import datetime import pandas as pd import math from odps import ODPS from threading import Timer from get_data import get_data_from_odps from utils import filter_video_status from db_helper import RedisHelper from config import set_config from log import Log config_, _ = set_config() log_ = Log() project = 'loghubods' table = 'video_each_hour_update' features = [ 'videoid', 'lastonehour_preview', # 过去1小时预曝光人数 'lastonehour_view', # 过去1小时曝光人数 'lastonehour_play', # 过去1小时播放人数 'lastonehour_share', # 过去1小时分享人数 'lastonehour_return', # 过去1小时分享，过去1小时回流人数 'lastonehour_preview_total_final', # 过去1小时预曝光次数 'lastonehour_view_total_final', # 过去1小时曝光次数 'lastonehour_play_total_final', # 过去1小时播放次数 'lastonehour_share_total_final', # 过去1小时分享次数 'lastonehour_show', # 过去1小时video_show人数 'lastonehour_show_total_final', # 过去1小时video_show次数 'platform_return', ] def h_data_check(project, table, now_date): """检查数据是否准备好""" odps = ODPS( access_id=config_.ODPS_CONFIG['ACCESSID'], secret_access_key=config_.ODPS_CONFIG['ACCESSKEY'], project=project, endpoint=config_.ODPS_CONFIG['ENDPOINT'], connect_timeout=3000, read_timeout=500000, pool_maxsize=1000, pool_connections=1000 ) try: dt = datetime.datetime.strftime(now_date, '%Y%m%d%H') sql = f'select * from {project}.{table} where dt = {dt}' with odps.execute_sql(sql=sql).open_reader() as reader: data_count = reader.count except Exception as e: data_count = 0 return data_count def get_rov_redis_key(now_date): # 获取rov模型结果存放key redis_helper = RedisHelper() now_dt = datetime.datetime.strftime(now_date, '%Y%m%d') key_name = f'{config_.RECALL_KEY_NAME_PREFIX}{now_dt}' if not redis_helper.key_exists(key_name=key_name): pre_dt = datetime.datetime.strftime(now_date - datetime.timedelta(days=1), '%Y%m%d') key_name = f'{config_.RECALL_KEY_NAME_PREFIX}{pre_dt}' return key_name def get_feature_data(now_date): """获取特征数据""" dt = datetime.datetime.strftime(now_date, '%Y%m%d%H') # dt = '2022041310' records = get_data_from_odps(date=dt, project=project, table=table) feature_data = [] for record in records: item = {} for feature_name in features: item[feature_name] = record[feature_name] feature_data.append(item) feature_df = pd.DataFrame(feature_data) return feature_df def cal_score(df, param): """ 计算score :param df: 特征数据 :param param: 规则参数 :return: """ # score计算公式: sharerate*backrate*logback*ctr # sharerate = lastonehour_share/(lastonehour_play+1000) # backrate = lastonehour_return/(lastonehour_share+10) # ctr = lastonehour_play/(lastonehour_view+1000), 对ctr限最大值：K2 = 0.6 if ctr > 0.6 else ctr # score = sharerate * backrate * LOG(lastonehour_return+1) * K2 df = df.fillna(0) df['share_rate'] = df['lastonehour_share'] / (df['lastonehour_play'] + 1000) df['back_rate'] = df['lastonehour_return'] / (df['lastonehour_share'] + 10) df['log_back'] = (df['lastonehour_return'] + 1).apply(math.log) if param.get('view_type', None) == 'pre-view': df['ctr'] = df['lastonehour_play'] / (df['lastonehour_preview'] + 1000) elif param.get('view_type', None) == 'video-show': df['ctr'] = df['lastonehour_play'] / (df['lastonehour_show'] + 1000) else: df['ctr'] = df['lastonehour_play'] / (df['lastonehour_view'] + 1000) df['K2'] = df['ctr'].apply(lambda x: 0.6 if x > 0.6 else x) df['score'] = df['share_rate'] * df['back_rate'] * df['log_back'] * df['K2'] df['platform_return_rate'] = df['platform_return'] / df['lastonehour_return'] df = df.sort_values(by=['score'], ascending=False) return df def cal_score2(df): # score2计算公式: score = lastonehour_return/(lastonehour_view+1000) df = df.fillna(0) df['score'] = df['lastonehour_return'] / (df['lastonehour_view'] + 1000) df['platform_return_rate'] = df['platform_return'] / df['lastonehour_return'] df = df.sort_values(by=['score'], ascending=False) return df def