services/anomaly_service.py

import numpy as np  # Import NumPy for statistical calculations

# Function to track fault counts and calculate statistics
def track_faults(fault_counts):
    if not fault_counts:
        return {"mean": 0, "std": 0}  # Return default values if no data
    mean = np.mean(fault_counts)  # Calculate mean of fault counts
    std = np.std(fault_counts)  # Calculate standard deviation
    return {"mean": mean, "std": std}

# Function to predict potential fault escalation
def predict_fault(fault_counts, threshold=2.5):
    if len(fault_counts) < 5:  # Require at least 5 frames for prediction
        return False
    stats = track_faults(fault_counts)  # Get statistics
    recent_counts = fault_counts[-5:]  # Look at last 5 frames
    for count in recent_counts:
        if stats["std"] > 0 and (count - stats["mean"]) / stats["std"] > threshold:  # Check for spike
            return True
    return False  # No escalation detected