Merge pull request #47 from hudson-and-thames/develop

Update version to 0.4.0

arbitragelab/codependence/codependence_matrix.py

@@ -15,6 +15,7 @@
 from arbitragelab.codependence.correlation import distance_correlation
 from arbitragelab.codependence.gnpr_distance import spearmans_rho, gpr_distance, gnpr_distance
 from arbitragelab.codependence.optimal_transport import optimal_transport_dependence
+from arbitragelab.util import devadarsh
 
 
 # pylint: disable=invalid-name
@@ -49,6 +50,9 @@ def get_dependence_matrix(df: pd.DataFrame, dependence_method: str, theta: float
                                   Sets the relative area of correlation in a copula. [from 0 to 1] (0.2 by default)
     :return: (pd.DataFrame) Dependence matrix.
     """
+
+    devadarsh.track('get_dependence_matrix')
+
     # Get the feature names.
     features_cols = df.columns.values
     n = df.shape[1]
@@ -110,6 +114,9 @@ def get_distance_matrix(X: pd.DataFrame, distance_metric: str = 'angular') -> pd
     :param distance_metric: (str) The distance metric to be used for generating the distance matrix.
     :return: (pd.DataFrame) Distance matrix.
     """
+
+    devadarsh.track('get_distance_matrix')
+
     if distance_metric == 'angular':
         distfun = lambda x: ((1 - x).round(5) / 2.) ** .5
     elif distance_metric == 'abs_angular':

arbitragelab/cointegration_approach/coint_sim.py

@@ -41,6 +41,7 @@ def __init__(self, ts_num: int, ts_length: int):
         self.ts_num = ts_num
         self.ts_length = ts_length
         self.__price_params, self.__coint_params = self.initialize_params()
+
         devadarsh.track('CointegrationSimulation')
 
     @staticmethod

arbitragelab/cointegration_approach/multi_coint.py

@@ -15,6 +15,7 @@
 import pandas as pd
 
 from arbitragelab.cointegration_approach.johansen import JohansenPortfolio
+from arbitragelab.util import devadarsh
 
 
 class MultivariateCointegration:
@@ -38,6 +39,8 @@ def __init__(self, asset_df: pd.DataFrame, trade_df: pd.DataFrame):
         self.__trade_df = trade_df
         self.__coint_vec = None
 
+        devadarsh.track('MultivariateCointegration')
+
     @staticmethod
     def _missing_impute(price_df: pd.DataFrame, nan_method: str = 'ffill', order: int = 3) -> pd.DataFrame:
         """

arbitragelab/cointegration_approach/signals.py

@@ -10,6 +10,7 @@
 import pandas as pd
 from sklearn.linear_model import LinearRegression
 
+from arbitragelab.util import devadarsh
 
 def get_half_life_of_mean_reversion(data: pd.Series) -> float:
     """
@@ -19,6 +20,8 @@ def get_half_life_of_mean_reversion(data: pd.Series) -> float:
     :return: (float) Half-life of mean reversion
     """
 
+    devadarsh.track('get_half_life_of_mean_reversion')
+
     reg = LinearRegression(fit_intercept=True)
 
     training_data = data.shift(1).dropna().values.reshape(-1, 1)
@@ -46,6 +49,8 @@ def linear_trading_strategy(portfolio_series: pd.Series, sma_window: int = None,
     :return: (pd.DataFrame) Mean-reverting portfolio series and target allocation on each day.
     """
 
+    devadarsh.track('linear_trading_strategy')
+
     if sma_window is None:
         # The book suggests to use window = speed of reversion
         sma_window = int(get_half_life_of_mean_reversion(portfolio_series))
@@ -83,6 +88,8 @@ def bollinger_bands_trading_strategy(portfolio_series: pd.Series, sma_window: in
     :return: (pd.DataFrame) Mean-reverting portfolio series and target allocation on each day.
     """
 
+    devadarsh.track('bollinger_bands_trading_strategy')
+
     if exit_z_score >= entry_z_score:
         raise ValueError('Exit Z-score can not be bigger than entry Z-Score.')
 

