rov-offline/relevant_top_videos.py

import pandas as pd
import json
import time
from config import set_config
from db_helper import RedisHelper

config_, _ = set_config()


def get_relevant_videos_with_excel():
    filepath = './data/relevant.xlsx'
    df = pd.read_excel(filepath)
    relevant_videos = {}
    head_videos = []
    for i in range(len(df)):
        dt = df.iloc[i]['dt']
        head_vid = int(df.iloc[i]['headVid'])
        order = int(df.iloc[i]['order'])
        recommend_vid = int(df.iloc[i]['recommendVid'])
        start_time = int(df.iloc[i]['startTime'].timestamp())
        finish_time = int(df.iloc[i]['finishTime'].timestamp())
        item = {
            'order': order,
            'recommend_vid': recommend_vid,
            'start_time': start_time,
            'finish_time': finish_time
        }
        if head_vid in head_videos:
            relevant_videos[head_vid].append(item)
        else:
            relevant_videos[head_vid] = [item]
            head_videos.append(head_vid)

    print(head_videos)
    return relevant_videos


def update_relevant_videos_to_redis(relevant_videos):
    if not relevant_videos:
        return
    for head_vid, videos in relevant_videos.items():
        # 拼接key
        key_name = '{}{}'.format(config_.RELEVANT_VIDEOS_WITH_OP_KEY_NAME, head_vid)
        # 将数据转换为json
        videos_json = json.dumps(videos)
        # 以最晚结束的视频的结束时间 - 当前时间 + 5s 作为key的过期时间
        finish_time_list = [item['finish_time'] for item in videos]
        expire_time = max(finish_time_list) - int(time.time()) + 5
        if expire_time <= 0:
            return
        # 存入redis
        redis_helper = RedisHelper()
        redis_helper.set_data_to_redis(key_name=key_name, value=videos_json, expire_time=expire_time)


if __name__ == '__main__':
    relevant_videos = get_relevant_videos_with_excel()
    update_relevant_videos_to_redis(relevant_videos=relevant_videos)