voly 0.0.146__py3-none-any.whl → 0.0.147__py3-none-any.whl

voly/core/hd.py

@@ -8,7 +8,7 @@ import pandas as pd
 import numpy as np
 import datetime as dt
 from scipy import stats
-from typing import Dict, List, Tuple, Optional, Union, Any
+from typing import Dict, List, Tuple, Optional, Union, Any, Callable
 from voly.utils.logger import logger, catch_exception
 from voly.exceptions import VolyError
 from voly.core.rnd import get_all_moments
@@ -27,20 +27,25 @@ def get_historical_data(currency: str,
     Fetch historical OHLCV data for a cryptocurrency.
 
     Parameters:
-    ----------
+    -----------
     currency : str
-        The cryptocurrency to fetch data for (e.g., 'BTC', 'ETH').
+        The cryptocurrency to fetch data for (e.g., 'BTC', 'ETH')
     lookback_days : str
         The lookback period in days, formatted as '90d', '30d', etc.
     granularity : str
-        The time interval for data points (e.g., '15m', '1h', '1d').
+        The time interval for data points (e.g., '15m', '1h', '1d')
     exchange_name : str
-        The exchange to fetch data from (default: 'binance').
+        The exchange to fetch data from (default: 'binance')
 
     Returns:
-    -------
-    pd.DataFrame: Historical price data with OHLCV columns.
+    --------
+    pd.DataFrame
+        Historical price data with OHLCV columns and datetime index
     """
+    # Validate inputs
+    if not lookback_days.endswith('d'):
+        raise VolyError("lookback_days should be in format '90d', '30d', etc.")
+
     try:
         # Get the exchange class from ccxt
         exchange_class = getattr(ccxt, exchange_name.lower())
@@ -49,30 +54,40 @@ def get_historical_data(currency: str,
         raise VolyError(f"Exchange '{exchange_name}' not found in ccxt. Please check the exchange name.")
 
     # Form the trading pair symbol
-    symbol = currency + '/USDT'
+    symbol = f"{currency}/USDT"
 
     # Convert lookback_days to timestamp
-    if lookback_days.endswith('d'):
-        days_ago = int(lookback_days[:-1])
-        date_start = (dt.datetime.now() - dt.timedelta(days=days_ago)).strftime('%Y-%m-%d %H:%M:%S')
-    else:
-        raise VolyError("lookback_days should be in format '90d', '30d', etc.")
-
+    days_ago = int(lookback_days[:-1])
+    date_start = (dt.datetime.now() - dt.timedelta(days=days_ago)).strftime('%Y-%m-%d %H:%M:%S')
     from_ts = exchange.parse8601(date_start)
-    ohlcv_list = []
-    ohlcv = exchange.fetch_ohlcv(symbol, granularity, since=from_ts, limit=1000)
-    ohlcv_list.append(ohlcv)
-
-    # Fetch all available data within the lookback period
-    while len(ohlcv) == 1000:
-        from_ts = ohlcv[-1][0]
-        new_ohlcv = exchange.fetch_ohlcv(symbol, granularity, since=from_ts, limit=1000)
-        if len(new_ohlcv) <= 1:
+
+    # Fetch data with pagination
+    ohlcv = []
+    last_ts = from_ts
+
+    logger.info(f"Fetching {currency} historical data from {exchange_name} for past {days_ago} days")
+
+    while True:
+        batch = exchange.fetch_ohlcv(symbol, granularity, since=last_ts, limit=1000)
+
+        if not batch or len(batch) == 0:
             break
-        ohlcv.extend(new_ohlcv[1:])  # Skip first element to avoid duplication
-        if len(new_ohlcv) < 1000:
+
+        if len(ohlcv) > 0 and batch[0][0] == ohlcv[-1][0]:
+            # Skip first element if it's a duplicate of the last from previous batch
+            batch = batch[1:]
+
+        if not batch:
+            break
+
+        ohlcv.extend(batch)
+        last_ts = batch[-1][0]
+
+        if len(batch) < 1000:
             break
 
+        logger.debug(f"Fetched {len(batch)} candles, total now: {len(ohlcv)}")
+
     # Convert to DataFrame
     df_hist = pd.DataFrame(ohlcv, columns=['date', 'open', 'high', 'low', 'close', 'volume'])
     df_hist['date'] = pd.to_datetime(df_hist['date'], unit='ms')
@@ -87,36 +102,76 @@ def get_historical_data(currency: str,
 @catch_exception
 def parse_window_length(window_length: str, df_hist: pd.DataFrame) -> int:
     """
-    Parse window length from string format (e.g., '7d', '30d') to number of data points.
+    Convert window length string (e.g., '30d') to number of data points.
 
     Parameters:
     -----------
     window_length : str
         Window length in days, formatted as '7d', '30d', etc.
     df_hist : pd.DataFrame
-        Historical data DataFrame with datetime index.
+        Historical data DataFrame with datetime index
 
     Returns:
     --------
     int
-        Number of data points corresponding to the window length.
+        Number of data points corresponding to the window length
     """
+    # Validate inputs
     if not isinstance(window_length, str) or not window_length.endswith('d'):
         raise VolyError("window_length should be in format '7d', '30d', etc.")
 