arbitragelab/cointegration_approach/sparse_mr_portfolio.py

@@ -31,6 +31,7 @@
 from sklearn.preprocessing import normalize, StandardScaler
 
 from arbitragelab.optimal_mean_reversion import OrnsteinUhlenbeck
+from arbitragelab.util import devadarsh
 
 
 class SparseMeanReversionPortfolio:
@@ -58,6 +59,8 @@ def __init__(self, assets: pd.DataFrame):
         self._standardized = standard_data
         self._standardized.columns = self._assets.columns
 
+        devadarsh.track('SparseMeanReversionPortfolio')
+
     @property
     def assets(self) -> pd.DataFrame:
         """

arbitragelab/copula_approach/__init__.py

@@ -4,10 +4,13 @@
 
 from arbitragelab.copula_approach.copula_calculation import (
     find_marginal_cdf, ml_theta_hat, log_ml, sic, aic, hqic, construct_ecdf_lin, fit_nu_for_t_copula, scad_penalty,
-    scad_derivative, adjust_weights)
+    scad_derivative, adjust_weights, to_quantile)
 from arbitragelab.copula_approach.copula_generate import (
     Copula, Gumbel, Frank, Clayton, Joe, N13, N14, Gaussian, Student, Switcher)
 from arbitragelab.copula_approach.copula_strategy import CopulaStrategy
 from arbitragelab.copula_approach.copula_strategy_basic import BasicCopulaStrategy
 from arbitragelab.copula_approach.copula_strategy_mpi import CopulaStrategyMPI
 from arbitragelab.copula_approach.copula_generate_mixedcopula import (MixedCopula, CFGMixCop, CTGMixCop)
+from arbitragelab.copula_approach.vine_copula_partner_selection import PartnerSelection
+from arbitragelab.copula_approach.vinecop_generate import (RVineCop, CVineCop)
+from arbitragelab.copula_approach.vinecop_strategy import CVineCopStrat

arbitragelab/copula_approach/copula_calculation.py

@@ -23,8 +23,9 @@
 <https://www.tandfonline.com/doi/pdf/10.1080/01621459.2013.873366?casa_token=sey8HrojSgYAAAAA:TEMBX8wLYdGFGyM78UXSYm6hXl1Qp_K6wiLgRJf6kPcqW4dYT8z3oA3I_odrAL48DNr3OSoqkQsEmQ>`__
 """
 # pylint: disable = invalid-name
-from typing import Callable
+from typing import Callable, Tuple
 import numpy as np
+import pandas as pd
 import scipy.stats as ss
 from scipy.interpolate import interp1d
 from scipy.optimize import minimize
@@ -102,6 +103,36 @@ def bounded_ecdf(x):
     # Vectorize it to work with arrays.
     return np.vectorize(bounded_ecdf)
 
+def to_quantile(data: pd.DataFrame) -> Tuple[pd.DataFrame, list]:
+    """
+    Convert the data frame to quantile by row.
+
+    Not in place. Also returns the marginal cdfs of each column. This can work with more than just 2 columns.
+
+    The method returns:
+
+        - quantile_data: (pd.DataFrame) The calculated quantile data in a data frame with the original indexing.
+        - cdf_list: (list) The list of marginal cumulative density functions.
+
+    :param data: (pd.DataFrame) The original data in DataFrame.
+    :return: (tuple)
+        quantile_data: (pd.DataFrame) The calculated quantile data in a data frame with the original indexing.
+        cdf_list: (list) The list of marginal cumulative density functions.
+    """
+
+    column_count = len(data.columns)  # Number of columns.
+    cdf_lst = [None] * column_count  # List to store all marginal cdf functions.
+    quantile_data_lst = [None] * column_count  # List to store all quantile data in pd.Series.
+
+    # Loop through all columns.
+    for i in range(column_count):
+        cdf_lst[i] = construct_ecdf_lin(data.iloc[:, i])
+        quantile_data_lst[i] = data.iloc[:, i].map(cdf_lst[i])
+
+    quantile_data = pd.concat(quantile_data_lst, axis=1)  # Form the quantile DataFrame.
+
+    return quantile_data, cdf_lst
+
 def ml_theta_hat(x: np.array, y: np.array, copula_name: str) -> float:
     """
     Calculate empirical theta (theta_hat) for a type of copula by pseudo-maximum likelihood.

