sarpyx 0.1.5__py3-none-any.whl → 0.1.6__py3-none-any.whl

sarpyx/processor/core/focus_v2.py

@@ -1,1625 +0,0 @@
-import argparse
-from typing import Dict, Any, Optional, Union, Tuple, Callable, List
-try:
-    import torch
-except ImportError:
-    print('Unable to import torch module')
-    torch = None
-import pickle
-import pandas as pd
-import numpy as np
-from scipy.interpolate import interp1d
-import math
-from pathlib import Path 
-import matplotlib.pyplot as plt
-import copy 
-import gc
-from functools import wraps
-import psutil
-import time
-from os import environ
-
-try:
-    import zarr
-    import numcodecs
-    ZARR_AVAILABLE = True
-except ImportError:
-    print('Warning: zarr not available, falling back to pickle for saving')
-    ZARR_AVAILABLE = False
-
-# ---------- Import custom modules ----------
-from .code2physical import range_dec_to_sample_rate
-from .transforms import perform_fft_custom 
-from . import constants as cnst
-from ..utils.viz import dump
-
-
-# ---------- Global settings ----------
-environ['OMP_NUM_THREADS'] = '12' # Set OpenMP threads for parallel processing
-__VTIMING__ = False
-
-
-
-# ---------- Decorators and utility functions ----------
-def timing_decorator(func: Callable) -> Callable:
-    """Decorator to measure and print function execution time.
-    
-    Args:
-        func: The function to measure.
-        
-    Returns:
-        The wrapped function with timing measurement.
-    """
-    @wraps(func)
-    def wrapper(*args, **kwargs):
-        start_time = time.time()
-        result = func(*args, **kwargs)
-        elapsed_time = time.time() - start_time
-        if __VTIMING__:
-            print(f'Elapsed time for {func.__name__}: {elapsed_time:.4f} seconds')
-        else:
-            # Only print if __VTIMING__ is enabled
-            pass
-        return result
-    return wrapper
-
-def print_memory() -> None:
-    """Print current RAM memory usage percentage."""
-    print(f'RAM memory usage: {psutil.virtual_memory().percent}%')
-
-def flush_mem(func: Callable) -> Callable:
-    """Decorator for memory-efficient operations with monitoring.
-    
-    Args:
-        func: The function to wrap.
-        
-    Returns:
-        The wrapped function with memory monitoring and cleanup.
-    """
-    @wraps(func)
-    def wrapper(*args, **kwargs):
-        # Monitor memory before
-        initial_memory = psutil.virtual_memory().percent
-        
-        # Execute function
-        result = func(*args, **kwargs)
-        
-        # Force garbage collection
-        gc.collect()
-        
-        # Monitor memory after
-        final_memory = psutil.virtual_memory().percent
-        
-        # Print memory change if verbose
-        if hasattr(args[0], '_verbose') and args[0]._verbose:
-            print(f'Memory usage: {initial_memory:.1f}% -> {final_memory:.1f}% '
-                  f'(Δ{final_memory - initial_memory:+.1f}%)')
-        
-        return result
-    return wrapper
-
-def cleanup_variables(*variables: Any) -> None:
-    """Explicitly delete variables and run garbage collection.
-    
-    Args:
-        *variables: Variables to delete.
-    """
-    for var in variables:
-        del var
-    gc.collect()
-
-def initialize_params(
-    device: Optional[torch.device] = None,
-    slant_range_vec: Optional[np.ndarray] = None,
-    D: Optional[np.ndarray] = None,
-    c: Optional[float] = None,
-    len_range_line: Optional[int] = None,
-    range_sample_freq: Optional[float] = None,
-    wavelength: Optional[float] = None
-) -> Dict[str, Any]:
-    """Initialize processing parameters dictionary.
-    
-    Args:
-        device: PyTorch device for computation.
-        slant_range_vec: Slant range vector.
-        D: Cosine of instantaneous squint angle.
-        c: Speed of light.
-        len_range_line: Length of range line.
-        range_sample_freq: Range sampling frequency.
-        wavelength: Radar wavelength.
-        
-    Returns:
-        Dictionary containing all parameters.
-    """
-    return {key: value for key, value in locals().items()}
-
-def multiply_inplace(
-    a: Union[np.ndarray, torch.Tensor], 
-    b: Union[np.ndarray, torch.Tensor]
-) -> Union[np.ndarray, torch.Tensor]:
-    """Multiply two arrays element-wise in-place with broadcasting support.
-    
-    Args:
-        a: First array (modified in-place).
-        b: Second array.
-        
-    Returns:
-        Reference to modified first array.
-        
-    Raises:
-        ValueError: If arrays have incompatible shapes for broadcasting.
-    """
-    if hasattr(a, 'shape') and hasattr(b, 'shape'):
-        # Check if shapes are compatible for broadcasting
-        if a.shape != b.shape and b.size != 1 and a.size != 1:
-            # For 2D array * 1D array, the 1D array should match one of the 2D dimensions
-            if len(a.shape) == 2 and len(b.shape) == 1:
-                if b.shape[0] == a.shape[1]:
-                    # Broadcasting along range dimension - use numpy broadcasting
-                    pass  # NumPy will handle this automatically
-                elif b.shape[0] == a.shape[0]:
-                    # Need to reshape for azimuth dimension broadcasting
-                    b = b.reshape(-1, 1)
-                else:
-                    raise ValueError(f'1D array length ({b.shape[0]}) does not match either dimension of 2D array {a.shape}')
-    
-    # Perform in-place multiplication
-    try:
-        if isinstance(a, np.ndarray):
-            np.multiply(a, b, out=a)
-        else:  # torch tensor
-            a.mul_(b)
-        return a
-    except (ValueError, RuntimeError) as e:
-        raise ValueError(f'Arrays have incompatible shapes for in-place broadcasting: {a.shape} and {b.shape}. '
-                       f'Original error: {str(e)}') from e
-
-def multiply(
-    a: Union[np.ndarray, torch.Tensor], 
-    b: Union[np.ndarray, torch.Tensor],
-    debug: bool = False,
-) -> Union[np.ndarray, torch.Tensor]:
-    """Multiply two arrays element-wise with broadcasting support.
-    
-    Args:
-        a: First array.
-        b: Second array.
-        
-    Returns:
-        Element-wise multiplication result.
-        
-    Raises:
-        ValueError: If arrays have incompatible shapes for broadcasting.
-    """
-    if hasattr(a, 'shape') and hasattr(b, 'shape'):
-        # Check if shapes are compatible for broadcasting
-        if a.shape != b.shape and b.size != 1 and a.size != 1:
-            # Try to understand the broadcasting scenario
-            if debug:
-                print(f'Debug: Attempting to multiply arrays with shapes {a.shape} and {b.shape}')
-            
-            # For 2D array * 1D array, the 1D array should match one of the 2D dimensions
-            if len(a.shape) == 2 and len(b.shape) == 1:
-                if debug:
-                    if b.shape[0] == a.shape[1]:
-                        print(f'Debug: Broadcasting 1D array along range dimension (axis=1)')
-                    elif b.shape[0] == a.shape[0]:
-                        print(f'Debug: Need to reshape 1D array for azimuth dimension (axis=0)')
-                        b = b.reshape(-1, 1)  # Reshape for broadcasting along azimuth
-                    else:
-                        raise ValueError(f'1D array length ({b.shape[0]}) does not match either dimension of 2D array {a.shape}')
-            
-            # Allow broadcasting for compatible shapes
-            try:
-                result = a * b
-                if debug:
-                    print(f'Debug: Broadcasting successful, result shape: {result.shape}')
-                return result
-            except (ValueError, RuntimeError) as e:
-                print(f'Debug: Broadcasting failed with error: {str(e)}')
-                raise ValueError(f'Arrays have incompatible shapes for broadcasting: {a.shape} and {b.shape}. '
-                               f'Original error: {str(e)}') from e
-    
-    return a * b
-
-@flush_mem
-@timing_decorator
-def ifft2d(radar_data: Union[np.ndarray, torch.Tensor], backend: str = 'numpy', verbose: bool = False) -> Union[np.ndarray, torch.Tensor]:
-    """Perform memory-efficient 2D inverse FFT on radar data.