+    if len(df_hist) < 2:
+        raise VolyError("Historical data must contain at least 2 points to calculate granularity")
+
     # Extract number of days
     days = int(window_length[:-1])
 
-    # Calculate time delta between consecutive data points
-    if len(df_hist) > 1:
-        avg_delta = (df_hist.index[-1] - df_hist.index[0]) / (len(df_hist) - 1)
-        # Convert to days and get points per day
-        days_per_point = avg_delta.total_seconds() / (24 * 60 * 60)
-        # Calculate number of points for the window
-        n_points = int(days / days_per_point)
-        return max(n_points, 10)  # Ensure at least 10 points
-    else:
-        raise VolyError("Not enough data points in df_hist to calculate granularity.")
+    # Calculate average time delta between data points
+    avg_delta = (df_hist.index[-1] - df_hist.index[0]).total_seconds() / (len(df_hist) - 1)
+
+    # Convert to days and calculate points per window
+    days_per_point = avg_delta / (24 * 60 * 60)
+    n_points = int(days / days_per_point)
+
+    # Ensure minimum number of points
+    return max(n_points, 10)
+
+
+def get_param_names(model_type: str, distribution: str) -> List[str]:
+    """
+    Get parameter names for a volatility model and distribution.
+
+    Parameters:
+    -----------
+    model_type : str
+        Type of volatility model ('garch' or 'egarch')
+    distribution : str
+        Distribution type ('normal', 'studentst', or 'skewstudent')
+
+    Returns:
+    --------
+    List[str]
+        List of parameter names
+    """
+    # GARCH(1,1) parameters
+    if model_type.lower() == 'garch':
+        if distribution.lower() == 'normal':
+            return ['mu', 'omega', 'alpha[1]', 'beta[1]']
+        elif distribution.lower() == 'studentst':
+            return ['mu', 'omega', 'alpha[1]', 'beta[1]', 'nu']
+        elif distribution.lower() == 'skewstudent':
+            return ['mu', 'omega', 'alpha[1]', 'beta[1]', 'nu', 'lambda']
+
+    # EGARCH(1,1,1) parameters
+    elif model_type.lower() == 'egarch':
+        if distribution.lower() == 'normal':
+            return ['mu', 'omega', 'alpha[1]', 'gamma[1]', 'beta[1]']
+        elif distribution.lower() == 'studentst':
+            return ['mu', 'omega', 'alpha[1]', 'gamma[1]', 'beta[1]', 'nu']
+        elif distribution.lower() == 'skewstudent':
+            return ['mu', 'omega', 'alpha[1]', 'gamma[1]', 'beta[1]', 'nu', 'lambda']
+
+    raise VolyError(f"Invalid model_type '{model_type}' or distribution '{distribution}'")
 
 
 @catch_exception
@@ -127,85 +182,90 @@ def fit_volatility_model(log_returns: np.ndarray,
                          window_length: str = '30d',
                          n_fits: int = 400) -> Dict[str, Any]:
     """
-    Fit a volatility model (GARCH or EGARCH) to log returns.
+    Fit a volatility model (GARCH or EGARCH) to historical returns.
 
     Parameters:
     -----------
     log_returns : np.ndarray
-        Array of log returns
+        Array of log returns (percent)
     df_hist : pd.DataFrame
-        DataFrame with historical price data
+        Historical price data
     model_type : str
         Type of volatility model ('garch' or 'egarch')
     distribution : str
         Distribution type ('normal', 'studentst', or 'skewstudent')
     window_length : str
-        Length of each window as a string (e.g., '30d')
+        Length of sliding window in days (e.g., '30d')
     n_fits : int
-        Number of sliding windows
+        Number of sliding windows to fit
 
     Returns:
     --------
     Dict[str, Any]
-        Dictionary with model parameters and processes
+        Dictionary with model parameters and fitting results
     """
     # Parse window length
     window_points = parse_window_length(window_length, df_hist)
 
+    # Validate data
     if len(log_returns) < window_points + n_fits:
         raise VolyError(f"Not enough data points. Need at least {window_points + n_fits}, got {len(log_returns)}")
 
-    # Adjust window sizes if necessary to avoid over-fitting
+    # Adjust window sizes to avoid overfitting
     n_fits = min(n_fits, max(100, len(log_returns) // 3))
     window_points = min(window_points, max(20, len(log_returns) // 3))
 
-    start = window_points + n_fits
-    end = n_fits
+    # Calculate start and end indices for sliding windows
+    start_idx = window_points + n_fits
+    end_idx = n_fits
 
-    # Different number of parameters based on model type and distribution
+    # Get parameter names for the model
     param_names = get_param_names(model_type, distribution)
     n_params = len(param_names)
 
+    # Initialize arrays for parameters and innovations
     parameters = np.zeros((n_fits, n_params))
     z_process = []
 
-    logger.info(
-        f"Fitting {model_type.upper()} model with {distribution} distribution using {n_fits} windows of {window_length}...")
+    logger.info(f"Fitting {model_type.upper()} model with {distribution} distribution "
+                f"using {n_fits} windows of {window_length}")
 
