data_cleaning/task/summary_funnel.py
2021-02-01 18:22:59 +08:00

59 lines
2.4 KiB
Python

import traceback
from pymongo import UpdateOne
from pydantic import Field
import pandas as pd
from .task import Task
from utils import *
from model import GBaseModel, IntStr, MdbObjectId
class SummaryFunnel(Task):
"""
功能分析
"""
class Model(GBaseModel):
id: MdbObjectId = Field(..., title="id", alias='_id')
cdate: int = Field(..., title='当天0点')
def cleaning(self, cursor_list):
funnel_conf = self.local_db['attr'].find_one({'pname': 'funnel'}, {'data': True})
if not funnel_conf:
ddsend_msg(f'{self.game_name} 请设置漏斗配置')
return
step_list = list({step['cond']['step'] for step in funnel_conf['data']['steps'] if step['cond']['step']})
for cursor in cursor_list: # type:dict
for source_coll, ts in cursor.items(): # type:str,dict
if ts['cursor_st'] == ts['cursor_et']:
continue
logger.info(f'开始处理{self.game_name} 处理 {source_coll} 游标 {ts}')
where = {
'_event_name': 'Guide',
'step': {'$in': step_list},
'_event_time': {
'$gte': ts['cursor_st'],
'$lt': ts['cursor_et'],
}
}
bulk_data = []
for item in self.local_db[source_coll].find(where): # 所有字段
try:
item['cdate'] = int(pd.Timestamp(item['_event_time'], unit='s', tz=self.timezone) \
.normalize().timestamp())
model = self.Model(**item)
data = model.dict(by_alias=True)
data.update(item)
bulk_data.append(UpdateOne({'_id': data['_id']}, {'$set': data}, upsert=True))
except Exception as e:
msg = traceback.format_exc()
ddsend_msg(f'{self.game_name}.{source_coll}字段异常 {msg}')
logger.error(repr(e))
if bulk_data:
self.local_db[self.dest_coll].bulk_write(bulk_data, ordered=False)
self.remote_db[self.dest_coll].bulk_write(bulk_data, ordered=False)
self.set_cursor(cursor_st=ts['cursor_st'], cursor_et=ts['cursor_et'])