cal_score3(df): # score3计算公式: # score = lastonehour_share_total_final/(lastonehour_view+1000) # + 0.03 * lastonehour_return/(lastonehour_share_total_final+1) df = df.fillna(0) df['share_rate'] = df['lastonehour_share_total_final'] / (df['lastonehour_view'] + 1000) df['back_rate'] = df['lastonehour_return'] / (df['lastonehour_share_total_final'] + 1) df['score'] = df['share_rate'] + 0.03 * df['back_rate'] df['platform_return_rate'] = df['platform_return'] / df['lastonehour_return'] df = df.sort_values(by=['score'], ascending=False) return df def video_rank(df, now_date, now_h, rule_key, param): """ 获取符合进入召回源条件的视频，与每日更新的rov模型结果视频列表进行合并 :param df: :param now_date: :param now_h: :param rule_key: 小时级数据进入条件 :param param: 小时级数据进入条件参数 :return: """ # 获取rov模型结果 redis_helper = RedisHelper() key_name = get_rov_redis_key(now_date=now_date) initial_data = redis_helper.get_all_data_from_zset(key_name=key_name, with_scores=True) log_.info(f'initial data count = {len(initial_data)}') # 获取符合进入召回源条件的视频，进入条件：小时级回流>=20 && score>=0.005 return_count = param.get('return_count') score_value = param.get('score_rule') platform_return_rate = param.get('platform_return_rate', 0) h_recall_df = df[(df['lastonehour_return'] >= return_count) & (df['score'] >= score_value) & (df['platform_return_rate'] >= platform_return_rate)] h_recall_df['videoid'] = h_recall_df['videoid'].astype(int) h_recall_videos = h_recall_df['videoid'].to_list() log_.info(f'h_recall videos count = {len(h_recall_videos)}') # 视频状态过滤 filtered_videos = filter_video_status(h_recall_videos) log_.info('filtered_videos count = {}'.format(len(filtered_videos))) # 写入对应的redis h_video_ids = [] h_recall_result = {} for video_id in filtered_videos: score = h_recall_df[h_recall_df['videoid'] == video_id]['score'] h_recall_result[int(video_id)] = float(score) h_video_ids.append(int(video_id)) h_recall_key_name = \ f"{config_.RECALL_KEY_NAME_PREFIX_BY_H}{rule_key}.{datetime.datetime.strftime(now_date, '%Y%m%d')}.{now_h}" if len(h_recall_result) > 0: redis_helper.add_data_with_zset(key_name=h_recall_key_name, data=h_recall_result, expire_time=23 * 3600) # 清空线上过滤应用列表 redis_helper.del_keys(key_name=f"{config_.H_VIDEO_FILER}{rule_key}") dup_to_redis(h_video_ids, now_date, now_h, rule_key) # 去重更新rov模型结果，并另存为redis中 # initial_data_dup = {} # for video_id, score in initial_data: # if int(video_id) not in h_video_ids: # initial_data_dup[int(video_id)] = score # log_.info(f"initial data dup count = {len(initial_data_dup)}") # initial_key_name = \ # f"{config_.RECALL_KEY_NAME_PREFIX_DUP_H}{rule_key}.{datetime.datetime.strftime(now_date, '%Y%m%d')}.{now_h}" # if len(initial_data_dup) > 0: # redis_helper.add_data_with_zset(key_name=initial_key_name, data=initial_data_dup, expire_time=23 * 3600) def dup_to_redis(h_video_ids, now_date, now_h, rule_key): """将小时级数据与其他召回视频池去重，存入对应的redis""" redis_helper = RedisHelper() # ##### 去重小程序相对24h数据更新结果，并另存为redis中 rule_24h_key_name = f"{config_.RECALL_KEY_NAME_PREFIX_BY_24H}rule2.{datetime.datetime.strftime(now_date, '%Y%m%d')}.{now_h}" if redis_helper.key_exists(key_name=rule_24h_key_name): rule_24h_data = redis_helper.get_all_data_from_zset(key_name=rule_24h_key_name, with_scores=True) log_.info(f'rule_24h data count = {len(rule_24h_data)}') rule_24h_dup = {} for video_id, score in rule_24h_data: if int(video_id) not in h_video_ids: rule_24h_dup[int(video_id)] = score h_video_ids.append(int(video_id)) log_.info(f"rule_24h data dup count = {len(rule_24h_dup)}") rule_24h_dup_key_name = \ f"{config_.RECALL_KEY_NAME_PREFIX_DUP_24H_H}{rule_key}.{datetime.datetime.strftime(now_date, '%Y%m%d')}.