+    # Fit models with sliding windows
     for i in range(n_fits):
+        # Log progress
         if i % (n_fits // 10) == 0:
             logger.info(f"Fitting progress: {i}/{n_fits}")
 
-        # Skip if we don't have enough data
-        if end - i - 1 < 0 or start - i - 1 > len(log_returns):
+        # Check if we have enough data for this window
+        if end_idx - i - 1 < 0 or start_idx - i - 1 > len(log_returns):
             continue
 
-        window = log_returns[end - i - 1:start - i - 1]
+        # Extract window data
+        window = log_returns[end_idx - i - 1:start_idx - i - 1]
 
-        # Skip windows that are too small or have invalid data
+        # Skip invalid windows
         if len(window) < 10 or np.isnan(window).any() or np.isinf(window).any():
             continue
 
-        # Mean-center the data to improve numerical stability
+        # Mean-center the data for numerical stability
         data = window - np.mean(window)
 
         try:
-            # Configure model based on type and distribution
+            # Configure and fit model
             if model_type.lower() == 'garch':
                 model = arch_model(data, vol='GARCH', p=1, q=1, dist=distribution.lower())
             else:  # egarch
                 model = arch_model(data, vol='EGARCH', p=1, o=1, q=1, dist=distribution.lower())
 
+            # Fit with optimization settings
             fit_result = model.fit(disp='off', options={'maxiter': 1000})
 
-            # Extract parameters based on model type and distribution
+            # Extract parameters
             params_dict = fit_result.params.to_dict()
-
-            # Extract parameter values in correct order
             param_values = [params_dict.get(param, 0) for param in param_names]
             parameters[i, :] = param_values
 
-            # Get last innovation (standardized residual)
+            # Extract standardized residuals (innovations)
             residuals = fit_result.resid
             conditional_vol = fit_result.conditional_volatility
 
@@ -217,11 +277,11 @@ def fit_volatility_model(log_returns: np.ndarray,
         except Exception as e:
             logger.warning(f"Model fit failed for window {i}: {str(e)}")
 
-    # Clean up any failed fits
+    # Check if we have enough successful fits
     if len(z_process) < n_fits / 2:
         raise VolyError(f"Too many model fits failed ({len(z_process)}/{n_fits}). Check your data.")
 
-    # Filter out rows with zeros (failed fits)
+    # Remove failed fits
     valid_rows = ~np.all(parameters == 0, axis=1)
     parameters = parameters[valid_rows]
 
@@ -240,100 +300,95 @@ def fit_volatility_model(log_returns: np.ndarray,
     }
 
 
-def get_param_names(model_type: str, distribution: str) -> List[str]:
+@catch_exception
+def create_innovation_sampler(vol_model: Dict[str, Any]) -> Callable:
     """
-    Get parameter names based on model type and distribution.
+    Create a function to sample innovations based on the volatility model.
 
     Parameters:
     -----------
-    model_type : str
-        Type of volatility model ('garch' or 'egarch')
-    distribution : str
-        Distribution type ('normal', 'studentst', or 'skewstudent')
+    vol_model : Dict[str, Any]
+        Volatility model information from fit_volatility_model()
 
     Returns:
     --------
-    List[str]
-        List of parameter names
+    Callable
+        Function that returns random innovations when called
     """
-    if model_type.lower() == 'garch':
-        if distribution.lower() == 'normal':
-            return ['mu', 'omega', 'alpha[1]', 'beta[1]']
-        elif distribution.lower() == 'studentst':
-            return ['mu', 'omega', 'alpha[1]', 'beta[1]', 'nu']
-        else:  # skewstudent
-            return ['mu', 'omega', 'alpha[1]', 'beta[1]', 'nu', 'lambda']
-    else:  # egarch
-        if distribution.lower() == 'normal':
-            return ['mu', 'omega', 'alpha[1]', 'gamma[1]', 'beta[1]']
-        elif distribution.lower() == 'studentst':
-            return ['mu', 'omega', 'alpha[1]', 'gamma[1]', 'beta[1]', 'nu']
-        else:  # skewstudent
-            return ['mu', 'omega', 'alpha[1]', 'gamma[1]', 'beta[1]', 'nu', 'lambda']
+    distribution = vol_model['distribution']
+    z_process = vol_model['z_process']
+
+    if distribution.lower() == 'normal':
+        # Use standard normal for normal distribution
+        def sample_innovation(size=1):
+            return np.random.normal(0, 1, size=size)
+    else:
+        # Use KDE for non-normal distributions to capture empirical distribution
+        kde = stats.gaussian_kde(z_process, bw_method='silverman')
+        z_range = np.linspace(min(z_process), max(z_process), 1000)
+        z_prob = kde(z_range)
+        z_prob = z_prob / np.sum(z_prob)
+
+        def sample_innovation(size=1):
+            return np.random.choice(z_range, size=size, p=z_prob)
+
+    return sample_innovation
 
 
 @catch_exception
-def simulate_volatility_paths(vol_model: Dict[str, Any],
+def generate_volatility_paths(vol_model: Dict[str, Any],
                               horizon: int,
                               simulations: int = 5000) -> Tuple[np.ndarray, float]:
     """
-    Simulate future paths using a fitted volatility model.
+    Simulate future price paths using a fitted volatility model.
 