arbitragelab/copula_approach/copula_generate_mixedcopula.py

@@ -21,8 +21,10 @@
 import numpy as np
 import pandas as pd
 import matplotlib.pyplot as plt
+
 import arbitragelab.copula_approach.copula_generate as cg
 import arbitragelab.copula_approach.copula_calculation as ccalc
+from arbitragelab.util import devadarsh
 
 
 class MixedCopula(ABC):
@@ -225,6 +227,8 @@ def __init__(self, cop_params: tuple = None, weights: tuple = None):
 
         self.copulas = [self.clayton_cop, self.t_cop, self.gumbel_cop]
 
+        devadarsh.track('CTGMixCop')
+
     def fit(self, data: pd.DataFrame, max_iter: int = 25, gamma_scad: float = 0.6, a_scad: float = 6,
             weight_margin: float = 1e-2) -> float:
         """
@@ -631,6 +635,8 @@ def __init__(self, cop_params: list = None, weights: list = None):
 
         self.copulas = [self.clayton_cop, self.frank_cop, self.gumbel_cop]
 
+        devadarsh.track('CFGMixCop')
+
     def fit(self, data: pd.DataFrame, max_iter: int = 25, gamma_scad: float = 0.6, a_scad: float = 6,
             weight_margin: float = 1e-2) -> float:
         """

arbitragelab/copula_approach/copula_strategy_basic.py

@@ -19,9 +19,11 @@
 import numpy as np
 import pandas as pd
 import scipy.stats as ss
+
 import arbitragelab.copula_approach.copula_generate as cg
 import arbitragelab.copula_approach.copula_calculation as ccalc
 import arbitragelab.copula_approach.copula_generate_mixedcopula as cgmix
+from arbitragelab.util import devadarsh
 
 
 class BasicCopulaStrategy:
@@ -73,6 +75,8 @@ def __init__(self, copula: Union[cg.Copula, cgmix.MixedCopula] = None, open_thre
         self._short_count = 0
         self._exit_count = 0
 
+        devadarsh.track('BasicCopulaStrategy')
+
     @staticmethod
     def to_quantile(data: pd.DataFrame) -> Tuple[pd.DataFrame, list]:
         """

arbitragelab/copula_approach/copula_strategy_mpi.py

@@ -12,9 +12,11 @@
 from typing import Callable, Sequence, Union
 import numpy as np
 import pandas as pd
+
 from arbitragelab.copula_approach.copula_strategy_basic import BasicCopulaStrategy
 import arbitragelab.copula_approach.copula_generate as cg
 import arbitragelab.copula_approach.copula_generate_mixedcopula as cgmix
+from arbitragelab.util import devadarsh
 
 
 class CopulaStrategyMPI(BasicCopulaStrategy):
@@ -62,6 +64,8 @@ def __init__(self, copula: Union[cg.Copula, cgmix.MixedCopula] = None,
         self._short_count = 0
         self._exit_count = 0
 
+        devadarsh.track('CopulaStrategyMPI')
+
     @staticmethod
     def to_returns(pair_prices: pd.DataFrame, fill_init_nan: Sequence[float] = (0, 0)) -> pd.DataFrame:
         r"""

arbitragelab/copula_approach/pairs_selection.py

@@ -11,6 +11,8 @@
 import scipy.stats as ss
 import pandas as pd
 
+from arbitragelab.util import devadarsh
+
 
 class PairsSelector:
     r"""
@@ -27,6 +29,8 @@ def __init__(self):
         Class initiation.
         """
 
+        devadarsh.track('PairsSelector')
+
     def rank_pairs(self, stocks_universe: pd.DataFrame, method: str = 'kendall tau',
                    nan_option: Union[str, None] = 'forward fill',
                    keep_num_pairs: Union[int, None] = None) -> pd.Series: