ranking

ranking.py

    def calculate_ranking(self, item_keys, values):
        '''
        Emit items with similarity in key for ranking:

        19,0.4    70,1
        19,0.6    21,2
        21,0.6    19,2
        21,0.9    70,1
        70,0.4    19,1
        70,0.9    21,1

        '''
        similarity, n = values
        item_x, item_y = item_keys
        if int(n) > 0:
            yield (item_x, similarity), (item_y, n)

    def top_similar_items(self, key_sim, similar_ns):
        '''
        For each item emit K closest items in comma separated file:

        De La Soul;A Tribe Called Quest;0.6;1
        De La Soul;2Pac;0.4;2

        '''
        item_x, similarity = key_sim
        for item_y, n in similar_ns:
            print '%s;%s;%f;%d' % (item_x, item_y, similarity, n)