     Parameters:
     -----------
     vol_model : Dict[str, Any]
-        Dict with volatility model parameters
+        Volatility model information from fit_volatility_model()
     horizon : int
-        Number of steps to simulate
+        Number of time steps to simulate
     simulations : int
         Number of paths to simulate
 
     Returns:
     --------
     Tuple[np.ndarray, float]
-        Simulated returns and drift
+        Array of simulated returns and the drift term
     """
+    # Extract model information
     parameters = vol_model['parameters']
-    z_process = vol_model['z_process']
     model_type = vol_model['model_type']
     distribution = vol_model['distribution']
     param_names = vol_model['param_names']
 
-    # Use mean parameters as starting point
+    # Get mean parameters
     pars = vol_model['avg_params'].copy()
     bounds = vol_model['std_params'].copy()
 
-    # Create dictionary for easier parameter access
+    # Create parameter dictionary for easier access
     param_dict = {name: value for name, value in zip(param_names, pars)}
 
-    # Log parameters in a structured way
+    # Log parameters
     param_str = ", ".join([f"{name}={param_dict.get(name, 0):.6f}" for name in param_names])
     logger.info(f"{model_type.upper()} parameters: {param_str}")
 
-    # Create sampling function based on distribution
-    if distribution.lower() == 'normal':
-        # Use standard normal for normal distribution
-        def sample_innovation(size=1):
-            return np.random.normal(0, 1, size=size)
-    else:
-        # Use KDE for non-normal distributions to capture empirical distribution
-        kde = stats.gaussian_kde(z_process, bw_method='silverman')  # original code didnt have bw_method
-        z_range = np.linspace(min(z_process), max(z_process), 1000)
-        z_prob = kde(z_range)
-        z_prob = z_prob / np.sum(z_prob)
+    # Create innovation sampler
+    sample_innovation = create_innovation_sampler(vol_model)
 
-        def sample_innovation(size=1):
-            return np.random.choice(z_range, size=size, p=z_prob)
-
-    # Simulate paths
+    # Initialize results array
     simulated_returns = np.zeros(simulations)
     mu = param_dict.get('mu', 0)
 
+    logger.info(f"Simulating {simulations} paths for horizon {horizon}")
+
+    # Simulate paths
     for i in range(simulations):
+        # Log progress
         if (i + 1) % (simulations // 10) == 0:
             logger.info(f"Simulation progress: {i + 1}/{simulations}")
 
-        # Optionally vary parameters between simulations
+        # Vary parameters periodically for robustness
         if (i + 1) % (simulations // 20) == 0:
             # Create parameter variations based on their estimated distribution
             sim_params = {}
@@ -356,25 +411,30 @@ def simulate_volatility_paths(vol_model: Dict[str, Any],
 
         # Initialize volatility based on model type
         if model_type.lower() == 'garch':
+            # Extract GARCH parameters
             omega = sim_params.get('omega', 0)
             alpha = sim_params.get('alpha[1]', 0)
             beta = sim_params.get('beta[1]', 0)
 
-            # Initialize GARCH volatility (unconditional variance)
-            sigma2 = omega / (1 - alpha - beta) if alpha + beta < 1 else omega / 0.99
+            # Initialize with unconditional variance
+            persistence = alpha + beta
+            sigma2 = omega / (1 - persistence) if persistence < 1 else omega / 0.99
+
         else:  # egarch
+            # Extract EGARCH parameters
             omega = sim_params.get('omega', 0)
             beta = sim_params.get('beta[1]', 0)
 
-            # Initialize EGARCH volatility
+            # Initialize log variance
             log_sigma2 = omega / (1 - beta) if beta < 1 else omega / 0.99
             sigma2 = np.exp(log_sigma2)
 
+        # Initialize return sum
         returns_sum = 0
 
-        # Simulate path step by step
+        # Simulate path
         for _ in range(horizon):
-            # Sample a random innovation
+            # Sample innovation
             z = sample_innovation()
 
             # Update returns and volatility based on model type
@@ -384,29 +444,265 @@ def simulate_volatility_paths(vol_model: Dict[str, Any],
                 returns_sum += e + mu
 
                 # Update GARCH volatility
-                sigma2 = sim_params.get('omega', 0) + sim_params.get('alpha[1]', 0) * e ** 2 + sim_params.get('beta[1]',
-                                                                                                              0) * sigma2
+                sigma2 = (sim_params.get('omega', 0) +
+                          sim_params.get('alpha[1]', 0) * e ** 2 +
+                          sim_params.get('beta[1]', 0) * sigma2)
+
             else:  # egarch
                 # Calculate return
                 e = z * np.sqrt(sigma2)
                 returns_sum += e + mu
 
-                # Update EGARCH volatility
-                abs_z = abs(z)
+                # Extract EGARCH parameters
                 gamma = sim_params.get('gamma[1]', 0)
                 alpha = sim_params.get('alpha[1]', 0)
                 beta = sim_params.get('beta[1]', 0)
                 omega = sim_params.get('omega', 0)
 
-                # EGARCH update equation
+                # Update EGARCH volatility
+                abs_z = abs(z)
                 log_sigma2 = omega + beta * log_sigma2 + alpha * (abs_z - np.sqrt(2 / np.pi)) + gamma * z
                 sigma2 = np.exp(log_sigma2)
 