{now_h}" if len(rule_24h_dup) > 0: redis_helper.add_data_with_zset(key_name=rule_24h_dup_key_name, data=rule_24h_dup, expire_time=23 * 3600) # ##### 去重小程序模型更新结果，并另存为redis中 model_key_name = get_rov_redis_key(now_date=now_date) model_data = redis_helper.get_all_data_from_zset(key_name=model_key_name, with_scores=True) log_.info(f'model data count = {len(model_data)}') model_data_dup = {} for video_id, score in model_data: if int(video_id) not in h_video_ids: model_data_dup[int(video_id)] = score h_video_ids.append(int(video_id)) log_.info(f"model data dup count = {len(model_data_dup)}") model_data_dup_key_name = \ f"{config_.RECALL_KEY_NAME_PREFIX_DUP_H}{rule_key}.{datetime.datetime.strftime(now_date, '%Y%m%d')}.{now_h}" if len(model_data_dup) > 0: redis_helper.add_data_with_zset(key_name=model_data_dup_key_name, data=model_data_dup, expire_time=23 * 3600) def rank_by_h(now_date, now_h, rule_params): # 获取特征数据 feature_df = get_feature_data(now_date=now_date) # rank for key, value in rule_params.items(): log_.info(f"rule = {key}, param = {value}") # 计算score cal_score_func = value.get('cal_score_func', 0) if cal_score_func == 2: score_df = cal_score2(df=feature_df) elif cal_score_func == 3: score_df = cal_score3(df=feature_df) else: score_df = cal_score(df=feature_df, param=value) video_rank(df=score_df, now_date=now_date, now_h=now_h, rule_key=key, param=value) # to-csv score_filename = f"score_{key}_{datetime.datetime.strftime(now_date, '%Y%m%d%H')}.csv" score_df.to_csv(f'./data/{score_filename}') # to-logs log_.info({"date": datetime.datetime.strftime(now_date, '%Y%m%d%H'), "redis_key_prefix": config_.RECALL_KEY_NAME_PREFIX_BY_H, "rule_key": key, "score_df": score_df[['videoid', 'score']]}) def h_rank_bottom(now_date, now_h, rule_key): """未按时更新数据，用上一小时结果作为当前小时的数据""" log_.info(f"rule_key = {rule_key}") # 获取rov模型结果 redis_helper = RedisHelper() if now_h == 0: redis_dt = datetime.datetime.strftime(now_date - datetime.timedelta(days=1), '%Y%m%d') redis_h = 23 else: redis_dt = datetime.datetime.strftime(now_date, '%Y%m%d') redis_h = now_h - 1 key_prefix_list = [config_.RECALL_KEY_NAME_PREFIX_BY_H] for key_prefix in key_prefix_list: key_name = f"{key_prefix}{rule_key}.{redis_dt}.{redis_h}" initial_data = redis_helper.get_all_data_from_zset(key_name=key_name, with_scores=True) if initial_data is None: initial_data = [] final_data = dict() h_video_ids = [] for video_id, score in initial_data: final_data[video_id] = score h_video_ids.append(int(video_id)) # 存入对应的redis final_key_name = \ f"{key_prefix}{rule_key}.{datetime.datetime.strftime(now_date, '%Y%m%d')}.{now_h}" if len(final_data) > 0: redis_helper.add_data_with_zset(key_name=final_key_name, data=final_data, expire_time=23 * 3600) # 清空线上过滤应用列表 redis_helper.del_keys(key_name=f"{config_.H_VIDEO_FILER}{rule_key}") dup_to_redis(h_video_ids, now_date, now_h, rule_key) def h_timer_check(): rule_params = config_.RULE_PARAMS # return_count_list = [20, 10] now_date = datetime.datetime.today() log_.info(f"now_date: {datetime.datetime.strftime(now_date, '%Y%m%d%H')}") now_h = datetime.datetime.now().hour now_min = datetime.datetime.now().minute # if now_h == 0: # for key, _ in rule_params.items(): # h_rank_bottom(now_date=now_date, now_h=now_h, rule_key=key) # return # 查看当前小时更新的数据是否已准备好 h_data_count = h_data_check(project=project, table=table, now_date=now_date) if h_data_count > 0: log_.info(f'h_data_count = {h_data_count}') # 数据准备好，进行更新 rank_by_h(now_date=now_date, now_h=now_h, rule_params=rule_params) elif now_min > 50: log_.info('h_recall data is None, use bottom data!') for key, _ in rule_params.items(): h_rank_bottom(now_date=now_date, now_h=now_h, rule_key=key) else: # 数据没准备好，1分钟后重新检查 Timer(60, h_timer_check).start() if __name__ == '__main__': # df1 = get_feature_data() # res = cal_score(df=df1) # video_rank(df=res, now_date=datetime.datetime.today()) # rank_by_h() h_timer_check()