def delete_empty_dicts():
db = CDB()
db.load(file_name)
swipes = db.get_data()
keys = list(swipes.keys())
for key in keys:
if len(swipes[key].keys()) == 0:
swipes.pop(key)
d = Data()
d.set_data(swipes)
db.set_data(d)
db.save(file_name)
def get_days_over_100(file_name):
#Function to Remove Data from days before day 100.
#This function is required because the Message data
#supplied to us starts 100 days after the swipe and question data.
db = CDB()
db.load(file_name)
swipes = db.get_data()
keys = swipes.keys()
for key in keys:
days = sorted(swipes[key].keys())
for day in days:
if int(day) < 100:
swipes[key].pop(day)
d = Data()
d.set_data(swipes)
db.set_data(d)
db.save(file_name)
def regularize_combine():
f = open("hmm_data.txt", "w")
swipes_db = CDB()
swipes_db.load("swipes.db") #Load swipe data
questions_db = CDB()
questions_db.load("questions.db") #load question data
messages_db = CDB()
messages_db.load("messages.db") #load message data
swipes = swipes_db.get_data(
) #Data on the number of other users each user has swiped on each day
questions = questions_db.get_data(
) #Data on the number of questions each user has answered on each day
messages = messages_db.get_data(
) #Data on the number of messages each user has sent on each day
keys = swipes.keys()
for key in keys: #Where key is the userid of a user
#Begin finding the earliest and latest days in the sequence for a specific user
earliest_days = [
] #List of earliest days of activity. Will contain days of earliest swipe, earliest question answered, and earliest message sent
latest_days = [
] #List of latest days of activity. Will contain days of latest swipe, latest question answered, and latest message sent
key_swipes = swipes[key].keys(
) #A list of the days that the user swiped on at least 1 other user
swipe_days = sorted(swipes[key].keys(
)) #Sorted list of days that the user swiped on at least 1 other user
earliest_days += [
int(swipe_days[0])
] #Extract earliest day that the user swiped on at least 1 other user and add it to list of earliest days
latest_days += [
int(swipe_days[-1])
] #Extract latest day that the user swiped on at least 1 other user and add it to list of latest days
key_questions = {
} #Will be added to later if the user has answered any questions
key_messages = {
} #Will be added to later if the user has sent any messages
if key in questions: #Check if the user has answered any questions
key_questions = questions[key].keys(
) #List of days that the user has answered at least one question
question_days = sorted(questions[key].keys())
earliest_days += [
int(question_days[0])
] #Extract earliest day that the user answered at least 1 question and add it to list of earliest days
latest_days += [
int(question_days[-1])
] #Extract latest day that the user answered at least 1 question and add it to list of latest days
if key in messages: #Check if the user has sent any messages
key_messages = messages[key].keys(
) #List of days that the user has sent at least one message
messages_days = sorted(messages[key].keys())
earliest_days += [int(messages_days[0])]
latest_days += [int(messages_days[-1])]
earliest = min(
earliest_days
) #Find the earliest day the user has done any of the above activities
latest = max(
latest_days
) #Find the latest day the user has done any of the above activities
mess = [
] #Sequence of number of messages sent from earliest day to latest day
swip = [
] #Sequence of number of users swiped from earliest day to latest day
quest = [
] #Sequence of number of questions answered from earliest day to latest day
for i in range(earliest, latest +
1): #Iterate through every day from earliest to latest
i = str(i)
if i in key_swipes:
#Check if the user has swiped any other users on this day
swip += [swipes[key][i]
] #if they have, add the number to the sequence
else:
#If they haven't, add 0 to the sequence
swip += [0]
if i in key_questions:
quest += [questions[key][i]]
else:
quest += [0]
if i in key_messages:
mess += [messages[key][i]]
else:
mess += [0]
swip = np.array(swip)
mess = np.array(mess)
quest = np.array(quest)
participation = swip + quest #Participation is defined as the number of other users swiped plus the number of questions answered
#normalize participation
part_mean = np.mean(participation)
part_stdv = np.std(participation)
participation = participation - part_mean
if part_stdv != 0:
participation = participation / part_stdv
#Normalize messages
mess_mean = np.mean(mess)
mess_stdv = np.std(mess)
mess = mess - mess_mean
if mess_stdv != 0:
mess = mess / mess_stdv
#Activity is the mean of participation and messages sent
activity = mess + participation
activity = activity / 2
activity = list(activity)
#Save the sequence to hmm_data.txt
print(json.dumps(activity), file=f)
f.close()