def get_regression_features(zone_choice,
arrival_choice=-1,
cut_choice=-1,
cap_mode_choice=2):
zone_path, zone_file_prefix = get_zone_output_path(zone_choice, root)
df = pd.read_csv(os.path.join(zone_path, zone_file_prefix + 'Summary.csv'))
df = unstack_summary_df(df, zone=zone_choice).dropna(subset=['capacity'])
df = df.fillna(0)
df['cut'] = to_categories(df['cut'])
if cut_choice in [0, 1]:
df = df[df['cut'] == cut_choice]
elif cut_choice == 2:
df = df[df['cut'] != cut_choice]
if arrival_choice != -1:
df = df[df['arrival'] == arrival_choice]
y = df["discount"]
if cap_mode_choice == 0:
x = df[['eco', 'capacity']]
elif cap_mode_choice == 1:
x = df[['eco', 'capacity', 'capacity_avg_day']]
else:
x = df[['eco', 'capacity', 'capacity_avg_day', 'capacity_avg_global']]
return x, y, df
def create_summary_by_zone(zones):
"""
Creates summary file in zone folder
Args:
zones (iterable): Zone number
"""
for zone in zones:
zone_path, zone_file_prefix = get_zone_output_path(zone, data_path)
df_avail, df_order, df_steer, df_cap = load_data_file_for_zone(
zone, data_path)
slots_observed, cslots_observed = get_slots_observed(
zone_path,
zone_file_prefix,
df_avail,
df_order.columns,
check_existing=False)
slots_active, cslots_active, slots_offered, cslots_offered = get_slots_active(
zone_path, zone_file_prefix, df_order, slots_observed)
summary = summarize(zone_path, zone_file_prefix, df_avail, df_order,
df_steer, df_cap, slots_offered, cslots_offered)
summary = summary.reset_index()
summary['EVENT_DTM'] = pd.to_datetime(summary['EVENT_DTM'])
plot_arrivals(zone_path, zone_file_prefix, summary, '60min', 'mean',
'ALL')
def create_eco_and_discount_from_steering(zone):
"""
Reads steering file and creates eco and discount files from it
Args:
zone (str): Zone number in float
Returns:
(pd.DataFrame): eco_df, discount_df
"""
zone_path, zone_file_prefix = get_zone_output_path(zone, data_path)
df_avail, df_order, df_steer, df_cap = load_data_file_for_zone(
zone, data_path)
def split_summary_by_day_and_cut():
"""
Returns:
"""
zlist = ['700.0', '500.0']
for zone in zlist:
zone_path, zone_file_prefix = get_zone_output_path(zone, data_path)
summary_zone = pd.read_csv(
os.path.join(zone_path, 'Zone_' + zone[:-2] + '_Summary.csv'))
summary_zone['EVENT_DTM'] = pd.to_datetime(summary_zone['EVENT_DTM'])
slots_zone_offered, cslots_zone_offered = read_offered_slots(zone_path)
for day in ['0', '1', '2', '3', '4', '5', '6']:
plot_arrivals(
os.path.join(zone_path, day),
os.path.join(zone_file_prefix, '_Arrival_Day_' + day),
summary_zone, '60min', 'mean', day)
# save offered slots for each day
# summary df by day
for cutcat in ['BEFORE_CUT1', 'BEFORE_CUT2', 'MISSED_BOTH_CUTS']:
print(zone, day, cutcat)
cut_path, cut_prefix = get_cut_path(day, cutcat, zone_path,
zone_file_prefix)
summary = summary_zone.loc[
(summary_zone['ARRIVAL_CAT'] == cutcat)
& (summary_zone['ARRIVAL_DAY'] == int(day))]
slots_active, cslots_active, slots_offered, cslots_offered = get_slots_active(
cut_path, cut_prefix,
summary[slots_zone_offered + ['SLOT_CHOICE']],
slots_zone_offered)
cols = summary.columns[0:16].tolist() + ['SLOTS_AVAILABLE', 'NO_PURCHASE'] + slots_offered + \
['CNO_PURCHASE'] + cslots_offered + [col + '_Eco' for col in slots_offered] + \
[col + '_Discount' for col in slots_offered] + [col + '_Capacity' for col in slots_offered]
summary[cols].to_csv(os.path.join(cut_path,
cut_prefix + 'Summary.csv'),
index=False)
del slots_offered, cslots_offered, slots_active, cslots_active,
if pd.size(summary):
plot_arrivals(cut_path, cut_prefix, summary[cols], '60min',
'mean', day)
plt.close()
del summary, cut_path, cut_prefix
def get_regression_features_ranked(zone_choice,
arrival_choice=-1,
cut_choice=-1,
cap_mode_choice=2):
zone_path, zone_file_prefix = get_zone_output_path(zone_choice, root)
df = pd.DataFrame()
path = os.path.join(zone_path, 'RankData',