-    
-    Args:
-        radar_data: Input radar data array.
-        backend: Backend to use ('numpy' or 'torch').
-        verbose: Whether to print verbose output.
-        
-    Returns:
-        Processed radar data after 2D inverse FFT.
-        
-    Raises:
-        ValueError: If backend is not supported.
-    """
-    if verbose:
-        print('Performing 2D inverse FFT...')
-    
-    # Inverse FFT along azimuth dimension first
-    if backend == 'numpy':
-        radar_data = np.fft.ifft(radar_data, axis=0)
-    elif backend == 'torch':
-        radar_data = torch.fft.ifft(radar_data, dim=0)
-    else:
-        raise ValueError(f'Unsupported backend: {backend}')
-    
-    # Then inverse FFT along range dimension
-    if backend == 'numpy':
-        radar_data = np.fft.ifftshift(np.fft.ifft(radar_data, axis=1), axes=1)
-    elif backend == 'torch':
-        radar_data = torch.fft.ifft(radar_data, dim=1)
-        radar_data = torch.fft.ifftshift(radar_data, dim=1)
-    else:
-        raise ValueError(f'Unsupported backend: {backend}')
-    
-    if verbose:
-        print(f'2D inverse FFT completed, data shape: {radar_data.shape}')
-        print_memory()
-    
-    return radar_data
-
-@flush_mem
-@timing_decorator
-def iff_azimuth(
-    radar_data: Union[np.ndarray, torch.Tensor], 
-    backend: str = 'numpy', 
-    verbose: bool = False
-) -> Union[np.ndarray, torch.Tensor]:
-    """Perform memory-efficient inverse FFT along azimuth dimension.
-    
-    Args:
-        radar_data: Input radar data array.
-        backend: Backend to use ('numpy' or 'torch').
-        verbose: Whether to print verbose output.
-        
-    Returns:
-        Processed radar data after inverse FFT along azimuth dimension.
-        
-    Raises:
-        ValueError: If backend is not supported.
-    """
-    if verbose:
-        print('Performing inverse FFT along azimuth dimension...')
-    
-    if backend == 'numpy':
-        radar_data = np.fft.ifft(radar_data, axis=0)
-    elif backend == 'torch':
-        radar_data = torch.fft.ifft(radar_data, dim=0)
-    else:
-        raise ValueError(f'Unsupported backend: {backend}')
-    
-    if verbose:
-        print(f'Inverse FFT along azimuth completed, data shape: {radar_data.shape}')
-        print_memory()
-    
-    return radar_data
-
-# -------- Processing Class ----------
-class CoarseRDA:
-    """Memory-efficient Coarse Range Doppler Algorithm processor for SAR data.
-    
-    This class implements a memory-optimized coarse Range Doppler Algorithm for processing
-    synthetic aperture radar (SAR) data, specifically designed for Sentinel-1 data.
-    
-    The processing pipeline follows these main steps:
-    1. Initialization and data loading
-    2. Transmission replica generation
-    3. 2D FFT transformation
-    4. Range compression
-    5. Range Cell Migration Correction (RCMC)
-    6. Azimuth compression
-    7. Final inverse transformations
-    
-    Memory optimizations include:
-    - In-place operations where possible
-    - Explicit memory cleanup
-    - Memory usage monitoring
-    - Efficient data type usage
-    """
-
-    # ==================== INITIALIZATION METHODS ====================
-    
-    def __init__(
-        self, 
-        raw_data: Dict[str, Any], 
-        verbose: bool = False, 
-        backend: str = 'numpy',
-        memory_efficient: bool = True
-    ) -> None:
-        """Initialize the CoarseRDA processor.
-        
-        Args:
-            raw_data: Dictionary containing 'echo', 'ephemeris', and 'metadata'.
-            verbose: Whether to print verbose output.
-            backend: Backend to use ('numpy', 'torch', or 'custom').
-            memory_efficient: Whether to enable memory-efficient mode.
-            
-        Raises:
-            ValueError: If invalid backend is specified.
-            AssertionError: If required data is missing.
-        """
-        self._validate_inputs(raw_data, backend)
-        self._initialize_settings(verbose, backend, memory_efficient)
-        self._load_data(raw_data)
-        self._setup_backend()
-        self._initialize_transmission_replica()
-
-    def _validate_inputs(self, raw_data: Dict[str, Any], backend: str) -> None:
-        """Validate input parameters.
-        
-        Args:
-            raw_data: Dictionary containing radar data.
-            backend: Processing backend.
-            
-        Raises:
-            AssertionError: If required data is missing.
-            ValueError: If invalid backend is specified.
-        """
-        assert isinstance(raw_data, dict), 'raw_data must be a dictionary'
-        assert 'echo' in raw_data, 'raw_data must contain "echo" key'
-        assert 'ephemeris' in raw_data, 'raw_data must contain "ephemeris" key'
-        assert 'metadata' in raw_data, 'raw_data must contain "metadata" key'
-        
-        valid_backends = {'numpy', 'torch', 'custom'}
-        if backend not in valid_backends:
-            raise ValueError(f'Backend must be one of {valid_backends}, got {backend}')
-
-    def _initialize_settings(self, verbose: bool, backend: str, memory_efficient: bool) -> None:
-        """Initialize processor settings.
-        
-        Args:
-            verbose: Whether to print verbose output.
-            backend: Processing backend.
-            memory_efficient: Whether to enable memory-efficient mode.
-        """
-        self._backend = backend
-        self._verbose = verbose
-        self._memory_efficient = memory_efficient
-        
-        if self._verbose:
-            print(f'Memory efficient mode: {"enabled" if memory_efficient else "disabled"}')
-
-    def _load_data(self, raw_data: Dict[str, Any]) -> None:
-        """Load and preprocess input data with memory optimization.
-        
-        Args:
-            raw_data: Dictionary containing radar data.
-        """
-        # Use views instead of copies where possible
-        self.radar_data = raw_data['echo']
-        
-        # Create a copy for ephemeris as we need to modify it
-        self.ephemeris = raw_data['ephemeris'].copy()
-        self.ephemeris['time_stamp'] /= 2**24
-        
-        # Use reference for metadata as we don't modify it
-        self.metadata = raw_data['metadata']
-        
-        # Initialize dimensions - these should remain constant throughout processing
-        self.len_az_line, self.len_range_line = self.radar_data.shape
-        
-        if self._verbose:
-            print(f'Loaded radar data with shape: {self.radar_data.shape}')
-            print(f'Azimuth lines: {self.len_az_line}, Range lines: {self.len_range_line}')
-            print(f'Data type: {self.radar_data.dtype}')
-            print(f'Memory usage: {self.radar_data.nbytes / 1024**3:.2f} GB')
-
-    def _setup_backend(self) -> None:
-        """Set up processing backend and device configuration."""
-        if self._backend == 'torch':
-            if torch is None:
-                raise ImportError('PyTorch is required for torch backend but not available')
-            self.device = getattr(
-                self.radar_data, 
-                'device', 
-                torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-            )
-            if self._verbose:
-                print(f'Selected device: {self.device}')
-
-    def _initialize_transmission_replica(self) -> None:
-        """Initialize transmission replica during setup."""
-        self._generate_tx_replica()
-
-    # ==================== TRANSMISSION REPLICA METHODS ====================
-
-    @timing_decorator
-    def _generate_tx_replica(self) -> None:
-        """Generate transmission replica based on metadata parameters.
-        
-        This method creates the transmission replica used for range compression
-        based on the radar system parameters extracted from metadata.
-        """
-        if self._verbose:
-            print('Generating transmission replica...')
-        
-        # Extract range decimation and calculate sample frequency
-        rgdec = int(self.metadata['range_decimation'].unique()[0])
-        if self._verbose:
-            print(f'Range decimation code: {rgdec}')
-        
-        self.range_sample_freq = range_dec_to_sample_rate(rgdec)
-        if self._verbose:
-            print(f'Range sample frequency: {self.range_sample_freq:.2f} Hz')
-
-        # Extract transmission parameters
-        tx_params = self._extract_tx_parameters()
-        
-        # Generate replica signal
-        self._create_replica_signal(tx_params)
-        
-        if self._verbose:
-            print(f'Replica length: {self.replica_len}')
-            print('Transmission replica generated successfully!')
