#!/usr/bin/env python
# -*- coding:utf-8 -*- 
# @Time    : 2018/4/5 8:09
# @Author  : liujiantao
# @Site    : 特征集成
# @File    : feature_integrate.py
# @Software: PyCharm
import time

from data_helper import DataHelper
from ljt_feature_etl02 import LjtFeatureEtl02
from threading_util import ThreadingUtil
import pandas as pd
import numpy as np
from config import *

config = Config()


class FeatureIntegrate(object):
    """
    特征汇总
    """

    def __init__(self):
        self.d_h = DataHelper()

    def get_all_features(self, path=config.path_train01):
        start = time.time()
        data = self.d_h.get_data(path)
        min_Y = data[data['Y'] != 0]['Y'].min()
        max_Y = data[data['Y'] != 0]['Y'].max()
        ft_Liyang = LjtFeatureEtl02(self.d_h, data)
        user_Y_list = self.d_h.get_user_Y_list(data)
        all_features_list = []
        pre_features_list = []
        act_class_Y = []
        act_Y = []
        all_Y = []
        pre_Y = []
        mt = ThreadingUtil()
        g_func_list = []
        # 用户行程占比
        g_func_list.append({"func": ft_Liyang.get_features01, "args": (data,)})
        g_func_list.append({"func": ft_Liyang.get_features02, "args": (data,)})
        g_func_list.append({"func": ft_Liyang.get_features03, "args": (data,)})
        g_func_list.append({"func": ft_Liyang.get_features04, "args": (data,)})
        mt.set_thread_func_list(g_func_list)
        mt.start()
        m = 0
        n = 0
        m11 = [[row[col] for row in mt.data_list] for col in range(len(mt.data_list[0]))]
        for i in range(user_Y_list.__len__()):
            current_list = m11[i]
            Y = user_Y_list[i]
            all_Y.append(Y)
            if Y == 0:
                if m >= 1 * n: continue
                act_class_Y.append(Y)
                all_features_list.append(current_list)
                act_Y.append(Y)
                m += 1
            else:
                act_class_Y.append(1)
                pre_features_list.append(current_list)
                all_features_list.append(current_list)
                act_Y.append(Y)
                pre_Y.append(Y)
                n += 1
        self.d_h.print_str += "feature_intergrate cost time: " + str(time.time() - start) + " | "
        return all_features_list, act_class_Y, pre_features_list, pre_Y, act_Y, all_Y, min_Y,max_Y

    def get_test_features(self, path=config.path_test01):
        """

        :param path:
        :return:
        """
        data = self.d_h.get_test_data(path)
        userid_list = self.d_h.get_userlist(data)
        ft_Liyang = LjtFeatureEtl02(self.d_h, data)
        mt = ThreadingUtil()
        g_func_list = []
        g_func_list.append({"func": ft_Liyang.get_features01, "args": (data,)})
        g_func_list.append({"func": ft_Liyang.get_features02, "args": (data,)})
        g_func_list.append({"func": ft_Liyang.get_features03, "args": (data,)})
        g_func_list.append({"func": ft_Liyang.get_features04, "args": (data,)})
        mt.set_thread_func_list(g_func_list)
        mt.start()
        test_features = [[row[col] for row in mt.data_list] for col in range(len(mt.data_list[0]))]
        return userid_list, test_features