SimRecall/get3HotRecall.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/3_days_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")
        if len(items)<11:
            continue
        vid = int(items[1])
        view_users = int(items[4])
        view_pv = int(items[5])
        play_users = int(items[6])
        play_pv = int(items[7])
        share_users = int(items[8])
        share_pv = int(items[9])
        return_users = int(items[10])
        #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_pv
            share_users = item_info[3]+share_users
            share_pv = item_info[4]+share_pv
            return_users = item_info[5]+return_users 
            data_dict[vid] = (view_users, view_pv, play_users, play_pv, share_users, share_pv, return_users)
    #print(data_dict.items())
    info_dict = {}
    data_path = "./data/3_days_recall_hot_"+nowdate
    f = open(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]
        if view_pv<100 or view_users<10:
            continue
        share_pv = v[5]
        share_users = v[4]
        play_users = v[2]
        play_pv = v[3]
        #print("return_users:", return_users) 
        k_score = float(return_users)/(float(view_users)+30)
        #print(k_score)
        share_score = float(share_pv)/(float(view_pv)+100)
        backrate = float(return_users)/(float(view_users)+30)
        ctr_score = float(play_pv)/float(view_pv+100)
        if ctr_score<=0.5:
            continue
        #print(k, k_score, share_score*backrate, share_score, backrate) 
        score_info = [k_score, share_score*backrate, share_score, backrate, ctr_score, view_users, view_pv, play_users, play_pv, share_users, share_pv, return_users]
        info_dict[k] = score_info
    sorted_v = sorted(info_dict.items(), key=lambda s:s[1][1], reverse=True) 
    print("sorted_v:", sorted_v[:30])
    recall_name = "hot_3day:"
    hot_recall_res = []
    for item in sorted_v[:5]:
        hot_recall_res.append((item[0], item[1][1]))
    if len(hot_recall_res)>5:
        score_info = json.dumps(hot_recall_res)
        print("score_info:", score_info)
        redis_helper.set_data_to_redis(recall_name, score_info, 60*60*24*15)
        f.write(recall_name+"\t"+score_info+"\n")
    f.close()
    #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()