Generating a random number based on a distribution function

We can generate random numbers in an interval [a,b] easily if we want to make it uniformely:

A=rand()*(b-a) + a

where rand() is a function which can generate a uniform random number between 0 and 1. so A is a random number in [a,b].

For generating a random number based on a distribution function like y=x-x^2, i faced a problem.

I would like to use a method mentioned here. But i am not interested to use the python function inverse_cdf(np.random.uniform()).

I can compute the CDF of function "y" by an integration over 0 and X and i call it "f". But when i put the rand() function(a number between 0 and 1) into the inverse function of f, i get a complex number! It means: A=f^(-1) (rand()) returns a complex number.

Is it a correct way for generating a random number based on a distribution function?