-
-    def _extract_tx_parameters(self) -> Dict[str, float]:
-        """Extract transmission parameters from metadata.
-        
-        Returns:
-            Dictionary containing transmission parameters.
-        """
-        txpsf = self.metadata['tx_pulse_start_freq'].unique()[0]
-        txprr = self.metadata['tx_ramp_rate'].unique()[0]
-        txpl = self.metadata['tx_pulse_length'].unique()[0]
-        
-        if self._verbose:
-            print(f'TX pulse start frequency: {txpsf:.2f} Hz')
-            print(f'TX ramp rate: {txprr:.2f} Hz/s')
-            print(f'TX pulse length: {txpl:.6f} s')
-        
-        return {'start_freq': txpsf, 'ramp_rate': txprr, 'pulse_length': txpl}
-
-    def _create_replica_signal(self, tx_params: Dict[str, float]) -> None:
-        """Create the replica signal from transmission parameters.
-        
-        Args:
-            tx_params: Dictionary containing transmission parameters.
-        """
-        txpsf = tx_params['start_freq']
-        txprr = tx_params['ramp_rate'] 
-        txpl = tx_params['pulse_length']
-        
-        # Generate replica
-        self.num_tx_vals = int(txpl * self.range_sample_freq)
-        if self._verbose:
-            print(f'Number of TX values: {self.num_tx_vals}')
-        
-        tx_replica_time_vals = np.linspace(-txpl/2, txpl/2, num=self.num_tx_vals)
-        phi1 = txpsf + txprr * txpl / 2
-        phi2 = txprr / 2
-        
-        if self._verbose:
-            print(f'Phase parameters - phi1: {phi1:.2f}, phi2: {phi2:.2e}')
-        
-        self.tx_replica = np.exp(
-            2j * np.pi * (phi1 * tx_replica_time_vals + phi2 * tx_replica_time_vals**2)
-        )
-        self.replica_len = len(self.tx_replica)
-
-    # ==================== FFT METHODS ====================
-
-    @flush_mem
-    @timing_decorator
-    def fft2d(self, w_pad: Optional[int] = None, executors: int = 12) -> None:
-        """Perform memory-efficient 2D FFT on radar data in range and azimuth dimensions.
-
-        Args:
-            w_pad: Width padding for range FFT (ignored for dimension preservation).
-            executors: Number of executors for custom backend.
-            
-        Raises:
-            ValueError: If backend is not supported.
-        """
-        if self._verbose:
-            print(f'FFT input data shape: {self.radar_data.shape}')
-            print_memory()
-        
-        if self._backend == 'numpy':
-            self._fft2d_numpy_efficient()
-        elif self._backend == 'custom':
-            self._fft2d_custom(executors)
-        elif self._backend == 'torch':
-            self._fft2d_torch_efficient()
-        else:
-            raise ValueError(f'Backend {self._backend} not supported')
-        
-        # Verify dimensions are preserved
-        expected_shape = (self.len_az_line, self.len_range_line)
-        if self.radar_data.shape != expected_shape:
-            raise RuntimeError(f'FFT changed radar data shape from {expected_shape} to {self.radar_data.shape}')
-        
-        if self._verbose:
-            print(f'FFT output data shape: {self.radar_data.shape}')
-            print('- FFT performed successfully!')
-            print_memory()
-
-    def _fft2d_numpy_efficient(self) -> None:
-        """Perform memory-efficient 2D FFT using NumPy backend preserving original dimensions.
-        
-        Uses in-place operations and memory cleanup for better efficiency.
-        """
-        # Store original shape for verification
-        original_shape = self.radar_data.shape
-        if self._verbose:
-            print(f'Original radar data shape: {original_shape}')
-        
-        # Ensure data is contiguous and maintain original precision
-        if not self.radar_data.flags.c_contiguous:
-            if self._verbose:
-                print('Making data contiguous...')
-            self.radar_data = np.ascontiguousarray(self.radar_data)
-        
-        # FFT each range line (axis=1) - EXACT SAME as original
-        if self._verbose:
-            print(f'Performing FFT along range dimension (axis=1)...')
-        
-        # Use same approach as original - no dtype changes
-        self.radar_data = np.fft.fft(self.radar_data, axis=1)
-        
-        if self._verbose:
-            print(f'First FFT along range dimension completed, shape: {self.radar_data.shape}')
-            print_memory()
-        
-        # FFT each azimuth line (axis=0) with fftshift - EXACT SAME as original
-        if self._verbose:
-            print(f'Performing FFT along azimuth dimension (axis=0) with fftshift...')
-        
-        # Use same approach as original
-        self.radar_data = np.fft.fftshift(np.fft.fft(self.radar_data, axis=0), axes=0)
-        
-        if self._verbose:
-            print(f'Second FFT along azimuth dimension completed, shape: {self.radar_data.shape}')
-            print_memory()
-        
-        # Verify shape preservation
-        assert self.radar_data.shape == original_shape, \
-            f'FFT changed shape from {original_shape} to {self.radar_data.shape}'
-
-    def _fft2d_torch_efficient(self) -> None:
-        """Perform memory-efficient 2D FFT using PyTorch backend preserving dimensions.
-        
-        Uses in-place operations where possible.
-        """
-        original_shape = self.radar_data.shape
-        
-        if self._verbose:
-            print('Performing memory-efficient PyTorch FFT...')
-            print_memory()
-        
-        # FFT each range line (axis=1) - in-place when possible
-        if self._memory_efficient:
-            temp = torch.fft.fft(self.radar_data, dim=1)
-            self.radar_data.copy_(temp)
-            del temp
-            torch.cuda.empty_cache() if torch.cuda.is_available() else None
-        else:
-            self.radar_data = torch.fft.fft(self.radar_data, dim=1)
-        
-        # FFT each azimuth line (axis=0) with fftshift
-        if self._memory_efficient:
-            temp = torch.fft.fft(self.radar_data, dim=0)
-            self.radar_data.copy_(temp)
-            del temp
-            
-            temp = torch.fft.fftshift(self.radar_data, dim=0)
-            self.radar_data.copy_(temp)
-            del temp
-            torch.cuda.empty_cache() if torch.cuda.is_available() else None
-        else:
-            self.radar_data = torch.fft.fftshift(
-                torch.fft.fft(self.radar_data, dim=0), 
-                dim=0
-            )
-        
-        # Verify shape preservation
-        assert self.radar_data.shape == original_shape, \
-            f'Torch FFT changed shape from {original_shape} to {self.radar_data.shape}'
-
-    @flush_mem
-    @timing_decorator
-    def ifft_range(self) -> None:
-        """Perform memory-efficient inverse FFT along range dimension."""
-        if self._backend == 'numpy':
-            # Use EXACT SAME approach as original
-            self.radar_data = np.fft.ifftshift(np.fft.ifft(self.radar_data, axis=1), axes=1)
-        elif self._backend == 'torch':
-            self.radar_data = torch.fft.ifft(self.radar_data, dim=1)
-            self.radar_data = torch.fft.ifftshift(self.radar_data, dim=1)
-        else:
-            raise ValueError(f'Unsupported backend: {self._backend}')
-
-    @flush_mem
-    @timing_decorator
-    def ifft_azimuth(self) -> None:
-        """Perform memory-efficient inverse FFT along azimuth dimension."""
-        if self._backend == 'numpy':
-            # Use EXACT SAME approach as original
-            self.radar_data = np.fft.ifft(self.radar_data, axis=0)
-        elif self._backend == 'torch':
-            self.radar_data = torch.fft.ifft(self.radar_data, dim=0)
-        else:
-            raise ValueError(f'Unsupported backend: {self._backend}')
-
-    # ==================== FILTER GENERATION METHODS ====================
-
-    def _apply_window(self, signal: np.ndarray, window_type: str = 'hamming') -> np.ndarray:
-        """Apply windowing function to signal.
-        
-        Args:
-            signal: Input signal array.
-            window_type: Type of window ('hamming', 'hanning', 'blackman', 'rectangular').
-            
-        Returns:
-            Windowed signal array.
-            
-        Raises:
-            ValueError: If window type is not supported.
