Question about the data cleaning function

[scottshufe]

Hi, thanks for your excellent work!

I find that the data cleaning process does not seem to successfully filter out users with less than 10 interactions.

casper/src/util.py

Lines 20 to 28 in 23012d8

	users = users[counts>=10]
	sequence_dic = {int(user):[] for user in set(raw_data[:,0])}
	new_data = []
	for i in range(raw_data.shape[0]):
	if int(raw_data[i,0]) in user_dic:
	new_data.append([int(raw_data[i,0]),int(raw_data[i,1]),raw_data[i,2],0])
	new_data = np.array(new_data)

Although the variable "users" in Line 20 refers to those users having more than 10 interactions, it has not been used in the following codes. Could you please check this issue?

I suggest modifying this part of codes as follows:

users = users[counts >= 10]
# sequence_dic = {int(user): [] for user in set(raw_data[:, 0])}
sequence_dic = {int(user): [] for user in users}
new_data = []
for i in range(raw_data.shape[0]):
    # if int(raw_data[i, 0]) in user_dic:
    if int(raw_data[i, 0]) in users:
        new_data.append([int(raw_data[i, 0]), int(raw_data[i, 1]), raw_data[i, 2], 0])

new_data = np.array(new_data)

Thank you very much.