S_Прибыль

#В файле visits.csv зафиксированы даты посещения клиента.
#Поля:
#id - уникальный идентификатор клиента
#start_dt - дата визита (первый визит совпадает с датой регистрации клиента)

#В файле purchases.csv находится информация о совершенных покупках клиентом.
#Поля:
#buy_ts — дата и время покупки
#id - уникальный идентификатор клиента
#revenue — доход в y.e.

#Посчитайте LTV на 30 день с момента прихода клиента. Постройте график динамики LTV по дням.
#Для начала возьмём небольшую когорту -- 50 пользователей -- затем увеличим количество

#(1)Берем когорту пользователей и для каждого пользователя считаем прибыль в динамике по дням с момента регистрации.

import csv
calendar = {
    2016: {1: 31, 2: 29, 3: 31, 4: 30, 5: 31, 6: 30, 7: 31, 8: 31, 9: 30, 10: 31, 11: 30, 12: 31},
    2017: {1: 31, 2: 28, 3: 31, 4: 30, 5: 31, 6: 30, 7: 31, 8: 31, 9: 30, 10: 31, 11: 30, 12: 31}
}

def count_end_date(start_date, calendar_):
    year = start_date[0]
    month = start_date[1]
    day = start_date[2]
    end_year = year
    end_month = month
    end_day = day
    for i in range(30):
        end_day += 1
        if end_year not in calendar_.keys():
            continue
        if end_day > calendar_[end_year][end_month]:
            end_day = 1
            end_month +=  1
            if end_month == 13:
                end_month = 1
            if end_month == 1:
                end_year += 1
    end_date = [end_year, end_month, end_day]
    return end_date

def check_period( date, start_date, end_date, calendar_):
    start_year = start_date[0]
    start_month = start_date[1]
    start_day = start_date[2]
    end_year = end_date[0]
    end_month = end_date[1]
    end_day = end_date[2]
    this_year = date[0]
    this_month = date[1]
    this_day = date[2]
    ans = False
    if this_month == start_month:
        if this_day <= calendar_[start_year][start_month]:
            ans = True
    elif this_month == end_month:
        if this_day <= end_day:
            ans = True
    return ans

from_date = {}
to_date = {}


def recycle(line): #get [id, date, income] from a row of csv
    if len(line) == 1 and type(line[0]) == str and '"' in line[0]:
        line = line[0]
        sz = len(line)
        num_quotes = 0
        income = ""
        for i in reversed(range(sz)):
            if line[i] == '"':
                num_quotes += 1
            elif num_quotes == 1:
                income += line[i]
            elif num_quotes == 2:
                break
        if income == '':
            return
        income = income[::-1]
        income = income.replace(',', '.')
        income = float(income)

        date = line.split()[0]
        date = date.split('.')
        day = int(date[0])
        month = int(date[1])
        year = int(date[2])
        date = [year, month, day]
        id = line.split(',')
        id = id[1]
    elif len(line) == 1 and type(line[0]) == str and '"' not in line[0]:
        #['11.01.2017 10:52:11,2636761148,11']
        date, id, income = line[0].split(',')
        income = float(income)
        date = list(int(i) for i in date.split()[0].split('.')[::-1])
    else:
        income = float(line[-1])
        id = line[-2]
        date = line[0].split()[0].split('.')
        day = int(date[0])
        month = int(date[1])
        year = int(date[2])
        date = [year, month, day]
    return [id, date, income]

dynamic_income_per_user = {}
with open("visits.csv", 'r', encoding = 'utf-8') as file:
    reader = csv.reader(file, delimiter = ',', lineterminator = '\r')
    u = 0
    for line in reader:
        u+=1
        id = line[0]
        date = line[1]
        date = list(int(i) for i in date.split('-'))  #[year, month, day]
        if id not in from_date.keys():
            from_date[id] = date
            to_date[id] = count_end_date(date, calendar)
            dynamic_income_per_user[id] = [0] * 30


def num_of_day_from_start(start_date, date, calendar_):
    start_year = start_date[0]
    start_month = start_date[1]
    start_day = start_date[2]
    this_year = date[0]
    this_month = date[1]
    this_day = date[2]
    days_till_start = 0
    days_till_this = 0
    #how many days till start
    for year in calendar_:
        for month in calendar_[year]:
            if month == start_month and year == start_year:
                days_till_start += start_day
                break
            else:
                days_till_start += calendar_[year][month]
    #how many days till this date
    for year in calendar_:
        for month in calendar_[year]:
            if month == this_month and year == this_year:
                days_till_this += this_day
                break
            else:
                days_till_this += calendar_[year][month]
    days_from_start_till_this_date = days_till_this - days_till_start
    return days_from_start_till_this_date


with open("purchases.csv", 'r', encoding = 'utf-8') as file:
    reader = csv.reader(file, delimiter = ';', lineterminator = '\r')
    for line in reader:
        purchase = recycle(line) #[id, date, income]
        id = purchase[0]
        date = purchase[1]
        income = purchase[2]
        start_date = from_date[id]
        end_date = to_date[id]
        in_30_days_from_start = check_period(date, start_date, end_date, calendar)
        if in_30_days_from_start == True:
            day_from_start = num_of_day_from_start(start_date, date, calendar)
            if day_from_start not in range(0,30):
               continue
            print(day_from_start)
            if id in dynamic_income_per_user:
                dynamic_income_per_user[id][day_from_start] += income

#теперь у нас есть данные по "динамической" прибыли о каждого покупателя
print(dynamic_income_per_user)
#Идея на завтра: записать DictWriter --> дни от 1 до 30 с регистрации -->DictReader -->считать сумму и т.д .
#ИЛИ можно обойтись без этого???
#Кстати, почему код don't run but debug???