Question about survival time data simulation methodology in examples

[shuhxi]

Hi, thanks for making this repo available with all the detailed examples! I noticed that in the simulation files in examples/data (e.g.Simulation_Linear/utils.py), the survival times $T_0$ and $T_1$ are generated as:

T0 = np.exp((np.dot(x2, gamma1.T) + np.dot(x1, gamma1.T)))
T1 = np.exp((np.dot(x2, gamma2.T) + np.dot(x1, gamma1.T)))

This approach seems to result in log-normally distributed $T_0$ and $T_1$. I'm curious about this since, in works like Alaa et al., 2017 (which is cited in the BITES paper), survival times are assumed to be exponentially distributed and sampled e.g. using np.random.exponential in pycox. Could you please clarify the reasoning behind this approach for generating survival times?

Thanks again for your great work!