-        """
-        if window_type.lower() == 'rectangular' or window_type is None:
-            return signal
-        
-        signal_length = len(signal)
-        
-        if window_type.lower() == 'hamming':
-            window = np.hamming(signal_length)
-        elif window_type.lower() == 'hanning':
-            window = np.hanning(signal_length)
-        elif window_type.lower() == 'blackman':
-            window = np.blackman(signal_length)
-        else:
-            raise ValueError(f'Unsupported window type: {window_type}. '
-                           f'Supported types: hamming, hanning, blackman, rectangular')
-        
-        return signal * window
-
-    def _get_reference_function(self, replica_chirp: np.ndarray, window_type: str = 'hamming') -> np.ndarray:
-        """Generate reference function from replica chirp following C++ implementation.
-        
-        Args:
-            replica_chirp: Input replica chirp signal.
-            window_type: Type of window to apply.
-            
-        Returns:
-            Reference function for range compression.
-        """
-        # Apply windowing to replica chirp first (as in C++ implementation)
-        windowed_replica = self._apply_window(replica_chirp, window_type)
-        
-        # Calculate energy normalization (matching C++ implementation)
-        # C++ code: norm *= n / norm_size, then energy = accumulate(norm)
-        norm = np.abs(windowed_replica)
-        norm_size = len(norm)
-        norm_squared = (norm * norm) / norm_size  # This matches C++ norm *= n / norm_size
-        energy = np.sum(norm_squared)
-        
-        if self._verbose:
-            print(f'Calculated energy for normalization: {energy:.6f}')
-        
-        # Create reference function
-        reference = windowed_replica.copy()
-        
-        # Apply FFT and conjugate (matching C++ implementation)
-        reference = np.fft.fft(reference)
-        reference = np.conjugate(reference)
-        
-        # Normalize by energy
-        if energy > 0:
-            reference = reference / energy
-        else:
-            if self._verbose:
-                print('⚠️  Warning: Zero energy detected, skipping normalization')
-        
-        return reference
-
-    @flush_mem
-    @timing_decorator
-    def get_range_filter(self, pad_w: int = 0, window_type: str = 'hamming') -> np.ndarray:
-        """Compute memory-efficient range filter for radar data compression.
-    
-        Args:
-            pad_w: Width padding (ignored - filter always matches radar data dimensions).
-            window_type: Type of window to apply ('hamming', 'hanning', 'blackman', 'rectangular').
-            
-        Returns:
-            Range filter array exactly matching radar data range dimension.
-            
-        Raises:
-            AssertionError: If filter dimensions are invalid.
-            ValueError: If window type is not supported.
-        """
-        # Use exact radar data dimensions - no padding considerations
-        current_range_dim = self.radar_data.shape[1]
-        
-        if self._verbose:
-            print(f'Creating range filter for radar data shape: {self.radar_data.shape}')
-            print(f'Range dimension: {current_range_dim}')
-            print(f'TX replica length: {self.num_tx_vals}')
-            print(f'Window type: {window_type}')
-        
-        # Create range filter with exact radar data range dimension
-        range_filter = np.zeros(current_range_dim, dtype=complex)
-        
-        # Place replica in center of filter
-        if current_range_dim >= self.num_tx_vals:
-            index_start = (current_range_dim - self.num_tx_vals) // 2
-            index_end = index_start + self.num_tx_vals
-            
-            if self._verbose:
-                print(f'Placing replica at indices [{index_start}:{index_end}] in filter of length {current_range_dim}')
-            
-            range_filter[index_start:index_end] = self.tx_replica
-        else:
-            # If range dimension is smaller than replica, truncate replica
-            if self._verbose:
-                print(f'⚠️  Range dimension ({current_range_dim}) < replica length ({self.num_tx_vals}), truncating replica')
-            
-            replica_start = (self.num_tx_vals - current_range_dim) // 2
-            replica_end = replica_start + current_range_dim
-            range_filter[:] = self.tx_replica[replica_start:replica_end]
-        
-        # Generate reference function using C++ approach - apply windowing to the placed replica
-        range_filter = self._get_reference_function(range_filter, window_type)
-        
-        if self._verbose:
-            print(f'Range filter shape: {range_filter.shape}')
-            print(f'Range filter energy after normalization: {np.sum(np.abs(range_filter)**2):.6f}')
-        
-        # Ensure filter exactly matches radar data range dimension
-        assert range_filter.shape[0] == current_range_dim, \
-            f'Filter shape mismatch: expected {current_range_dim}, got {range_filter.shape[0]}'
-        
-        return range_filter
-
-    @flush_mem
-    @timing_decorator
-    def get_rcmc(self) -> np.ndarray:
-        """Calculate memory-efficient Range Cell Migration Correction filter.
-
-        Returns:
-            RCMC filter array matching radar data dimensions.
-        """
-        self._compute_effective_velocities()
-        
-        self.wavelength = cnst.TX_WAVELENGTH_M
-        
-        # Generate azimuth frequency values for the entire azimuth line length
-        self.az_freq_vals = np.arange(
-            -self.az_sample_freq/2, 
-            self.az_sample_freq/2, 
-            self.az_sample_freq/self.len_az_line
-        )
-        
-        # Ensure we have exactly the right number of frequency values
-        if len(self.az_freq_vals) != self.len_az_line:
-            self.az_freq_vals = np.linspace(
-                -self.az_sample_freq/2, 
-                self.az_sample_freq/2, 
-                self.len_az_line, 
-                endpoint=False
-            )
-        
-        if self._verbose:
-            print(f'Azimuth frequency values shape: {self.az_freq_vals.shape}')
-            print(f'Effective velocities shape: {self.effective_velocities.shape}')
-        
-        # Take mean effective velocity across range for each azimuth line
-        # This reduces from (56130, 25724) to (56130,)
-        mean_effective_velocities = np.mean(self.effective_velocities, axis=1)
-        
-        if self._verbose:
-            print(f'Mean effective velocities shape: {mean_effective_velocities.shape}')
-        
-        # Cosine of instantaneous squint angle
-        # Broadcasting: (56130,) with (56130,) -> (56130,)
-        self.D = np.sqrt(
-            1 - (self.wavelength**2 * self.az_freq_vals**2) / 
-                (4 * mean_effective_velocities**2)
-        )
-        
-        if self._verbose:
-            print(f'D (cosine squint angle) shape: {self.D.shape}')
-        
-        # Create RCMC filter with CURRENT radar data dimensions (should be original dimensions)
-        current_range_dim = self.radar_data.shape[1]
-        
-        range_freq_vals = np.linspace(
-            -self.range_sample_freq/2, 
-            self.range_sample_freq/2, 
-            num=current_range_dim
-        )
-        
-        if self._verbose:
-            print(f'Range frequency values shape: {range_freq_vals.shape}')
-            print(f'Current radar data range dimension: {current_range_dim}')
-            print(f'Slant range vec shape: {self.slant_range_vec.shape}')
-        
-        # Calculate RCMC shift - use first slant range value for reference
-        rcmc_shift = self.slant_range_vec[0] * (1/self.D - 1)
-        
-        if self._verbose:
-            print(f'RCMC shift shape: {rcmc_shift.shape}')
-        
-        # IMPORTANT: Use same calculation method as original - no chunked processing
-        # Broadcasting for final filter calculation
-        range_freq_2d = range_freq_vals[np.newaxis, :]  # (1, current_range_dim)
-        rcmc_shift_2d = rcmc_shift[:, np.newaxis]       # (56130, 1)
-        
-        rcmc_filter = np.exp(4j * np.pi * range_freq_2d * rcmc_shift_2d / self.c)
-        
-        if self._verbose:
-            print(f'Final RCMC filter shape: {rcmc_filter.shape}')
-        
-        return rcmc_filter
-
-    @flush_mem
-    @timing_decorator
-    def get_azimuth_filter(self) -> np.ndarray:
-        """Calculate memory-efficient azimuth compression filter.
-        
-        Returns:
-            Azimuth filter array matching radar data dimensions.
-        """
-        if self._verbose:
-            print(f'Computing azimuth filter...')