+        # Store final return
         simulated_returns[i] = returns_sum
 
     return simulated_returns, mu * horizon
 
 
+@catch_exception
+def prepare_domains(domain_params: Tuple[float, float, int],
+                    s: float,
+                    return_domain: str) -> Dict[str, np.ndarray]:
+    """
+    Prepare domain arrays for different representations.
+
+    Parameters:
+    -----------
+    domain_params : Tuple[float, float, int]
+        (min_log_moneyness, max_log_moneyness, num_points)
+    s : float
+        Spot price
+    return_domain : str
+        Domain for results
+
+    Returns:
+    --------
+    Dict[str, np.ndarray]
+        Dictionary of domain arrays
+    """
+    # Create log-moneyness grid
+    LM = np.linspace(domain_params[0], domain_params[1], domain_params[2])
+
+    # Calculate other domains
+    M = np.exp(LM)  # Moneyness
+    R = M - 1  # Returns
+    K = s / M  # Strike prices
+
+    # Calculate grid spacing
+    dx = LM[1] - LM[0]
+
+    return {
+        'log_moneyness': LM,
+        'moneyness': M,
+        'returns': R,
+        'strikes': K,
+        'dx': dx
+    }
+
+
+@catch_exception
+def calculate_basic_density(df_hist: pd.DataFrame,
+                            t: float,
+                            r: float,
+                            n_periods: int,
+                            domains: Dict[str, np.ndarray],
+                            bandwidth: str = 'silverman') -> Dict[str, np.ndarray]:
+    """
+    Calculate historical density using KDE of historical returns.
+
+    Parameters:
+    -----------
+    df_hist : pd.DataFrame
+        Historical price data
+    t : float
+        Time to maturity in years
+    r : float
+        Risk-free rate
+    n_periods : int
+        Number of periods to scale returns
+    domains : Dict[str, np.ndarray]
+        Domain arrays
+    bandwidth : str
+        KDE bandwidth method
+
+    Returns:
+    --------
+    Dict[str, np.ndarray]
+        Dictionary of PDFs in different domains
+    """
+    # Extract domains
+    LM = domains['log_moneyness']
+    M = domains['moneyness']
+    R = domains['returns']
+    K = domains['strikes']
+    dx = domains['dx']
+
+    # Filter historical data for the maturity's lookback period
+    start_date = pd.Timestamp.now() - pd.Timedelta(days=int(t * 365.25))
+    maturity_hist = df_hist[df_hist.index >= start_date].copy()
+
+    if len(maturity_hist) < 10:
+        raise VolyError(f"Not enough historical data for maturity (t={t:.4f})")
+
+    # Calculate scaled returns
+    maturity_hist['log_returns'] = np.log(maturity_hist['close'] / maturity_hist['close'].shift(1)) * np.sqrt(n_periods)
+    maturity_hist = maturity_hist.dropna()
+    returns = maturity_hist['log_returns'].values
+
+    if len(returns) < 2:
+        raise VolyError(f"Not enough valid returns for maturity (t={t:.4f})")
+
+    # Girsanov adjustment to shift to risk-neutral measure
+    mu_scaled = returns.mean()
+    sigma_scaled = returns.std()
+    expected_risk_neutral_mean = (r - 0.5 * sigma_scaled ** 2) * np.sqrt(t)
+    adjustment = mu_scaled - expected_risk_neutral_mean
+    adj_returns = returns - adjustment
+
+    # Create PDF with KDE
+    kde = stats.gaussian_kde(adj_returns, bw_method=bandwidth)
+    pdf_lm = kde(LM)
+
+    # Normalize the PDF
+    pdf_lm = pdf_lm / np.trapz(pdf_lm, LM)
+
+    # Transform to other domains
+    pdf_m = pdf_lm / M
+    pdf_k = pdf_lm / K
+    pdf_r = pdf_lm / (1 + R)
+
+    # Calculate CDF
+    cdf = np.cumsum(pdf_lm * dx)
+    cdf = cdf / cdf[-1]
+
+    return {
+        'log_moneyness': pdf_lm,
+        'moneyness': pdf_m,
+        'returns': pdf_r,
+        'strikes': pdf_k,
+        'cdf': cdf
+    }
+
+
+@catch_exception
+def calculate_volatility_density(vol_model: Dict[str, Any],
+                                 s: float,
+                                 t: float,
+                                 r: float,
+                                 n_periods: int,
+                                 tau_days: float,
+                                 domains: Dict[str, np.ndarray],
+                                 simulations: int = 5000,
+                                 bandwidth: str = 'silverman') -> Tuple[Dict[str, np.ndarray], Dict[str, Any]]:
+    """
+    Calculate historical density using volatility model simulation.
+
+    Parameters:
+    -----------
+    vol_model : Dict[str, Any]
+        Volatility model from fit_volatility_model()
