"""
Bollinger Bands indicator implementation.
"""

import pandas as pd

from ..base import BaseIndicator


class BollingerBandsIndicator(BaseIndicator):
    """
    Bollinger Bands technical indicator.
    
    Calculates a set of lines plotted two standard deviations away from a simple moving average.
    Handles sparse data appropriately without interpolation.
    """
    
    def calculate(self, df: pd.DataFrame, period: int = 20,
                 std_dev: float = 2.0, price_column: str = 'close') -> pd.DataFrame:
        """
        Calculate Bollinger Bands.
        
        Args:
            df: DataFrame with OHLCV data
            period: Number of periods for moving average (default 20)
            std_dev: Number of standard deviations (default 2.0)
            price_column: Price column to use ('open', 'high', 'low', 'close')
            
        Returns:
            DataFrame with Bollinger Bands values and metadata, indexed by timestamp
        """
        # Validate input data
        if not self.validate_dataframe(df, period):
            return pd.DataFrame()
        try:
            df = df.copy()
            df['middle_band'] = df[price_column].rolling(window=period, min_periods=period).mean()
            df['std'] = df[price_column].rolling(window=period, min_periods=period).std()
            df['upper_band'] = df['middle_band'] + (df['std'] * std_dev)
            df['lower_band'] = df['middle_band'] - (df['std'] * std_dev)
            # Only keep rows with valid bands, and only 'timestamp', 'upper_band', 'middle_band', 'lower_band' columns
            result_df = df.loc[df['middle_band'].notna() & df['upper_band'].notna() & df['lower_band'].notna(), ['timestamp', 'upper_band', 'middle_band', 'lower_band']].copy()
            result_df.set_index('timestamp', inplace=True)
            return result_df
        except Exception:
            return pd.DataFrame()