import pandas as pd
import util
from statsmodels import api as sm
from matplotlib import pyplot as plt
import datetime as dt
pid = 5
test_week_num = 6
alpha = 0.7
start_date = dt.datetime.strptime("2017-06-04", "%Y-%m-%d")
end_date = dt.datetime.strptime("2017-12-31", "%Y-%m-%d")
df_date = util.generate_time_index_df(start_date, end_date)
f = pd.read_csv("input/" + str(pid) + ".csv")
df_day = util.handle_original_df_day_sn(f, df_date)
df_week = util.handle_original_df_week_sn(df_day)
df = util.add_special_date(df_day, ahead_effect=7, behind_effect=5)
data = df[[
"log_sale_cnt",
"log_average_price",
# "sale_cnt",
# "average_price",
"special_date",
"ahead_special_date",
"behind_special_date"
]].values
sample_num = len(data)
str(i) for i in f[f.l3_gds_group_cd == category_id]["product_id"]
]
# for test
# category = [str(i) for i in [85, 83, 84, 285]]
# category = [str(i) for i in [8, 9, 10, 11, 252]]
# ============================== processing ==============================
# time index
start_date = dt.datetime.strptime("2017-06-04", "%Y-%m-%d")
end_date = dt.datetime.strptime("2017-12-31", "%Y-%m-%d")
df_date = util.generate_time_index_df(start_date, end_date)
df_all = [pd.read_csv("input/" + x + ".csv") for x in category]
df_day_all = [util.handle_original_df_day_sn(x, df_date) for x in df_all]
df_week_all = [util.handle_original_df_week_sn(x) for x in df_day_all]
# =============================== plot ===================================
num_products = len(category)
line_color = ["b", "g", "r", "y", "c", "k", "m"]
line_marker = [".", "*", "o", "x", "^", "v", "<", ">"]
line_kind = ["-", ":"]
line = [[
x[0] + x[1] + k for x in itertools.product(line_marker, line_color)
] for k in line_kind]
fig = plt.figure()
ax1 = fig.add_subplot(211)
line1 = []
for i in range(num_products):
line1.append(