-            print(f'Slant range vec shape: {self.slant_range_vec.shape}')
-            print(f'D shape: {self.D.shape}')
-            print(f'Wavelength: {self.wavelength}')
-        
-        # Use current radar data dimensions (should match original slant range vector)
-        current_range_dim = self.radar_data.shape[1]
-        
-        # Ensure slant range vector matches radar data dimensions
-        if current_range_dim != len(self.slant_range_vec):
-            if self._verbose:
-                print(f'⚠️  Warning: Current range dim ({current_range_dim}) != slant range vec length ({len(self.slant_range_vec)})')
-                print(f'This should not happen - using original slant range vector')
-            current_slant_range_vec = self.slant_range_vec
-        else:
-            current_slant_range_vec = self.slant_range_vec
-        
-        # IMPORTANT: Use same calculation method as original - no chunked processing
-        # Broadcasting for azimuth filter calculation
-        slant_range_2d = current_slant_range_vec[np.newaxis, :]  # (1, range_dim)
-        D_2d = self.D[:, np.newaxis]                             # (56130, 1)
-        
-        azimuth_filter = np.exp(4j * np.pi * slant_range_2d * D_2d / self.wavelength)
-        
-        if self._verbose:
-            print(f'Azimuth filter shape: {azimuth_filter.shape}')
-        
-        return azimuth_filter
-
-    # ==================== VELOCITY COMPUTATION METHODS ====================
-
-    @timing_decorator
-    def _compute_effective_velocities(self) -> None:
-        """Calculate effective spacecraft velocities for processing.
-        
-        This method computes the effective velocities needed for RCMC and
-        azimuth compression by combining spacecraft and ground velocities.
-        """
-        # Initialize timing and geometry parameters
-        self._initialize_timing_parameters()
-        
-        # Calculate spacecraft velocities and positions
-        space_velocities, positions = self._calculate_spacecraft_dynamics()
-        
-        # Compute effective velocities using Earth model
-        self._compute_ground_velocities(space_velocities, positions)
-
-    def _initialize_timing_parameters(self) -> None:
-        """Initialize timing and geometry parameters for velocity computation.
-        
-        Raises:
-            KeyError: If required metadata columns are missing.
-            ValueError: If metadata values are invalid.
-        """
-        self.c = cnst.SPEED_OF_LIGHT_MPS
-        
-        # Check for required columns with case-insensitive matching
-        metadata_columns = {col.lower(): col for col in self.metadata.columns}
-        
-        required_mappings = {
-            'pri': ['pri', 'pulse_repetition_interval'],
-            'rank': ['rank'],
-            'swst': ['swst', 'sampling_window_start_time', 'start_time']
-        }
-        
-        column_map = {}
-        for param, possible_names in required_mappings.items():
-            found_column = None
-            for name in possible_names:
-                if name.lower() in metadata_columns:
-                    found_column = metadata_columns[name.lower()]
-                    break
-            
-            if found_column is None:
-                available_cols = list(self.metadata.columns)
-                raise KeyError(
-                    f'Could not find column for {param}. Tried: {possible_names}. '
-                    f'Available columns: {available_cols}'
-                )
-            column_map[param] = found_column
-        
-        if self._verbose:
-            print('Column mapping:')
-            for param, col in column_map.items():
-                print(f'  {param} -> {col}')
-        
-        # Extract parameters with error handling
-        try:
-            self.pri = self.metadata[column_map['pri']].iloc[0]
-            rank = self.metadata[column_map['rank']].iloc[0] 
-            range_start_time_base = self.metadata[column_map['swst']].iloc[0]
-        except (IndexError, TypeError) as e:
-            raise ValueError(f'Error extracting metadata values: {str(e)}') from e
-        
-        # Validate values
-        assert self.pri > 0, f'PRI must be positive, got {self.pri}'
-        assert rank >= 0, f'Rank must be non-negative, got {rank}'
-        
-        if self._verbose:
-            print(f'PRI: {self.pri:.6f} s')
-            print(f'Rank: {rank}')
-            print(f'Base range start time: {range_start_time_base:.6f} s')
-        
-        # Calculate derived parameters
-        suppressed_data_time = 320 / (8 * cnst.F_REF)
-        range_start_time = range_start_time_base + suppressed_data_time
-        
-        # Sample rates
-        range_sample_period = 1 / self.range_sample_freq
-        self.az_sample_freq = 1 / self.pri
-        
-        if self._verbose:
-            print(f'Range start time: {range_start_time:.6f} s')
-            print(f'Azimuth sample frequency: {self.az_sample_freq:.2f} Hz')
-        
-        # Fast time and slant range vectors
-        sample_num_along_range_line = np.arange(0, self.len_range_line, 1)
-        fast_time_vec = range_start_time + (range_sample_period * sample_num_along_range_line)
-        self.slant_range_vec = ((rank * self.pri) + fast_time_vec) * self.c / 2
-        
-        if self._verbose:
-            print(f'Slant range vector shape: {self.slant_range_vec.shape}')
-            print(f'Slant range min/max: {self.slant_range_vec.min():.2f}/{self.slant_range_vec.max():.2f} m')
-
-    def _calculate_spacecraft_dynamics(self) -> Tuple[np.ndarray, np.ndarray]:
-        """Calculate spacecraft velocities and positions.
-        
-        Returns:
-            Tuple of (space_velocities, positions) arrays.
-            
-        Raises:
-            KeyError: If required columns are missing from ephemeris or metadata.
-            ValueError: If interpolation fails.
-        """
-        if self._verbose:
-            print('Calculating spacecraft dynamics...')
-            print(f'Ephemeris shape: {self.ephemeris.shape}')
-            print(f'Metadata shape: {self.metadata.shape}')
-        
-        # Spacecraft velocity calculations
-        ecef_vels = self.ephemeris.apply(
-            lambda x: math.sqrt(x['vx']**2 + x['vy']**2 + x['vz']**2), 
-            axis=1
-        )
-        
-        if self._verbose:
-            print(f'ECEF velocities shape: {ecef_vels.shape}')
-            print(f'ECEF velocities range: {ecef_vels.min():.2f} - {ecef_vels.max():.2f} m/s')
-        
-        # Extract arrays and ensure they are proper numpy arrays
-        time_stamps = self.ephemeris['time_stamp'].values
-        velocity_values = ecef_vels.values
-        x_values = self.ephemeris['x'].values
-        y_values = self.ephemeris['y'].values  
-        z_values = self.ephemeris['z'].values
-        
-        if self._verbose:
-            print(f'Time stamps shape: {time_stamps.shape}')
-            print(f'Time stamps range: {time_stamps.min():.6f} - {time_stamps.max():.6f}')
-            print(f' veocity values shape: {velocity_values.shape}')
-            print(f'Position arrays shapes: x={x_values.shape}, y={y_values.shape}, z={z_values.shape}')
-        
-        # Ensure arrays are sorted by time for interpolation
-        sort_indices = np.argsort(time_stamps)
-        time_stamps = time_stamps[sort_indices]
-        velocity_values = velocity_values[sort_indices]
-        x_values = x_values[sort_indices]
-        y_values = y_values[sort_indices]
-        z_values = z_values[sort_indices]
-        
-        # Calculate metadata time stamps
-        metadata_times = self.metadata.apply(
-            lambda x: x['coarse_time'] + x['fine_time'], 
-            axis=1
-        ).values
-        
-        if self._verbose:
-            print(f'Metadata times shape: {metadata_times.shape}')
-            print(f'Metadata times range: {metadata_times.min():.6f} - {metadata_times.max():.6f}')
-            print(f'Ephemeris time range: {time_stamps.min():.6f} - {time_stamps.max():.6f}')
-        
-        # Check if metadata times are within ephemeris time range
-        time_within_range = (metadata_times >= time_stamps.min()) & (metadata_times <= time_stamps.max())