zone_file_prefix + 'Arrival_Day_{}' + '_Summary.csv')
for days in range(7):
df = df.append(pd.read_csv(path.format(days)))
df = unstack_summary_df_ranked(
df, zone=zone_choice).dropna(subset=['capacity'])
df = df.fillna(0)
df['cut'] = to_categories(df['cut'])
if cut_choice in [0, 1]:
df = df[df['cut'] == cut_choice]
elif cut_choice == 2:
df = df[df['cut'] != cut_choice]
if arrival_choice != -1:
df = df[df['arrival'] == arrival_choice]
features = ['eco', 'arrival', 'cut', 'slot']
y = df["discount"]
if cap_mode_choice == 0:
x = df[features + ['capacity']]
elif cap_mode_choice == 1:
x = df[features + ['capacity', 'capacity_avg_day']]
else:
x = df[features +
['capacity', 'capacity_avg_day', 'capacity_avg_global']]
slot = x['slot'].str.split("_", n=1, expand=True)
x['day'] = slot[0]
x['slot'] = slot[1]
x = col_to_one_hot(x, 'slot', prefix='slot', drop_first=True)
x = col_to_one_hot(x, 'day', prefix='day', drop_first=True)
return x, y, df
len(gr_cols))
print('Iteration=', i, 'loglikelihood =', log_likeli, 'beta_disc',
beta_coef[-3], 'beta_eco', beta_coef[-2], 'beta_gr',
beta_coef[-1])
if np.linalg.norm(beta_coef[:-1] - beta[:-1]) < 10**-6 or i > 500:
predict_prob_df = pd.DataFrame(Q,
columns=['NO_PURCHASE'] +
slots_offered)
beta_df = pd.DataFrame([np.array(beta_coef)],
columns=['NO_PURCHASE'] + slots_offered +
['Discount', 'Eco', 'Gr'])
predict_prob_df.to_csv(Location + filename +
'predprobfeatures.csv')
beta_df.to_csv(Location + filename + 'betafeatures.csv')
del summary, predict_prob_df, design_df, features_df, assortment_df, choice_df, design, features, assortment, choice
break
return beta_df.iloc[0]
if __name__ == "__main__":
zone_path, zone_file_prefix = get_zone_output_path(zone, root)
summary = pd.read_csv(
'/Users/anupamtripathi/PycharmProjects/RA_/results/gr/500.0/gr_unranked_stacked_arrival_0_cut_-1_cap_2.csv'
)
summary = summary.drop(['NO_PURCHASE_Eco', 'NO_PURCHASE_Discount'], axis=1)
summary = summary.dropna()
slots_offered = pd.read_csv(
os.path.join(zone_path, zone_file_prefix + 'SlotsOfferedTitle.csv'))
MMfeaturesBoot(zone_path, zone_file_prefix, summary,
slots_offered['slotsOffered'].tolist())
def unstack_summary_df_ranked(summary_df,
zone,
root='data',
check_saved=False,
save=False): # do for summary
"""
Unstacks the summary dataframe
Args:
summary_df (dataframe): summary dataframe
Returns: unstacked dataframe
"""
zone_path, zone_file_prefix = get_zone_output_path(zone, root)
if os.path.exists(
os.path.join(zone_path, 'RankData',
zone_file_prefix + 'unstacked.csv')) and check_saved:
return pd.read_csv(os.path.join(zone_path, 'RankData',
zone_file_prefix + 'unstacked.csv'),
index_col=0)
summary_df['primary_key'] = summary_df['EVENT_DTM'].astype(
str) + '-' + summary_df['CUSTOMER_ID'].astype(str)
slots_per_day = u.get_slots_per_day_for_zone_ranked(zone, root)
slots, cslots = u.get_slots_observed_ranked(zone_path, zone_file_prefix,
summary_df)
# get capacities sums
summary_df['capacity_sum_global'] = summary_df.loc[:, slots[0] +
'_Capacity':slots[-1] +
'_Capacity'].fillna(
0).sum(axis=1)
for day in range(7):
slot_start = slot_end if day != 0 else list(
summary_df.columns).index(slots[0] + '_Capacity')
slot_end = slot_start + slots_per_day[day + 1]
summary_df['capacity_sum_day_' +
str(day + 1)] = summary_df.iloc[:,
slot_start:slot_end].fillna(
0).sum(axis=1)
summary_df = summary_df.fillna(0)
df_slots = pd.DataFrame()
for n in tqdm(slots + ['NO_PURCHASE']):
features = pd.DataFrame()
features['primary_key'] = summary_df['primary_key']
features['arrival'] = summary_df['ARRIVAL_DAY']
features['cut'] = summary_df['ARRIVAL_CAT']
features['slot'] = n
features['day'] = n[0]
if n != 'NO_PURCHASE':
features['capacity'] = summary_df[n + '_Capacity']
features['discount'] = summary_df[n + '_Discount']
features['eco'] = summary_df[n + '_Eco']
features['capacity_avg_global'] = (
