calculating similarity weights between two users for a user-based collaborative filtering method

Question

I want to build a movie recommendation system. I am not able to find answers on how to calculate the similarity weights. Here in the article, there is a rough estimation of