-        if not np.all(time_within_range):
-            out_of_range_count = np.sum(~time_within_range)
-            if self._verbose:
-                print(f'⚠️  Warning: {out_of_range_count} metadata times are outside ephemeris range')
-                print(f'   Will use boundary values for extrapolation')
-        
-        # Create interpolation functions with bounds_error=False and fill_value for extrapolation
-        try:
-            velocity_interp = interp1d(
-                time_stamps, velocity_values, 
-                kind='linear', 
-                bounds_error=False, 
-                fill_value=(velocity_values[0], velocity_values[-1])
-            )
-            x_interp = interp1d(
-                time_stamps, x_values, 
-                kind='linear', 
-                bounds_error=False, 
-                fill_value=(x_values[0], x_values[-1])
-            )
-            y_interp = interp1d(
-                time_stamps, y_values, 
-                kind='linear', 
-                bounds_error=False, 
-                fill_value=(y_values[0], y_values[-1])
-            )
-            z_interp = interp1d(
-                time_stamps, z_values, 
-                kind='linear', 
-                bounds_error=False, 
-                fill_value=(z_values[0], z_values[-1])
-            )
-        except ValueError as e:
-            raise ValueError(f'Failed to create interpolation functions: {str(e)}') from e
-        
-        # Interpolate at metadata time points
-        try:
-            space_velocities = velocity_interp(metadata_times)
-            x_interp_vals = x_interp(metadata_times)
-            y_interp_vals = y_interp(metadata_times)
-            z_interp_vals = z_interp(metadata_times)
-        except Exception as e:
-            raise ValueError(f'Interpolation failed: {str(e)}') from e
-        
-        # Ensure interpolated values are arrays and handle any remaining NaN values
-        space_velocities = np.asarray(space_velocities)
-        x_interp_vals = np.asarray(x_interp_vals)
-        y_interp_vals = np.asarray(y_interp_vals)
-        z_interp_vals = np.asarray(z_interp_vals)
-        
-        # Check for and handle NaN values
-        if np.any(np.isnan(space_velocities)):
-            nan_count = np.sum(np.isnan(space_velocities))
-            if self._verbose:
-                print(f'⚠️  Found {nan_count} NaN values in space_velocities, filling with nearest valid values')
-            
-            # Fill NaN values with nearest valid values
-            valid_mask = ~np.isnan(space_velocities)
-            if np.any(valid_mask):
-                # Forward fill
-                space_velocities = pd.Series(space_velocities).fillna(method='ffill').fillna(method='bfill').values
-            else:
-                # If all NaN, use average ephemeris velocity
-                space_velocities.fill(np.nanmean(velocity_values))
-        
-        # Handle NaN values in position components
-        for vals, name in [(x_interp_vals, 'x'), (y_interp_vals, 'y'), (z_interp_vals, 'z')]:
-            if np.any(np.isnan(vals)):
-                nan_count = np.sum(np.isnan(vals))
-                if self._verbose:
-                    print(f'⚠️  Found {nan_count} NaN values in {name}_interp_vals, filling with nearest valid values')
-                
-                valid_mask = ~np.isnan(vals)
-                if np.any(valid_mask):
-                    vals_series = pd.Series(vals).fillna(method='ffill').fillna(method='bfill')
-                    if name == 'x':
-                        x_interp_vals = vals_series.values
-                    elif name == 'y':
-                        y_interp_vals = vals_series.values
-                    else:
-                        z_interp_vals = vals_series.values
-        
-        if self._verbose:
-            print(f'Interpolated space_velocities shape: {space_velocities.shape}')
-            print(f'Interpolated position component shapes: x={x_interp_vals.shape}, y={y_interp_vals.shape}, z={z_interp_vals.shape}')
-        
-        # Create position array
-        positions = np.column_stack([x_interp_vals, y_interp_vals, z_interp_vals])
-        
-        if self._verbose:
-            print(f'Final space_velocities shape: {space_velocities.shape}')
-            print(f'Final positions shape: {positions.shape}')
-            print(f'Space velocities range: {space_velocities.min():.2f} - {space_velocities.max():.2f} m/s')
-            print(f'Position range - x: {positions[:, 0].min():.0f} to {positions[:, 0].max():.0f}')
-            print(f'Position range - y: {positions[:, 1].min():.0f} to {positions[:, 1].max():.0f}')
-            print(f'Position range - z: {positions[:, 2].min():.0f} to {positions[:, 2].max():.0f}')
-        
-        # Validate outputs
-        assert isinstance(space_velocities, np.ndarray), 'space_velocities must be numpy array'
-        assert isinstance(positions, np.ndarray), 'positions must be numpy array'
-        assert len(space_velocities.shape) == 1, f'space_velocities must be 1D, got shape {space_velocities.shape}'
-        assert len(positions.shape) == 2, f'positions must be 2D array, got shape {positions.shape}'
-        assert positions.shape[1] == 3, f'positions must have 3 columns (x,y,z), got {positions.shape[1]}'
-        assert space_velocities.shape[0] == positions.shape[0], f'velocity and position arrays must have same length'
-        
-        # Final check for NaN values after cleaning
-        assert not np.any(np.isnan(space_velocities)), 'NaN values still present in space_velocities after cleaning'
-        assert not np.any(np.isnan(positions)), 'NaN values still present in positions after cleaning'
-        
-        # Check for reasonable values
-        assert np.all(space_velocities > 1000), f'Space velocities too low (min: {space_velocities.min():.2f} m/s)'
-        assert np.all(space_velocities < 20000), f'Space velocities too high (max: {space_velocities.max():.2f} m/s)'
-        
-        position_magnitudes = np.linalg.norm(positions, axis=1)
-        assert np.all(position_magnitudes > 6e6), f'Position magnitudes too small (min: {position_magnitudes.min():.0f} m)'
-        assert np.all(position_magnitudes < 8e6), f'Position magnitudes too large (max: {position_magnitudes.max():.0f} m)'
-        
-        return space_velocities, positions
-
-    def _compute_ground_velocities(self, space_velocities: np.ndarray, positions: np.ndarray) -> None:
-        """Compute ground velocities and effective velocities.
-        
-        Args:
-            space_velocities: Spacecraft velocity magnitudes (1D array).
-            positions: Spacecraft position vectors (2D array, shape [N, 3]).
-            
-        Raises:
-            AssertionError: If input arrays have incompatible shapes.
-            ValueError: If calculations produce invalid results.
-        """
-        # Enhanced input validation
-        assert isinstance(space_velocities, np.ndarray), f'space_velocities must be numpy array, got {type(space_velocities)}'
-        assert isinstance(positions, np.ndarray), f'positions must be numpy array, got {type(positions)}'
-        assert len(positions.shape) == 2, f'positions must be 2D, got shape {positions.shape}'
-        assert positions.shape[1] == 3, f'positions must have 3 columns, got {positions.shape[1]}'
-        assert space_velocities.shape[0] == positions.shape[0], f'Array lengths must match: velocities={space_velocities.shape[0]}, positions={positions.shape[0]}'
-        
-        # Ensure arrays are proper numpy arrays with correct dtypes
-        space_velocities = np.asarray(space_velocities, dtype=np.float64)
-        positions = np.asarray(positions, dtype=np.float64)
-        
-        # Check for NaN/inf values
-        assert not np.any(np.isnan(space_velocities)), 'NaN values in space_velocities'
-        assert not np.any(np.isnan(positions)), 'NaN values in positions'
-        assert not np.any(np.isinf(space_velocities)), 'Infinite values in space_velocities'
-        assert not np.any(np.isinf(positions)), 'Infinite values in positions'
-        
-        if self._verbose:
-            print('Computing ground velocities...')
