目标函数

添加链接描述

添加链接描述

exchange = pd.read_csv('exchange.csv', skiprows = 1, sep = '\t')
exchange =  exchange[exchange['position'] == 'default']
exchange['item'] = exchange['item'].fillna('')
exchange['gold'] = exchange['gold'].fillna(0)
columns = ['charge_id', 'type', 'charge', 'us_price', 'gold',  'item']
bag_info = exchange[columns]
bag_info['charge_id'] = bag_info['charge_id'].astype(np.int32)
bag_info = bag_info.sort_values('charge_id')

default_charge = pd.DataFrame(data = np.array([0, 0, 0, 0, 0 ,'']).reshape(1, 6), columns = bag_info.columns)
bag_info = default_charge.append(bag_info, ignore_index = True)

import pandas as pd
import glob
data_dir = './data/'
file_paths = glob.glob(data_dir + '*.zip')
file_paths = sorted(file_paths)
df_list = []

print(len(file_paths))
for path in file_paths[-4:]:
    df_list.append(pd.read_csv(path, 
            names = ['index', 'player_id', 'giftID', 'state', 'offer_time'], 
            sep = ',', compression='zip', quotechar='"'))
log = pd.concat(df_list)
del df_list
log = log[log['giftID'].isin(exchange['charge_id'])]
print(log.shape)

import json
def parse_log(log):
    feature = log[['index', 'player_id', 'giftID', 'offer_time']]
    feature_dict = dict()
    keys = json.loads(log['state'].iloc[0]).keys()
    for key in keys:
        feature_dict[key] = []
    for line in log['state']:
        data = json.loads(line)
        for key in keys:
            feature_dict[key].append(data[key])
    
    for key in keys:
        feature[key] = feature_dict[key]
        
    return feature