+    s : float
+        Spot price
+    t : float
+        Time to maturity in years
+    r : float
+        Risk-free rate
+    n_periods : int
+        Number of periods to scale returns
+    tau_days : float
+        Days to maturity
+    domains : Dict[str, np.ndarray]
+        Domain arrays
+    simulations : int
+        Number of Monte Carlo simulations
+    bandwidth : str
+        KDE bandwidth method
+
+    Returns:
+    --------
+    Tuple[Dict[str, np.ndarray], Dict[str, Any]]
+        Dictionary of PDFs in different domains and model parameters
+    """
+    # Extract domains
+    LM = domains['log_moneyness']
+    M = domains['moneyness']
+    R = domains['returns']
+    K = domains['strikes']
+    dx = domains['dx']
+
+    # Simulate paths with the volatility model
+    horizon = max(1, int(tau_days))
+    simulated_returns, simulated_mu = generate_volatility_paths(
+        vol_model,
+        horizon,
+        simulations
+    )
+
+    # Scale the simulated returns to match target time horizon
+    scaling_factor = np.sqrt(n_periods / tau_days)
+    scaled_returns = simulated_returns * scaling_factor
+
+    # Risk-neutral adjustment
+    mu_scaled = scaled_returns.mean()
+    sigma_scaled = scaled_returns.std()
+    expected_risk_neutral_mean = (r - 0.5 * (sigma_scaled / 100) ** 2) * 100 * np.sqrt(t)
+    adjustment = mu_scaled - expected_risk_neutral_mean
+    risk_neutral_returns = scaled_returns - adjustment
+
+    # Convert to terminal prices
+    simulated_prices = s * np.exp(risk_neutral_returns / 100)
+
+    # Convert to moneyness domain (x-domain)
+    simulated_moneyness = s / simulated_prices
+
+    # Calculate PDF with KDE
+    kde = stats.gaussian_kde(simulated_moneyness, bw_method=bandwidth)
+    pdf_m = kde(M)
+
+    # Normalize the PDF
+    pdf_m = pdf_m / np.trapz(pdf_m, M)
+
+    # Transform to other domains
+    pdf_lm = pdf_m * M
+    pdf_k = pdf_lm / K
+    pdf_r = pdf_lm / (1 + R)
+
+    # Calculate CDF
+    cdf = np.cumsum(pdf_lm * dx)
+    cdf = cdf / cdf[-1]
+
+    # Prepare model parameters for moments
+    avg_params = vol_model['avg_params']
+    param_names = vol_model['param_names']
+    model_params = {name.replace('[1]', ''): value for name, value in zip(param_names, avg_params)}
+    model_params['model_type'] = vol_model['model_type']
+    model_params['distribution'] = vol_model['distribution']
+
+    # Add persistence for GARCH models
+    if vol_model['model_type'] == 'garch':
+        model_params['persistence'] = model_params.get('alpha', 0) + model_params.get('beta', 0)
+
+    return {
+        'log_moneyness': pdf_lm,
+        'moneyness': pdf_m,
+        'returns': pdf_r,
+        'strikes': pdf_k,
+        'cdf': cdf
+    }, model_params
+
+
+@catch_exception
 def get_hd_surface(model_results: pd.DataFrame,
                    df_hist: pd.DataFrame,
                    domain_params: Tuple[float, float, int] = (-1.5, 1.5, 1000),
@@ -427,29 +723,26 @@ def get_hd_surface(model_results: pd.DataFrame,
     df_hist : pd.DataFrame
         DataFrame with historical price data
     domain_params : Tuple[float, float, int]
-        Tuple of (min, max, num_points) for x-domain
+        (min_log_moneyness, max_log_moneyness, num_points)
     return_domain : str
-        Domain for x-axis values ('log_moneyness', 'moneyness', 'returns', 'strikes')
+        Domain for results ('log_moneyness', 'moneyness', 'returns', 'strikes')
     method : str
-        Method to use for HD estimation:
-        - 'garch': GARCH(1,1) model
-        - 'egarch': EGARCH(1,1,1) model with asymmetry
-        - 'basic': Simple histogram/KDE of historical returns
+        Method for HD estimation ('garch', 'egarch', 'basic')
     distribution : str
-        Distribution to use for volatility models ('normal', 'studentst', or 'skewstudent')
+        Distribution for volatility models ('normal', 'studentst', 'skewstudent')
     window_length : str
-        Length of sliding windows as string (e.g., '30d')
+        Length of sliding windows for model fitting (e.g., '30d')
     n_fits : int
-        Number of sliding windows for volatility model fitting
+        Number of sliding windows for model fitting
     simulations : int
-        Number of Monte Carlo simulations for volatility models
+        Number of Monte Carlo simulations
     bandwidth : str
-        KDE bandwidth method (default: 'silverman')
+        KDE bandwidth method
 