-            print(f'Space velocities shape: {space_velocities.shape}')
-            print(f'Positions shape: {positions.shape}')
-            print(f'Slant range vec shape: {self.slant_range_vec.shape}')
-            print(f'Input data ranges:')
-            print(f'  Space velocities: {space_velocities.min():.2f} - {space_velocities.max():.2f} m/s')
-            print(f'  Positions X: {positions[:, 0].min():.0f} - {positions[:, 0].max():.0f} m')
-            print(f'  Positions Y: {positions[:, 1].min():.0f} - {positions[:, 1].max():.0f} m')
-            print(f'  Positions Z: {positions[:, 2].min():.0f} - {positions[:, 2].max():.0f} m')
-        
-        # Earth model calculations
-        a = float(cnst.WGS84_SEMI_MAJOR_AXIS_M)
-        b = float(cnst.WGS84_SEMI_MINOR_AXIS_M)
-        
-        if self._verbose:
-            print(f'Earth model parameters: a={a:.0f} m, b={b:.0f} m')
-        
-        # Calculate spacecraft heights (magnitudes of position vectors)
-        H = np.linalg.norm(positions, axis=1)  # axis=1 for row-wise norm
-        H = np.asarray(H, dtype=np.float64)
-        
-        # Validate H calculation
-        assert H.shape == space_velocities.shape, f'H shape {H.shape} must match velocities shape {space_velocities.shape}'
-        assert not np.any(np.isnan(H)), 'NaN values in H (spacecraft heights)'
-        assert np.all(H > 0), 'All spacecraft heights must be positive'
-        
-        W = space_velocities / H
-        W = np.asarray(W, dtype=np.float64)
-        
-        # Calculate latitude using more robust method
-        xy_distance = np.sqrt(positions[:, 0]**2 + positions[:, 1]**2)
-        xy_distance = np.asarray(xy_distance, dtype=np.float64)
-        lat = np.arctan2(positions[:, 2], xy_distance)
-        lat = np.asarray(lat, dtype=np.float64)
-        
-        if self._verbose:
-            print(f'H (heights) shape: {H.shape}, range: {H.min():.0f} - {H.max():.0f} m')
-            print(f'W (angular velocities) shape: {W.shape}, range: {W.min():.6f} - {W.max():.6f} rad/s')
-            print(f'Latitudes range: {np.degrees(lat.min()):.2f} - {np.degrees(lat.max()):.2f} deg')
-        
-        # Validate intermediate calculations
-        assert not np.any(np.isnan(W)), 'NaN values in W (angular velocities)'
-        assert not np.any(np.isnan(lat)), 'NaN values in latitude'
-        
-        # Local Earth radius calculation with explicit numpy array operations
-        cos_lat = np.cos(lat)
-        sin_lat = np.sin(lat)
-        cos_lat = np.asarray(cos_lat, dtype=np.float64)
-        sin_lat = np.asarray(sin_lat, dtype=np.float64)
-        
-        # Ensure all terms are numpy arrays before sqrt operation
-        numerator = np.asarray(a**4 * cos_lat**2 + b**4 * sin_lat**2, dtype=np.float64)
-        denominator = np.asarray(a**2 * cos_lat**2 + b**2 * sin_lat**2, dtype=np.float64)
-        ratio = numerator / denominator
-        ratio = np.asarray(ratio, dtype=np.float64)
-        
-        local_earth_rad = np.sqrt(ratio)
-        local_earth_rad = np.asarray(local_earth_rad, dtype=np.float64)
-        
-        if self._verbose:
-            print(f'Local Earth radius range: {local_earth_rad.min():.0f} - {local_earth_rad.max():.0f} m')
-        
-        # Validate local Earth radius
-        assert not np.any(np.isnan(local_earth_rad)), 'NaN values in local_earth_rad'
-        assert np.all(local_earth_rad > 0), 'All local Earth radii must be positive'
-        
-        # Ensure slant_range_vec is also a proper numpy array
-        slant_range_vec = np.asarray(self.slant_range_vec, dtype=np.float64)
-        
-        # Broadcasting for slant range calculation
-        slant_range_2d = slant_range_vec[np.newaxis, :]  # Shape: [1, range_samples]
-        local_earth_rad_2d = local_earth_rad[:, np.newaxis]  # Shape: [azimuth_samples, 1]
-        H_2d = H[:, np.newaxis]  # Shape: [azimuth_samples, 1]
-        W_2d = W[:, np.newaxis]  # Shape: [azimuth_samples, 1]
-        
-        # Calculate cosine of look angle with explicit array operations
-        term1 = np.asarray(local_earth_rad_2d**2, dtype=np.float64)
-        term2 = np.asarray(H_2d**2, dtype=np.float64)
-        term3 = np.asarray(slant_range_2d**2, dtype=np.float64)
-        term4 = np.asarray(2 * local_earth_rad_2d * H_2d, dtype=np.float64)
-        
-        cos_beta = (term1 + term2 - term3) / term4
-        cos_beta = np.asarray(cos_beta, dtype=np.float64)
-        
-        # Clip to valid range for cosine
-        cos_beta = np.clip(cos_beta, -1.0, 1.0)
-        
-        if self._verbose:
-            print(f'cos_beta shape: {cos_beta.shape}')
-            print(f'cos_beta range: {cos_beta.min():.3f} - {cos_beta.max():.3f}')
-        
-        # Calculate ground velocities
-        ground_velocities = local_earth_rad_2d * W_2d * cos_beta
-        ground_velocities = np.asarray(ground_velocities, dtype=np.float64)
-        
-        if self._verbose:
-            print(f'Ground velocities shape: {ground_velocities.shape}')
-        
-        # Calculate effective velocities
-        space_velocities_2d = space_velocities[:, np.newaxis]  # Shape: [azimuth_samples, 1]
-        effective_vel_product = space_velocities_2d * ground_velocities
-        effective_vel_product = np.asarray(effective_vel_product, dtype=np.float64)
-        
-        # Ensure non-negative values before sqrt
-        effective_vel_product = np.maximum(effective_vel_product, 0.0)
-        
-        self.effective_velocities = np.sqrt(effective_vel_product)
-        self.effective_velocities = np.asarray(self.effective_velocities, dtype=np.float64)
-        
-        if self._verbose:
-            print(f'Effective velocities shape: {self.effective_velocities.shape}')
-            print(f'Effective velocities range: {self.effective_velocities.min():.2f} - {self.effective_velocities.max():.2f} m/s')
-        
-        # Final validation
-        assert not np.any(np.isnan(self.effective_velocities)), 'NaN values in effective velocities'
-        assert not np.any(np.isinf(self.effective_velocities)), 'Infinite values in effective velocities'
-        assert np.all(self.effective_velocities >= 0), 'All effective velocities must be non-negative'
-
-    
-
-    # ==================== MAIN PROCESSING METHODS ====================
-
-    @flush_mem
-    @timing_decorator
-    def data_focus(self) -> None:
-        """Perform memory-efficient complete SAR data focusing using Range Doppler Algorithm.
-        
-        This method orchestrates the complete SAR focusing process by calling
-        three main processing steps in sequence.
-        
-        Raises:
-            RuntimeError: If data dimensions change unexpectedly during processing.
-        """
-        if self._verbose:
-            print('Starting memory-efficient SAR data focusing...')
-            print(f'Initial radar data shape: {self.radar_data.shape}')
-            print_memory()
-        
-        # Store initial shape for verification
-        initial_shape = self.radar_data.shape
-        expected_shape = (self.len_az_line, self.len_range_line)
-        
-        assert initial_shape == expected_shape, \
-            f'Initial data shape {initial_shape} does not match expected {expected_shape}'
-        
-        
-        self.raw_data = copy.deepcopy(self.radar_data)
-        if self._verbose:
-            print(f'Raw radar data shape: {self.raw_data.shape}')
-            print_memory()
-        # ------------------------------------------------------------------------
-        # Step 1: 2D FFT transformation (preserves dimensions)
-        self.fft2d()
-        assert self.radar_data.shape == initial_shape, \
-            f'FFT changed data shape from {initial_shape} to {self.radar_data.shape}'
-        # ------------------------------------------------------------------------
-        
-        # Step 2: Range compression
-        self.range_compression()
-        self.range_compressed_data = ifft2d(copy.deepcopy(self.radar_data))
-        if self._verbose:
-            print(f'Range compressed data shape: {self.radar_data.shape}')
-            print_memory()
-        # ------------------------------------------------------------------------
-
-        # Step 3: Range Cell Migration Correction
-        self.rcmc()
-        self.rcmc_data = iff_azimuth(copy.deepcopy(self.radar_data))
-        if self._verbose:
-            print(f'RCMC data shape: {self.radar_data.shape}')
-            print_memory()
-        # ------------------------------------------------------------------------
-        # Step 4: Azimuth compression
-        self.azimuth_compression()
-        self.azimuth_compressed_data = self.radar_data
-        if self._verbose:
-            print(f'SAR data focusing completed successfully!')
-            print(f'Final radar data shape: {self.radar_data.shape}')
-            print_memory()
-    
-    @flush_mem
-    @timing_decorator
-    def range_compression(self) -> None:
-        """Perform memory-efficient range compression step.
-        
-        This method applies the range compression filter to compress the radar
-        signal in the range dimension while preserving data dimensions.
-        
-        Raises:
-            RuntimeError: If data dimensions change unexpectedly during processing.
