Implement equivalent to numpy.nan_to_num

[ricardoV94]

Description

A couple of switches should do the job, no need to implement a new Op

https://numpy.org/doc/stable/reference/generated/numpy.nan_to_num.html

[aspeers]

I'm interested in working on this.

[priyankavgalagali]

import numpy as np
example_array = np.array([np.nan, 1, np.inf, -np.inf, 3], dtype=float)

nan_replace = -1
posinf_replace = 1000
neginf_replace = -1000
import numpy as np

def custom_nan_to_num(arr, nan_val=0.0, posinf_val=None, neginf_val=None):
if posinf_val is None:
posinf_val = np.finfo(arr.dtype).max
if neginf_val is None:
neginf_val = np.finfo(arr.dtype).min

# Using np.where to replace values based on condition
arr = np.where(np.isnan(arr), nan_val, arr)
arr = np.where(np.isposinf(arr), posinf_val, arr)
arr = np.where(np.isneginf(arr), neginf_val, arr)

return arr

Apply the custom function

result = custom_nan_to_num(example_array, nan_val=nan_replace,
posinf_val=posinf_replace,
neginf_val=neginf_replace)

print(result)

Im new to this.....Can you let me know if its something like this or the function of numpy.nan_to_num is to be replaced?