     Returns:
     --------
     Dict[str, Any]
-        Dictionary containing pdf_surface, cdf_surface, x_surface, and moments
+        Dictionary with pdf_surface, cdf_surface, x_surface, and moments
     """
     # Validate inputs
     required_columns = ['s', 't', 'r']
@@ -460,11 +753,11 @@ def get_hd_surface(model_results: pd.DataFrame,
     if len(df_hist) < 2:
         raise VolyError("Not enough data points in df_hist")
 
-    # Determine granularity from df_hist
+    # Determine granularity from data
     minutes_diff = (df_hist.index[1] - df_hist.index[0]).total_seconds() / 60
     minutes_per_period = max(1, int(minutes_diff))
 
-    # Validate method and model parameters
+    # Validate method and distribution
     valid_methods = ['garch', 'egarch', 'basic']
     valid_distributions = ['normal', 'studentst', 'skewstudent']
 
@@ -477,16 +770,20 @@ def get_hd_surface(model_results: pd.DataFrame,
     if method in ['garch', 'egarch'] and distribution not in valid_distributions:
         raise VolyError(f"Invalid distribution: {distribution}. Must be one of {valid_distributions}")
 
-    # Calculate log returns from price history
+    # Validate return domain
+    valid_domains = ['log_moneyness', 'moneyness', 'returns', 'strikes']
+    if return_domain not in valid_domains:
+        raise VolyError(f"Invalid return_domain: {return_domain}. Must be one of {valid_domains}")
+
+    # Calculate log returns
     log_returns = np.log(df_hist['close'] / df_hist['close'].shift(1)) * 100
     log_returns = log_returns.dropna().values
 
-    # Fit volatility model if using GARCH or EGARCH
+    # Fit volatility model if needed
     vol_model = None
     if method in ['garch', 'egarch']:
-        model_type = method  # Use method as model_type
-        logger.info(
-            f"Using {model_type.upper()} with {distribution} distribution, {n_fits} fits, {simulations} simulations")
+        model_type = method
+        logger.info(f"Using {model_type.upper()} with {distribution} distribution")
 
         vol_model = fit_volatility_model(
             log_returns=log_returns,
@@ -496,8 +793,6 @@ def get_hd_surface(model_results: pd.DataFrame,
             window_length=window_length,
             n_fits=n_fits
         )
-    elif method == 'basic':
-        logger.info(f"Using basic returns-based KDE method with bandwidth {bandwidth}")
 
     # Initialize result containers
     pdf_surface = {}
@@ -507,155 +802,75 @@ def get_hd_surface(model_results: pd.DataFrame,
 