-        """
-        if self._verbose:
-            print('Starting range compression...')
-            print(f'Input radar data shape: {self.radar_data.shape}')
-            print_memory()
-        
-        # Store initial shape for verification
-        initial_shape = self.radar_data.shape
-        
-        # Legacy compatibility parameters
-        w_pad = 0
-        original_w = initial_shape[1]
-        
-        if self._verbose:
-            print(f'Processing with original_w={original_w}')
-        
-        # Perform range compression
-        self._perform_range_compression_efficient(w_pad, original_w)
-        
-        # Verify dimensions are preserved
-        assert self.radar_data.shape == initial_shape, \
-            f'Range compression changed data shape from {initial_shape} to {self.radar_data.shape}'
-        
-        if self._verbose:
-            print(f'Range compression completed successfully!')
-            print(f'Output radar data shape: {self.radar_data.shape}')
-            print_memory()
-    
-    @flush_mem
-    @timing_decorator
-    def rcmc(self) -> None:
-        """Perform memory-efficient Range Cell Migration Correction.
-        
-        This method applies the RCMC filter to correct for range cell migration
-        effects and performs inverse FFT in the range dimension.
-        
-        Raises:
-            RuntimeError: If data dimensions change unexpectedly during processing.
-        """
-        if self._verbose:
-            print('Starting Range Cell Migration Correction...')
-            print(f'Input radar data shape: {self.radar_data.shape}')
-            print_memory()
-        
-        # Store initial shape for verification
-        initial_shape = self.radar_data.shape
-        
-        # Perform RCMC
-        self._perform_rcmc_efficient()
-        
-        # Verify dimensions are preserved
-        assert self.radar_data.shape == initial_shape, \
-            f'RCMC changed data shape from {initial_shape} to {self.radar_data.shape}'
-        
-        if self._verbose:
-            print(f'RCMC completed successfully!')
-            print(f'Output radar data shape: {self.radar_data.shape}')
-            print_memory()
-    
-    @flush_mem
-    @timing_decorator
-    def azimuth_compression(self) -> None:
-        """Perform memory-efficient azimuth compression step.
-        
-        This method applies the azimuth compression filter to focus the radar
-        signal in the azimuth dimension and performs inverse FFT in azimuth.
-        
-        Raises:
-            RuntimeError: If data dimensions change unexpectedly during processing.
-        """
-        if self._verbose:
-            print('Starting azimuth compression...')
-            print(f'Input radar data shape: {self.radar_data.shape}')
-            print_memory()
-        
-        # Store initial shape for verification
-        initial_shape = self.radar_data.shape
-        
-        # Perform azimuth compression
-        self._perform_azimuth_compression_efficient()
-        
-        # Verify dimensions are preserved
-        assert self.radar_data.shape == initial_shape, \
-            f'Azimuth compression changed data shape from {initial_shape} to {self.radar_data.shape}'
-        
-        if self._verbose:
-            print(f'Azimuth compression completed successfully!')
-            print(f'Output radar data shape: {self.radar_data.shape}')
-            print_memory()
-
-    def _perform_range_compression_efficient(self, w_pad: int, original_w: int) -> None:
-        """Perform memory-efficient range compression step while preserving data dimensions.
-        
-        Args:
-            w_pad: Width padding (ignored - dimensions preserved).
-            original_w: Original width (for verification).
-            
-        Raises:
-            ValueError: If array shapes are incompatible.
-            AssertionError: If dimensions change unexpectedly.
-        """
-        if self._verbose:
-            print(f'Starting memory-efficient range compression...')
-            print(f'Radar data shape: {self.radar_data.shape}')
-            print_memory()
-        
-        # Store original shape for verification
-        original_shape = self.radar_data.shape
-        expected_shape = (self.len_az_line, self.len_range_line)
-        
-        # Verify we still have expected dimensions
-        assert original_shape == expected_shape, \
-            f'Unexpected radar data shape: {original_shape}, expected: {expected_shape}'
-        
-        # Get range filter with matching dimensions
-        range_filter = self.get_range_filter()
-        
-        if self._verbose:
-            print(f'Range filter shape: {range_filter.shape}')
-            print(f'Applying range compression filter...')
-        
-        # Apply range compression filter - USE SAME METHOD AS ORIGINAL
-        self.radar_data = multiply(self.radar_data, range_filter)
-        
-        # Cleanup filter
-        cleanup_variables(range_filter)
-        
-        # Verify dimensions are preserved
-        assert self.radar_data.shape == original_shape, \
-            f'Range compression changed data shape from {original_shape} to {self.radar_data.shape}'
-        
-        if self._verbose:
-            print(f'Range compression completed. Data shape: {self.radar_data.shape}')
-            print_memory()
-
-    def _perform_rcmc_efficient(self) -> None:
-        """Perform memory-efficient Range Cell Migration Correction."""
-        if self._verbose:
-            print('Starting memory-efficient RCMC...')
-            print_memory()
-        
-        rcmc_filter = self.get_rcmc()
-        
-        # Use SAME METHOD AS ORIGINAL
-        self.radar_data = multiply(self.radar_data, rcmc_filter)
-        
-        # Cleanup filter
-        cleanup_variables(rcmc_filter)
-        
-        # Inverse FFT in range
-        self.ifft_range()
-        
-        if self._verbose:
-            print('RCMC completed.')
-            print_memory()
-
-    def _perform_azimuth_compression_efficient(self) -> None:
-        """Perform memory-efficient azimuth compression step.
-        
-        Raises:
-            ValueError: If array shapes are incompatible.
-        """
-        if self._verbose:
-            print('Starting memory-efficient azimuth compression...')
-            print(f'Radar data shape before azimuth filter: {self.radar_data.shape}')
-            print_memory()
-        
-        # Get azimuth filter
-        azimuth_filter = self.get_azimuth_filter()
-        
-        if self._verbose:
-            print(f'Azimuth filter shape: {azimuth_filter.shape}')
-        
-        # Apply azimuth compression - USE SAME METHOD AS ORIGINAL
-        self.radar_data = multiply(self.radar_data, azimuth_filter)
-        
-        # Cleanup filter
-        cleanup_variables(azimuth_filter)
-        
-        if self._verbose:
-            print(f'Radar data shape after azimuth compression: {self.radar_data.shape}')
-        
-        # Inverse FFT in azimuth
-        self.ifft_azimuth()
-        
-        if self._verbose:
-            print(f'Final radar data shape: {self.radar_data.shape}')
-            print_memory()
-
-    # ==================== UTILITY METHODS ====================
-    
-
-    @timing_decorator
-    def save_file(self, save_path: Union[str, Path]) -> None:
-        """Save processed radar data to file.
-        
-        Args:
-            save_path: Path where to save the data.
-        """
-        dump(self.radar_data, save_path)
-        if self._verbose:
-            print(f'Data saved to {save_path}')
-
-
-    # ==================== EEEZY ====================
-    # For backward compatibility - keep original method name as alias
-    _prompt_tx_replica = _generate_tx_replica
-
-    # ==================== GRAPHIC METHODS ====================
-    def _display_slice(self, slice=(0, 4000, 0, 4000), 
-                       vmin=0, vmax=1000, 
-                       figsize=(12,12), 
-                       title = None,
-                       step: str = 'az') -> None:
-        """Display a slice of the radar data for visualization."""
-        
-        assert step in ['raw', 'range_compressed', 'rcmc', 'az_compressed'], \
-            'Invalid step. Choose from "raw", "range_compressed", "rcmc", or "az_compressed".'
-        
-        if step == 'raw':
-            data = self.raw_data
-        elif step == 'range_compressed':
-            data = self.range_compressed_data
-        elif step == 'rcmc':
-            data = self.rcmc_data
-        elif step == 'az_compressed':
-            data = self.radar_data
-        else:
-            raise ValueError(f'Invalid step: {step}. Choose from "raw", "range", "rcmc", or "az".')
-        
-        plt.figure(figsize=figsize)
-        plt.imshow(np.abs(data[slice[0]:slice[1], slice[2]:slice[3]]), vmin=vmin, vmax=vmax, cmap='viridis')
-        plt.axis('off')
-        if title:
-            plt.title(title)
-        plt.show()