tranlyvu/recommender

from recommender_engine import make_recommendation
from xlrd import open_workbook

"""
    This sample used jester dataset @ http://goldberg.berkeley.edu/jester-data/
"""

def make_preference_space(raw_data, start):
    
    """ create preference_space from raw data

        Args:
            raw_data
            start

        Returns:
            dict: keys are users and values are dictionaries of items and ratings
    """
    preference_space = {}
    for row in range(0,raw_data.nrows):    
        preference_space[str('user ' + str(start + row))] = {}    
        for col in range(1,raw_data.ncols):
            if raw_data.cell_value(row,col) == 99: continue
            preference_space[str('user ' + str(start + row))]['joke ' + str(col)] = float(raw_data.cell_value(row, col))        
    
    return preference_space


def make_preference_space_Jester(file_path):
    """
        The function takes in 3 jester dataset files and then merge the file
    """
    
    raw_data_1 = open_workbook(file_path + '/jester-data-1.xls').sheets()[0]
    prefs_space_1 = make_preference_space(raw_data_1, 0)
    raw_data_2 = open_workbook(file_path + '/jester-data-2.xls').sheets()[0]
    prefs_space_2 = make_preference_space(raw_data_2, raw_data_1.nrows)
    raw_data_3 = open_workbook(file_path + '/jester-data-3.xls').sheets()[0]
    prefs_space_3 = make_preference_space(raw_data_3,raw_data_1.nrows + raw_data_2.nrows)

    result = {}
    for space in [prefs_space_1, prefs_space_2, prefs_space_3]:
        result.update(space)
    
    return result
    

def jester():
    """
        Running the application using Jester dataset:
    """

    #Creating preference_space from Jester dataset
    preference_space = make_preference_space_Jester('../Jester')
    result = make_recommendation('user 36681', preference_space, 
                                recommender_approach = 'user_based', 
                                number_of_items_to_recommend = 5, 
                                similarity_measure = 'euclidean_distance')
    print(result)


if __name__=="__main__":
    jester()