Design user-based recommendation system based on gived two APIs that can do the following things:

• getFriends (Person p)-> [Person] returns all friends of person
• getVideos (Person p)-> [Video] returns all the favorite videos of the person

Recommend a video to a person, the algorithm is to find all the videos his friends like and sort by the number of likes.

Here, before you do the OOD problem, 1) you need to clarify the question with the interviewer, 2) discuss use case 3) defined and core object, 4) draw the class and 5) check.
I will simplify this OOD programming question as a simple algorithm and data structure problem to demo the simple implementation. Later, you could convert it to the OOD problem by changing a Hash map to API call.

Thus, we can simplify the APIs as two Dict[String, List[String]], one is userVideos with user names as keys and a list of all the video that the user has likes.
Another is userFriends with user names as keys and a list of all this friend. The task is to return a Dict[String, List[String]],where the key is a user name and the value is a list Top K recommendation video based on his friend, if there is the same number of likes, then the video with the lower alphabetical order comes first.

Time complexity: O(n^3)
Space complexity: O(n)

Example 1:
Input:
userVideos = {
"David": ["Django Unchained", "Fight Club", "Inception"],
"Emma": ["Predators", "Forrest Gump", "Alien"],
"Jack": ["Fight Club", "The Shawshank Redemption", "Life Is Beautiful", "Back to the Future", "Forrest Gump", "Alien"]}
userFriends = {
"David": ["Emma", "Jack"],
"Emma": ["David"],
"Jack": ["David"]}
k = 3
Output:
{
'Emma': ['Django Unchained', 'Fight Club', 'Inception'],
'Jack': ['Django Unchained', 'Inception'],
'David': ['Alien', 'Forrest Gump', 'Back to the Future']
}

Example 2:
Input:
userVideos = {
"David": ["Django Unchained", "Fight Club", "Inception"],
"Emma": ["Predators", "Forrest Gump", "Alien"],
"Jack": ["Fight Club", "The Shawshank Redemption", "Life Is Beautiful", "Back to the Future", "Forrest Gump", "Alien"]}
userFriends = {
"David": [],
"Emma": [],
"Jack": []}
k = 3
Output:
{
'Emma': [],
'Jack': [],
'David': []
}

Example 3
Input:
userVideos = {
"David": ["Django Unchained", "Fight Club", "Inception"],
"Emma": ["Predators", "Forrest Gump", "Alien"],
"Jack": ["Fight Club", "The Shawshank Redemption", "Life Is Beautiful", "Back to the Future", "Forrest Gump", "Alien"]}
userFriends = {
"David": ["Emma", "Jack"],
"Emma": ["David"],
"Jack": ["David"]}
k = 1
Output:
{
'Emma': ['Django Unchained'],
'Jack': ['Django Unchained'],
'David': ['Alien']}
}

import heapq
import collections

class Solution(object):
    def recommendation_system(self, userVideos, userFriends, k):
        """
        :type userSongs: Dict[List[str]]
        :type songGenres: Dict[List[str]]
        :rtype: Dict[List[str]]
        """

        res = {}

        for user, friend_value_list in userFriends.items():
            exist_value_dict = set(userVideos[user])
            favorite_value_dict = collections.Counter()

            for friend in friend_value_list:
                for video in userVideos[friend]:
                    if not video in exist_value_dict:
                        favorite_value_dict[video] += 1

            res[user] = heapq.nsmallest(k, favorite_value_dict, key=lambda x:(-favorite_value_dict[x], x))

        return res

def main():
    userVideos = { 
        "David": ["Django Unchained", "Fight Club", "Inception"],
        "Emma": ["Predators", "Forrest Gump", "Alien"],
        "Jack": ["Fight Club", "The Shawshank Redemption", "Life Is Beautiful", 
                 "Back to the Future", "Forrest Gump", "Alien"]}
    userFriends = { 
        "David": ["Emma", "Jack"],
        "Emma": ["David"], 
        "Jack": ["David"]}
    k = 3
    solution = Solution()
    res = solution.recommendation_system(userVideos, userFriends, k)
    print(res)

if __name__ == "__main__": 
    main()