     # Process each maturity
     for i in model_results.index:
-        # Get parameters for this maturity
-        s = model_results.loc[i, 's']  # Current spot price
-        r = model_results.loc[i, 'r']  # Risk-free rate
-        t = model_results.loc[i, 't']  # Time to maturity in years
-
-        # Get domain grids
-        LM = np.linspace(domain_params[0], domain_params[1], domain_params[2])
-        M = np.exp(LM)  # Moneyness
-        R = M - 1  # Returns
-        K = s / M  # Strike prices
-
-        # For time scaling calculations
-        tau_days_float = t * 365.25  # Exact number of days
-        n_periods = max(1, int(t * 365.25 * 24 * 60 / minutes_per_period))
-
-        logger.info(f"Processing HD for maturity {i} (t={t:.4f} years, {tau_days_float:.2f} days)")
-
-        if method == 'basic':
-            # Simple returns-based method
-            # Filter historical data for this maturity's lookback period
-            start_date = pd.Timestamp.now() - pd.Timedelta(days=int(t * 365.25))
-            maturity_hist = df_hist[df_hist.index >= start_date].copy()
-
-            if len(maturity_hist) < 10:
-                logger.warning(f"Not enough historical data for maturity {i}, skipping.")
-                continue
-
-            # Calculate scaled returns
-            maturity_hist['log_returns'] = np.log(maturity_hist['close'] / maturity_hist['close'].shift(1)) * np.sqrt(
-                n_periods)
-            maturity_hist = maturity_hist.dropna()
-
-            returns = maturity_hist['log_returns'].values
-            if len(returns) < 2:
-                logger.warning(f"Not enough valid returns for maturity {i}, skipping.")
-                continue
-
-            # Girsanov adjustment to shift to risk-neutral measure
-            mu_scaled = returns.mean()
-            sigma_scaled = returns.std()
-            expected_risk_neutral_mean = (r - 0.5 * sigma_scaled ** 2) * np.sqrt(t)
-            adjustment = mu_scaled - expected_risk_neutral_mean
-            adj_returns = returns - adjustment
-
-            # Create HD and normalize
-            f = stats.gaussian_kde(adj_returns, bw_method=bandwidth)
-            pdf_values = f(LM)
-
-            # Transform according to return domain
-            pdf_lm = pdf_values
-            pdf_m = pdf_lm / M
-            pdf_k = pdf_lm / K
-            pdf_r = pdf_lm / (1 + R)
-
-            # No model parameters to include
-            model_params = None
-
-        elif method in ['garch', 'egarch']:
-            # Volatility model-based method
-            if vol_model is None:
-                logger.warning(f"Volatility model fitting failed, skipping maturity {i}")
-                continue
-
-            # Simulate paths with the volatility model
-            horizon = max(1, int(tau_days_float))
-            simulated_returns, simulated_mu = simulate_volatility_paths(
-                vol_model,
-                horizon,
-                simulations
-            )
-
-            # Scale the simulated returns to match target time horizon
-            scaling_factor = np.sqrt(n_periods / tau_days_float)
-            scaled_returns = simulated_returns * scaling_factor
-
-            # Risk-neutral adjustment
-            mu_scaled = scaled_returns.mean()
-            sigma_scaled = scaled_returns.std()
-            expected_risk_neutral_mean = (r - 0.5 * (sigma_scaled / 100) ** 2) * 100 * np.sqrt(t)
-            adjustment = mu_scaled - expected_risk_neutral_mean
-            risk_neutral_returns = scaled_returns - adjustment
-
-            # Convert to terminal prices
-            simulated_prices = s * np.exp(risk_neutral_returns / 100)
-
-            # Convert to moneyness domain (x-domain)
-            simulated_moneyness = s / simulated_prices
-
-            # Perform KDE to get PDF
-            kde = stats.gaussian_kde(simulated_moneyness, bw_method=bandwidth)
-            pdf_values = kde(M)
-
-            # Transform according to return domain
-            pdf_m = pdf_values
-            pdf_lm = pdf_m * M
-            pdf_k = pdf_lm / K
-            pdf_r = pdf_lm / (1 + R)
-
-            # Include volatility model params in moments
-            avg_params = vol_model['avg_params']
-            param_names = vol_model['param_names']
-            model_params = {name.replace('[1]', ''): value for name, value in zip(param_names, avg_params)}
-            model_params['model_type'] = method
-            model_params['distribution'] = distribution
-
-            # Add persistence for GARCH models
-            if method == 'garch':
-                model_params['persistence'] = model_params.get('alpha', 0) + model_params.get('beta', 0)
-        else:
-            continue  # Skip if invalid method
-
-        # Ensure density integrates to 1
-        dx = LM[1] - LM[0]
-        total_area = np.sum(pdf_values * dx)
-        if total_area <= 0:
-            logger.warning(f"Invalid density (area <= 0) for maturity {i}, skipping.")
-            continue
-
-        pdf_values = pdf_values / total_area
-
-        # Calculate CDF
-        cdf = np.cumsum(pdf_lm * dx)
-        cdf = np.minimum(cdf / cdf[-1], 1.0)  # Ensure CDF is between 0 and 1
-
-        # Select appropriate domain and calculate moments
-        if return_domain == 'log_moneyness':
-            x = LM
-            pdf = pdf_lm
-            moments = get_all_moments(x, pdf, model_params)
-        elif return_domain == 'moneyness':
-            x = M
-            pdf = pdf_m
-            moments = get_all_moments(x, pdf, model_params)
-        elif return_domain == 'returns':
-            x = R
-            pdf = pdf_r
-            moments = get_all_moments(x, pdf, model_params)
-        elif return_domain == 'strikes':
-            x = K
-            pdf = pdf_k
+        try:
+            # Get parameters for this maturity
+            s = model_results.loc[i, 's']  # Spot price
+            r = model_results.loc[i, 'r']  # Risk-free rate
+            t = model_results.loc[i, 't']  # Time to maturity in years
+
+            # Calculate time scaling parameters
+            tau_days = t * 365.25  # Days to expiry
+            n_periods = max(1, int(tau_days * 24 * 60 / minutes_per_period))  # Number of periods
+
+            logger.info(f"Processing HD for maturity {i} (t={t:.4f} years, {tau_days:.2f} days)")
+
+            # Prepare domains
+            domains = prepare_domains(domain_params, s, return_domain)
+
+            # Calculate density based on method
+            if method == 'basic':
+                pdfs = calculate_basic_density(
+                    df_hist=df_hist,
+                    t=t,
+                    r=r,
+                    n_periods=n_periods,
+                    domains=domains,
+                    bandwidth=bandwidth
+                )
+                model_params = None
+
+            else:  # 'garch' or 'egarch'
+                if vol_model is None:
+                    logger.warning(f"Volatility model fitting failed, skipping maturity {i}")
+                    continue
+
+                pdfs, model_params = calculate_volatility_density(
+                    vol_model=vol_model,
+                    s=s,
+                    t=t,
+                    r=r,
+                    n_periods=n_periods,
+                    tau_days=tau_days,
+                    domains=domains,
+                    simulations=simulations,
+                    bandwidth=bandwidth
+                )
+
+            # Get domain arrays for output
+            if return_domain == 'log_moneyness':
+                x = domains['log_moneyness']
+                pdf = pdfs['log_moneyness']
+            elif return_domain == 'moneyness':
+                x = domains['moneyness']
+                pdf = pdfs['moneyness']
+            elif return_domain == 'returns':
+                x = domains['returns']
+                pdf = pdfs['returns']
+            elif return_domain == 'strikes':
+                x = domains['strikes']
+                pdf = pdfs['strikes']
+
+            # Calculate statistical moments
             moments = get_all_moments(x, pdf, model_params)
-        else:
-            raise VolyError(f"Unsupported return_domain: {return_domain}")
 
-        # Store results
-        pdf_surface[i] = pdf
-        cdf_surface[i] = cdf
-        x_surface[i] = x
-        all_moments[i] = moments
+            # Store results
+            pdf_surface[i] = pdf
+            cdf_surface[i] = pdfs['cdf']
+            x_surface[i] = x
+            all_moments[i] = moments
+
+        except Exception as e:
+            logger.warning(f"Failed to calculate HD for maturity {i}: {str(e)}")
 
     # Check if we have any valid results
     if not pdf_surface: