Enforcing numerical stability in numpy.log()

I have something like this:

import numpy as np

x = np.random.rand(10)  # I risk a zero entry
log_x = np.log(x)  # Risk of a divide by 0

What is the best way to enforce numerical stability here in the np.log call?

---

Current Solution

import numpy as np

def safe_log(x: np.ndarray) -> np.ndarray:
    while not x.all():
        x = np.nextafter(x, 1)
    return np.log(x)

source https://stackoverflow.com/questions/73101705/enforcing-numerical-stability-in-numpy-log