0]
+            
+        index_pct = f'{index}_pct'
+        ### ploting analysis
         figure, ax = plt.subplot_mosaic(
             [["scatter_total", "scatter_sample",'ts','ts']],
             layout="constrained",
             figsize=(18, 5)
         )
-        
-        ax['scatter_total'].scatter(self.merger_df.asset_return, self.merger_df.index_return)
-        b, a = np.polyfit(self.merger_df.asset_return, self.merger_df.index_return, 1)
+
+        ax['scatter_total'].scatter(self.merger_df.asset_return, self.merger_df[index_pct])
+    
+        huber_regr = HuberRegressor(fit_intercept = True)
+        huber_regr.fit(self.merger_df.asset_return.values.reshape(-1,1), self.merger_df[index_pct].values.reshape(-1,1))
+        b, a = huber_regr.coef_[0], huber_regr.intercept_
+    
+        # b, a = np.polyfit(self.merger_df.asset_return, self.merger_df[index_pct], 1)
         ax['scatter_total'].plot(self.merger_df.asset_return, b*self.merger_df.asset_return+a, color='red')
 
         ax['ts'].plot(self.merger_df.Date, self.merger_df.Close, color = 'grey', alpha = 0.3)
-        
+
         if subsample_ts:
             self.merger_df = self.merger_df.iloc[-subsample_ts:,:].dropna()
-        
+
         for i in range(0,len(self.merger_df)-sample_size,offset):
 
             merger_ = self.merger_df.sort_values('Date', ascending = False).iloc[i:i+sample_size,:]
-            x = merger_.index_return 
+            x = merger_[index_pct]
             y = merger_.asset_return
-            b, a = np.polyfit(x,y, 1)
-
+            # b, a = np.polyfit(x,y, 1)
+            huber_regr = HuberRegressor(fit_intercept = True)
+            huber_regr.fit(x.values.reshape(-1,1), y.values.reshape(-1,1))
+            b, a = huber_regr.coef_[0], huber_regr.intercept_
+            
             normalize = mcolors.Normalize(vmin=-1, vmax=1)
             colormap = cm.jet
 
@@ -2098,12 +2412,13 @@ class analyse_index(stock_eda_panel):
 
         scalarmappaple = cm.ScalarMappable(norm=normalize, cmap=colormap)
         scalarmappaple.set_array(x)
-        
-        plt.title(f'{self.asset} using index: {self.index}')
+
+        plt.title(f'{self.asset} using index: {index}')
         plt.colorbar(scalarmappaple)
-        
+
         if self.show_plot:
             plt.show()
+            
         if self.save_path:
             result_plot_name = f'market_best_fit.png'
             figure.savefig(self.save_path+result_plot_name)
@@ -2111,80 +2426,50 @@ class analyse_index(stock_eda_panel):
         if self.save_path and self.save_aws:
             # upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = f'market_plots/{self.asset}/'+result_plot_name,input_path = self.save_path+result_plot_name)
             upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = self.save_aws + result_plot_name, input_path = self.save_path + result_plot_name, aws_credentials = self.aws_credentials)
+            
         if not self.show_plot:
-            plt.close()
-        
+            plt.close() 
+            
+        if self.return_fig:
+            return figure
+            
     def get_betas(self,subsample_ts=False):
-        
-        self.process_data()
-        general_beta, a = np.polyfit(self.merger_df.asset_return, self.merger_df.index_return, 1)
-        general_r = stats.mstats.pearsonr(self.merger_df.asset_return, self.merger_df.index_return)[0]
-        
-        self.process_data()
-        if subsample_ts:
-            self.merger_df = self.merger_df.iloc[-subsample_ts:,:].dropna()
-        sample_beta, a = np.polyfit(self.merger_df.asset_return, self.merger_df.index_return, 1)
-        sample_r = stats.mstats.pearsonr(self.merger_df.asset_return, self.merger_df.index_return)[0]
-        
-        result = {
-            'general_beta':general_beta,
-            'general_r':general_r,
-            'sample_beta':sample_beta,
-            'sample_r':sample_r
-        }
-        
-        self.states_result = result
-        
-class evaluate_markets(analyse_index):
-    def __init__(self, stock_code, indexes):
-        self.stock_code = stock_code
-        self.indexes = indexes
-    def evaluate_best_market_fit(self,sample_size, offset,lag= 3, n_obs = 3500, verbose = False, plot_best = False):
-
-        results_dicts = dict()
+        """
+        get general beta and last sample beta, correlation score is included too
+    
+        Parameters
+        ----------
+        subsample_ts (int): subsample size of data 
+    
+        Returns
+        -------
+        None
+        """
+        result = list()
         for index in self.indexes:
-            betex = analyse_index(index = index,asset = self.stock_code,n_obs = n_obs, lag = lag)
-            betex.get_betas(sample_size)
-            results_dicts[index] = betex.states_result
-        pd_result = pd.DataFrame(results_dicts).T
-        pd_result['gen_r2'] = pd_result.general_r ** 2
-        pd_result['sampl_r2'] = pd_result.sample_r ** 2
-        self.stat_results = pd_result
-        
-        best_result = pd_result.sort_values('gen_r2',ascending = False).head(2).sort_values('sampl_r2',ascending = False).head(1)
-        best_fit_index = best_result.index.values[0]
-        
-        self.stat_results = self.stat_results.drop(columns = ['gen_r2','sampl_r2'])
-        
-        if verbose:
-            print(best_result)
-        if plot_best:
-            betex = analyse_index(index = best_fit_index,asset = self.stock_code, n_obs = n_obs, lag = lag)
-            betex.plot_betas(sample_size = sample_size, offset = offset, subsample_ts = False)
             
-        self.best_result = best_result
-        
-def get_relevant_beta(data_market, ticket_name,  show_plot = True, save_path = False, save_aws = False, aws_credentials = False):
-    """
-        data_market: pandas df
-        ticket_name: str name of the asset
-        show_plot: bool
-        save_path: str local path for saving e.g r'C:/path/to/the/file/'
-        save_aws: str remote key in s3 bucket path e.g. 'path/to/file/'
-        aws_credentials: dict
-        """
-    all_betas = data_market[data_market.asset == ticket_name].sort_values('general_r', ascending = False)
-    all_betas['gen_r2'] = all_betas.general_r ** 2
-    all_betas['sampl_r2'] = all_betas.sample_r ** 2
-    selection = all_betas.sort_values('gen_r2',ascending =False).head(2).sort_values('sampl_r2',ascending =False).head(1).drop(columns = ['gen_r2','sampl_r2'])
+            index_pct = f'{index}_pct'
+            huber_regr = HuberRegressor(fit_intercept = True)
+            huber_regr.fit(self.merger_df.asset_return.values.reshape(-1,1), self.merger_df[index_pct].values.reshape(-1,1))
+            general_beta, a = huber_regr.coef_[0], huber_regr.intercept_
+            general_r = stats.mstats.pearsonr(self.merger_df.asset_return, self.merger_df[index])[0]
+
+            dict_res = {
+                    'index':index,
+                    'general_beta':general_beta,
+                    'general_r':general_r,
+                }
     
-    if show_plot:
-        print(selection)
-    if save_path:
-        result_plot_name = f'market_best_fit.csv'
-        selection.to_csv(save_path+result_plot_name)
-
-    if save_path and save_aws:
-        # upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = f'market_plots/{ticket_name}/'+result_plot_name,input_path = save_path+result_plot_name)
-        upload_file_to_aws(bucket = 'VIRGO_BUCKET', key = save_aws + result_plot_name, input_path = save_path + result_plot_name, aws_credentials = aws_credentials)
-    return selection
+            if subsample_ts:
+                tmp_df = self.merger_df.iloc[-subsample_ts:,:].dropna()
+                huber_regr = HuberRegressor(fit_intercept = True)
+                huber_regr.fit(tmp_df.asset_return.values.reshape(-1,1), tmp_df[index_pct].values.reshape(-1,1))
+                sample_beta, a = huber_regr.coef_[0], huber_regr.intercept_
+                sample_r = stats.mstats.pearsonr(tmp_df.asset_return, tmp_df[index])[0]
+                dict_res['sample_beta'] = sample_beta
+                dict_res['sample_r'] = sample_r
+                
+            result.append(dict_res)
+    
+        self.states_result = result
+