#-*- coding: utf-8 -*-
#__author__ = "dukun"

import json
import time
import pymysql
import logging
import datetime
import numpy as np
import pandas as pd
import configparser
from celery_tasks import cel
#from DBUtils.PooledDB import PooledDB
from dbutils.pooled_db import PooledDB


#配置输出日志格式
LOG_FORMAT = "%(asctime)s %(filename)s[line:%(lineno)d] %(name)s %(levelname)s %(pathname)s %(message)s "

#配置输出时间的格式，注意月份和天数不要搞乱了
DATE_FORMAT = '%Y-%m-%d  %H:%M:%S %a ' 
logging.basicConfig(level=logging.INFO,
                    format=LOG_FORMAT,
                    datefmt = DATE_FORMAT ,
                    filename=r"./logs/channel.log" #有了filename参数就不会直接输出显示到控制台，而是直接写入文件
                    )



cf = configparser.RawConfigParser()
#cf = configparser.ConfigParser()
cf.read(r"./config.ini")  # 读取配置文件，如果写文件的绝对路径，就可以不用os模块
CHANNEL_IDS =  json.loads(cf.get("channel_ids","channel_ids"))

CHANNEL_HOST = cf.get("database1","host")
CHANNEL_PORT = int(cf.get("database1","port"))
CHANNEL_USER = cf.get("database1","user")
CHANNEL_PASSWORD = cf.get("database1","password")
CHANNEL_DB = cf.get("database1","db")

STAT_HOST = cf.get("database2","host")
STAT_PORT = int(cf.get("database2","port"))
STAT_USER = cf.get("database2","user")
STAT_PASSWORD = cf.get("database2","password")
STAT_DB = cf.get("database2","db")


#数据库连接信息
channel_db_config = {
    'host' : CHANNEL_HOST,
    'port' : CHANNEL_PORT,
    'user' : CHANNEL_USER,
    'password' : CHANNEL_PASSWORD,
    'db' : CHANNEL_DB,
    'charset' : 'utf8',
    'autocommit' : 1
}


stat_db_config = {
    'host' : STAT_HOST,
    'port' : STAT_PORT,
    'user' : STAT_USER,
    'password' : STAT_PASSWORD,
    'db' : STAT_DB,
    'charset' : 'utf8',
    'autocommit' : 1
}






#创建数据库连接池
def createPool(db_config):
   spool = PooledDB(pymysql, 5, **db_config) 
   return spool


@cel.task
def get_channel_use_time():
    conn = createPool(channel_db_config).connection()
    stat_conn = createPool(stat_db_config).connection()

    channel_sql = """SELECT trans_id,channel_name,flow_amount,use_time,status,cnt
			FROM	
			(SELECT trans_id, flow_amount,SUM(UNIX_TIMESTAMP(modify_date) - UNIX_TIMESTAMP(create_date)) use_time,
                        send_status status, COUNT(*) cnt   
			FROM 
      				mobile_flow_dispatch_rec 
			WHERE  UNIX_TIMESTAMP(now())-UNIX_TIMESTAMP(create_date) <= (2 * 3600)  
    				GROUP BY trans_id,status) t1  LEFT JOIN	access_channel_info 
			ON 	access_channel_info.channel_seq_id = t1.trans_id  
    			ORDER BY channel_name  ; """

    ins_sql = """ INSERT INTO channel_use_time VALUES(0,%s,%s,%s,%s,%s,%s,%s) """


    df = pd.read_sql(channel_sql,conn)
    #print(df)

    #channel_ids = CHANNEL_IDS
    
    ins_list = []
    faces = [30,50,100,200,300,500]
    date = datetime.datetime.now().strftime("%Y-%m-%d %H:%M")
    
    for channel_id in CHANNEL_IDS:
        channel_df = df[df['trans_id'] == channel_id]
        if channel_df.empty is True:
            continue

        channel_name = channel_df['channel_name'].iloc[0]
        operator,channel_name = get_operator(channel_name)
        for face in faces:
            face_df = channel_df[channel_df['flow_amount'] == face]
            if face_df.empty is True:
               continue

            succ_df = face_df[face_df['status'] == '2']
            fail_df = face_df[face_df['status'] == '4']

            #成功平均时长
            if succ_df.empty is False:
                succ_time = succ_df['use_time'].iloc[0]
                #print(succ_time)
                succ_count = succ_df['cnt'].iloc[0]
                succ_avg_time = float(succ_time/succ_count)
            else:
                succ_avg_time = 0

            if fail_df.empty is False:
                fail_time = fail_df['use_time'].iloc[0]
                fail_count = fail_df['cnt'].iloc[0]
                fail_avg_time = float(fail_time/fail_count)
            else:
                fail_avg_time = 0

            total_df = face_df.groupby('trans_id').agg({'use_time' : 'sum','cnt' : 'sum'})
            #print(total_df)
            total_time = total_df['use_time'].iloc[0]
            total_count = total_df['cnt'].iloc[0]
            total_avg_time = float(total_time / total_count)

            #print(total_avg_time)
            #print(succ_avg_time)
            #print(fail_avg_time)

            ins_list.append((date,channel_id,operator,channel_name,face,1,total_avg_time)) #合计平均时长
            ins_list.append((date,channel_id,operator,channel_name,face,2,succ_avg_time)) #成功平均时长
            ins_list.append((date,channel_id,operator,channel_name,face,3,fail_avg_time)) #失败平均时长
   
    
    stat_cursor = stat_conn.cursor()
    stat_cursor.executemany(ins_sql,ins_list)
    
    stat_conn.rollback()
    #print('数据回滚')



    conn.close()
    stat_conn.close()
    print(3333333333333333)

def get_operator(channel_name):
    operator = '移动'

    if '移动' in channel_name:
        operator = '移动'
    elif '联通' in channel_name:
        operator = '联通'
    elif '电信' in channel_name:
        operator = '电信'

    channel_name = channel_name.replace(operator,"")

    return operator,channel_name

	
get_channel_use_time()