SimRecall/calHourData.py

#coding utf-8
import sys
from operator import itemgetter
import json
import pandas as pd
#from db_help import RedisHelper

if __name__=="__main__":
    #1.load data
    nowdate=sys.argv[1]
    vlog='0'
    love_live = 4
    data_path = "./data/hour_video_data_"+nowdate
    f = open(data_path)
    #data = pd.read_csv(data_path, encoding="utf-8", sep='\t')
    #print(data)
    index = 0
    data_dict = {}
    #redis_helper = RedisHelper()
    while True:
        line = f.readline()
        if not line:
            break
        if index==0:
            index += 1
            continue
        index +=1
        items = line.strip().split("\t")
        #print(items)
        if len(items)<9:
            continue
        vid = items[1]
        view_users = int(items[2])
        view_pv = int(items[3])
        play_users = int(items[4])
        play_pv = int(items[5])
        share_users = int(items[6])
        share_pv = int(items[7])
        return_users = int(items[8])
        #print(items)
        if vid not in data_dict:
            data_dict[vid] = (view_users, view_pv, play_users, play_pv, share_users, share_pv, return_users)
        else:
            item_info = data_dict[vid]
            view_users = item_info[0]+view_users
            view_pv = item_info[1]+view_pv
            play_users = item_info[2]+play_users
            play_pv =  item_info[3]+play_pv
            share_users = item_info[4]+share_users
            share_pv = item_info[5]+share_pv
            return_users = item_info[6]+return_users 
            data_dict[vid] = (view_users, view_pv, play_users, play_pv, share_users, share_pv, return_users)
    #print(data_dict.items())
    f.close()
    info_dict = {}
    hour_data_path = "./data/sorted_hour_info_"+nowdate
    f = open(hour_data_path, 'w')
    for k, v in data_dict.items():
        #print(v)
        return_users = v[6]
        #print(return_users)
        view_users = v[0]
        view_pv = v[1]
        share_pv = v[5]
        share_users = v[4]
        play_users = v[2]
        #print("return_users:", return_users) 
        k_score = float(return_users)/(float(view_users)+5)
        k_score2 = float(return_users)/(float(view_pv)+5)
        #print(k_score)
        share_score = float(share_pv)/(float(view_pv)+5)
        backrate = float(return_users)/(float(view_users)+5)
        #print(k, k_score, share_score*backrate, share_score, backrate) 
        score_info = [k_score2, share_score*backrate, share_score, backrate, view_users, view_pv, play_users, play_pv, share_users, share_pv, return_users]
        score_info = json.dumps(score_info)
        info_dict[k] = score_info
        f.write(k+"\t"+score_info+"\n")
    #redis_helper.update_batch_setnx_key(info_dict, 60*60*24*15) 
    f.close()