summary_df['capacity_sum_global'] -
summary_df[n + '_Capacity']) / (len(slots) - 1)
features['capacity_avg_day'] = (
summary_df['capacity_sum_day_' + str(n[0])] -
summary_df[n + '_Capacity']) / (slots_per_day[int(n[0])] - 1)
else:
features['day'] = n
features['capacity'] = 1
features['discount'] = 0
features['eco'] = 0
features['capacity_avg_global'] = 1
features['capacity_avg_day'] = 1
features['order'] = summary_df[n]
features['avail'] = summary_df['C' + n]
df_slots = df_slots.append(features)
df_slots = u.col_to_one_hot(df_slots, 'slot', prefix='slot', delete=False)
if save:
df_slots.to_csv(
os.path.join(zone_path, 'RankData',
zone_file_prefix + 'unstacked.csv'))
return df_slots
def create_rank_data_by_zone(zone, computername):
slots_offered_rank = []
location, filename = get_zone_output_path(zone, computername)
hourlist = get_hour_list(zone)
summary = pd.read_csv(os.path.join(location, filename + 'Summary.csv'))
slots_offered, _ = u.get_slots_active_ranked(location, filename, summary)
for i in range(1, 8):
for j in list(hourlist.values()):
slots_offered_rank = np.append(slots_offered_rank,
str(i) + '_' + j)
summaries = []
for day in ['0', '1', '2', '3', '4', '5', '6']:
rank_folder = os.path.join(location, 'RankData')
rank_file_prefix = filename + 'Arrival_Day_{}_Summary.csv'.format(day)
summary_day = summary[summary['ARRIVAL_DAY'] == int(day)]
if day == '1':
daylist = {
'2': '1',
'3': '2',
'4': '3',
'5': '4',
'6': '5',
'0': '6',
'1': '7'
}
elif day == '2':
daylist = {
'2': '7',
'3': '1',
'4': '2',
'5': '3',
'6': '4',
'0': '5',
'1': '6'
}
elif day == '3':
daylist = {
'2': '6',
'3': '7',
'4': '1',
'5': '2',
'6': '3',
'0': '4',
'1': '5'
}
elif day == '4':
daylist = {
'2': '5',
'3': '6',
'4': '7',
'5': '1',
'6': '2',
'0': '3',
'1': '4'
}
elif day == '5':
daylist = {
'2': '4',
'3': '5',
'4': '6',
'5': '7',
'6': '1',
'0': '2',
'1': '3'
}
elif day == '6':
daylist = {
'2': '3',
'3': '4',
'4': '5',
'5': '6',
'6': '7',
'0': '1',
'1': '2'
}
elif day == '0':
daylist = {
'2': '2',
'3': '3',
'4': '4',
'5': '5',
'6': '6',
'0': '7',
'1': '1'
}
summary_day = summary_day.rename(
columns=lambda x: daylist[x[0]] + '_' + hourlist[x[2:15]] + x[15:]
if x in [col for col in slots_offered] + [
col + '_Eco' for col in slots_offered
] + [col + '_Discount' for col in slots_offered
] + [col + '_Capacity' for col in slots_offered] else x)
summary_day = summary_day.rename(
columns=lambda x: x[0] + daylist[x[1]] + '_' + hourlist[x[3:16]] +
x[16:] if x in ['C' + col for col in slots_offered] else x)
summary_day['SLOT_CHOICE'] = summary_day['SLOT_CHOICE'].fillna(
'NO_PURCHASE')
summary_day['SLOT_CHOICE'] = summary_day['SLOT_CHOICE'].apply(
lambda row: daylist[row[0]] + '_' + hourlist[row[2:15]] + row[15:]
if row in [col for col in slots_offered] else row)
summary_day.to_csv(os.path.join(
rank_folder, filename + 'Arrival_Day_{}_Summary.csv'.format(day)),
index=False)
summaries.append(summary_day)
summary = pd.concat(summaries, axis=0, sort=False)
print(slots_offered, list(summary.columns))
summary = summary[[
'EVENT_DTM', 'CUSTOMER_ID', 'ORDER_ID', 'ARRIVAL_CAT', 'ARRIVAL_DAY',
'DAY_OF_ORDER', 'CUT_OFF_1', 'CUT_OFF_2', 'SLOT_CHOICE', 'SLOT_STAMP',
'WINDOWS_STEERING', 'ZONE', 'ESTIMATED_SUBTOTAL', 'TOTAL_TIMESLOTS',
'CENSORED', 'TOTAL_ORDER'
] + [col for col in np.append('NO_PURCHASE', slots_offered_rank)] + [
'C' + col for col in np.append('NO_PURCHASE', slots_offered_rank)
] + [col + '_Eco' for col in slots_offered_rank] +
[col + '_Discount' for col in slots_offered_rank] +
[col + '_Capacity' for col in slots_offered_rank]]
summary['CUT2'] = summary['ARRIVAL_CAT'].apply(
lambda x: 1 if x == 'BEFORE_CUT2' else 0)
for day in range(0, 7):
summary[str(day) + '_ARRIVAL'] = summary['ARRIVAL_DAY'].apply(
lambda x: 1 if x == day else 0)
summary.to_csv(os.path.join(rank_folder, filename + 'SummaryNew.csv'),
index=False)