data.py

import numpy as np

def load_data(file_dir):
    # output: 
    # N: the number of user;
    # M: the number of item
    # data: the list of rating information
    user_ids_dict, rated_item_ids_dict = {},{}
    N, M, u_idx, i_idx = 0,0,0,0 
    data = []
    f = open(file_dir)
    for line in f.readlines():
        if '::' in line:
            u, i, r, _ = line.split('::')
        else:
            u, i, r, _ = line.split()
    
        if int(u) not in user_ids_dict:
            user_ids_dict[int(u)]=u_idx
            u_idx+=1
        if int(i) not in rated_item_ids_dict:
            rated_item_ids_dict[int(i)]=i_idx
            i_idx+=1
        data.append([user_ids_dict[int(u)],rated_item_ids_dict[int(i)],float(r)])
    
    f.close()
    N = u_idx
    M = i_idx

    return N, M, data, rated_item_ids_dict
	

def sequence2mat(sequence, N, M):
    # input:
    # sequence: the list of rating information
    # N: row number, i.e. the number of users
    # M: column number, i.e. the number of items
    # output:
    # mat: user-item rating matrix
    records_array = np.array(sequence)
    mat = np.zeros([N,M])
    row = records_array[:,0].astype(int)
    col = records_array[:,1].astype(int)
    values = records_array[:,2].astype(np.float32)
    mat[row,col]=values
    
    return mat