The code in craft_signal.py applies the inverse Fourier Transform on the bytes it reads from the flag.txt. So we only need to apply the Fourier Transform on the fftea data. This code thus recovers the flag :
import numpy as np
# Load the data from the "fftea" file
data = np.fromfile("challenge", dtype = np.complex64)
data = np.fromfile("fftea", dtype = np.complex64)
# Perform Fourier transform
original_data = np.fft.fft(data, n=64)
# Convert the complex64 array to bytes
flag = original_data.real.astype(np.complex64)
int_array = np.array([int(c.real) for c in flag])
aa = [chr(i) for i in int_array]
print(''.join(aa))