import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

# Path settings
BIT_FILE_PATH = r"G:\Users\Chipistil\Desktop\High-throughput-T2NP-Memristor-based-Dual-clock-edge-sampling-TRNG\1Mb.txt"
MAX_LAG = 10000
READ_COUNT = 1000000
SAVE_PATH = r"G:\Users\Chipistil\Desktop\High-throughput-T2NP-Memristor-based-Dual-clock-edge-sampling-TRNG\True Random Numbers Verification\03_autocorrelation_function_analysis_ACF_analysis\AHU Black Swan_ACF_plot.png"
EXCEL_PATH = r"G:\Users\Chipistil\Desktop\High-throughput-T2NP-Memristor-based-Dual-clock-edge-sampling-TRNG\True Random Numbers Verification\03_autocorrelation_function_analysis_ACF_analysis\AHU Black Swan_ACF_data.xlsx"

def calculate_acf(bit_data, max_lag):
    N = len(bit_data)
    mu = np.mean(bit_data)
    sigma = np.std(bit_data, ddof=1)
    
    lags = np.arange(0, max_lag + 1)
    acf_values = []
    
    for k in lags:
        sum_corr = np.sum((bit_data[:N - k] - mu) * (bit_data[k:] - mu))
        acf = sum_corr / ((N - k) * sigma ** 2)
        acf_values.append(acf)
    
    return lags, np.array(acf_values)

if __name__ == "__main__":
    with open(BIT_FILE_PATH, 'r') as f:
        content = f.read()
    bit_stream_full = np.array([int(c) for c in content if c in '01'], dtype=np.int64)
    if READ_COUNT is not None:
        bit_stream = bit_stream_full[:READ_COUNT]
    else:
        bit_stream = bit_stream_full
    print(f"✅ Data reading completed, bitstream length: {len(bit_stream):,}")
    
    lags, acf_vals = calculate_acf(bit_stream, MAX_LAG)
    print(f"✅ Autocorrelation calculation completed, lag range: 0 ~ {MAX_LAG}")
    
    N = len(bit_stream)
    conf_95_theoretical = 1.96 / np.sqrt(N)
    conf_95_actual = 1.96 * np.std(acf_vals[1:])
    
    USE_ACTUAL_CI = True
    
    if USE_ACTUAL_CI:
        conf_95 = conf_95_actual
        print(f"✅ 95% confidence interval (based on actual data): ±{conf_95:.6f}")
    else:
        conf_95 = conf_95_theoretical
        print(f"✅ 95% confidence interval (theoretical): ±{conf_95:.6f}")
    
    plt.rcParams['font.sans-serif'] = ['DejaVu Sans']
    plt.rcParams['axes.unicode_minus'] = False
    plt.figure(figsize=(12, 6), dpi=300)
    
    plt.scatter(lags[1:], acf_vals[1:], s=1, color='black', alpha=0.8)
    
    plt.axhline(y=conf_95, color='gray', linestyle='--', linewidth=1, label=f'95% CI: ±{conf_95:.6f}')
    plt.axhline(y=-conf_95, color='gray', linestyle='--', linewidth=1)
    
    plt.axhline(y=0, color='red', linestyle='-', linewidth=0.8)
    
    plt.xlabel('Lag', fontsize=12, fontweight='bold')
    plt.ylabel('Normalized Autocorrelation Coefficient (ACF)', fontsize=12, fontweight='bold')
    plt.title('1Mb Random Bitstream Autocorrelation Function', fontsize=14, fontweight='bold', pad=20)
    plt.legend(loc='upper right', fontsize=10)
    plt.grid(alpha=0.2, axis='y')
    plt.xlim(0, MAX_LAG)
    
    y_min = min(np.min(acf_vals[1:]), -conf_95) * 1.2
    y_max = max(np.max(acf_vals[1:]), conf_95) * 1.2
    plt.ylim(y_min, y_max)
    
    plt.tight_layout()
    plt.savefig(SAVE_PATH, dpi=300, bbox_inches='tight')
    
    df = pd.DataFrame({'Lag': lags[1:], 'ACF': acf_vals[1:]})
    df.to_excel(EXCEL_PATH, index=False, engine='openpyxl')
    print(f"✅ Data saved to: {EXCEL_PATH}")
    print(f"Lag data points: {len(lags[1:])}")
    
    print(f"✅ Autocorrelation plot saved to: {SAVE_PATH}")