Write a Python function to efficiently compute the rolling average of a massive array of sensor data.
Implement def rolling_average(data, window_size):, returning a list of averages for every contiguous window of exactly window_size values. Use a sliding-window approach so each value is added and removed at most once. Assume data is non-empty and 1 <= window_size <= len(data); return floating-point averages.
Example: data = [2, 4, 6, 8], window_size = 2 returns [3.0, 5.0, 7.0]. Example: data = [5, 5, 5], window_size = 3 returns [5.0].
def rolling_average(data, window_size):