mapreduce_job2

mapreduce_job2.py

    def count_max_of_mutual_friends(self, key, values):

        '''
        Prepare the dataset to yield the source and
        get the top suggestions.

        Input ({[friend1, friend2], numberOfMutualFriends}):

            ["fabio", "marcel"] 1
            ["fabiola", "marcel"]   1
            ["marcel", "patricia"]  1
            ["marcel", "paula"] 1
            ["carol", "marcel"] 2

        Output ({friend1,  [numberOfMutualFriends, friend2]},
                {friend2,  [numberOfMutualFriends, friend1]}):

            "fabio", [1,"marcel"]
            "marcel", [1,"fabio"]
            "fabiola", [1,"marcel"]
            "marcel", [1,"fabiola"]
            "marcel", [1,"patricia"]
            "patricia", [1,"marcel"]
            "marcel", [1,"paula"]
            "paula", [1,"marcel"]
            "marcel", [2,"carol"]
            "carol", [2,"marcel"]

        '''
        f1, f2 = key
        # for score in values:
        yield f1, (f2, int(values))
        yield f2, (f1, int(values))

    def top_recommendations(self, key, values):
        '''
        Get the TOP N recommendations for user.

        Input ({friend1,  [(numberOfMutualFriends, friend),
                           (numberOfMutualFriends2, friend)]}):

            "fabio", [[1,"marcel"]]
            "marcel", [[2,"carol"], [1,"fabio"], [1,"fabiola"], [1,"patricia"],
                        [1,"paula"]]
            "fabiola", [[1,"marcel"]]
            "patricia", [[1,"marcel"]]
            "paula", [[1,"marcel"]]
            "carol", [[2,"marcel"]]

        Output ({friend1,  [(numberOfMutualFriends, friend),
                           (numberOfMutualFriends2, friend)]}):

            Ex: Get the top 3 suggestions.

            "fabio", [[1,"marcel"]]
            "marcel", [[2,"carol"], [1,"fabio"], [1,"fabiola"]]
            "fabiola", [[1,"marcel"]]
            "patricia", [[1,"marcel"]]
            "paula", [[1,"marcel"]]
            "carol", [[2,"marcel"]]
        '''
        recommendations = []
        for idx, (item, score) in enumerate(values):
            recommendations.append((item, score))

        yield key, sorted(recommendations, key=lambda k: -k[1])[:TOP_N]