Python portfolio weight generation

I have a problem: I need to generate weights of a portfolio of n assets, and the weight of each asset must be a multiple of 1%, and at most 25%. I use the function np.random.randint. Then I divide the weights by the sum of the generated numbers, and encounter a problem during rounding some weights become less than 1%, and some greater than 25%, which contradicts the initial conditions

I used such function:

min_weight= 1 max_weight = 25 for asset in assets_for_portfolio: weight = np.random.randint(min_weight, max_weight) portfolio_weights = weight portfolio_weights /= portfolio_weights.sum()

But I encountered the problems I described