更新留存

api/api_v1/endpoints/data_auth.py

@@ -43,6 +43,7 @@ async def edit_data_auth(request: Request,
     await crud.data_auth.edit_data_auth(db, data_id)
     return schemas.Msg(code=0, msg='ok', data=data_id)
 
+
 @router.get("/quotas_map")
 async def quotas_map(
         request: Request,
@@ -103,7 +104,7 @@ async def my_event(request: Request,
                    current_user: schemas.UserDB = Depends(deps.get_current_user)
                    ) -> schemas.Msg:
     """获取自己的事件权限"""
-
+    event_list = []
     data_auth_id = await crud.authority.get_data_auth_id(db, game, request.user.username)
 
     my_data_auth = []
@@ -112,6 +113,7 @@ async def my_event(request: Request,
         if data_auth_id == '*':
             my_data_auth = await ck.distinct(game, 'event', '#event_name')
 
+
         else:
             # 设置了数据权限
             my_data_auth = await crud.data_auth.get(db, ObjectId(data_auth_id))
@@ -120,8 +122,6 @@ async def my_event(request: Request,
         if not my_data_auth:
             return schemas.Msg(code=0, msg='ok', data=[])
 
-    event_list = []
-
     event_show_name = await crud.event_mana.get_all_show_name(db, game)
     event_list.append({'id': 'event', 'title': '全部事件', 'category': []})
     for item in my_data_auth:
@@ -129,6 +129,7 @@ async def my_event(request: Request,
             'event_name': item,
             'event_desc': event_show_name.get(item, item)
         })
+    event_list[-1]['category'].append({'event_name': '*', 'event_desc': '任意事件'})
     event_list.sort()
     return schemas.Msg(code=0, msg='ok', data=event_list)
 
@@ -146,7 +147,7 @@ async def user_property(request: Request,
     data = json.loads(data)
     propertys = []
 
-    data_attr = await crud.data_attr.find_many(db, {'game':game, 'cat':'user'})
+    data_attr = await crud.data_attr.find_many(db, {'game': game, 'cat': 'user'})
     data_attr = {item['name']: item for item in data_attr}
 
     for k, v in data.items():
@@ -162,9 +163,6 @@ async def user_property(request: Request,
     return schemas.Msg(code=0, msg='ok', data=propertys)
 
 
-
-
-
 @router.post('/load_prop_quotas')
 async def load_prop_quotas(request: Request,
                            game: str,
@@ -181,7 +179,7 @@ async def load_prop_quotas(request: Request,
 
     all_filed = await rdb.get(f'{game}_event')
     all_filed = json.loads(all_filed)
-    data_attr = await crud.data_attr.find_many(db, {'game':game, 'cat':'event'})
+    data_attr = await crud.data_attr.find_many(db, {'game': game, 'cat': 'event'})
     data_attr = {item['name']: item for item in data_attr}
     event_props = []
     for item in event_prop_list:
@@ -274,7 +272,7 @@ async def load_filter_props(request: Request,
     all_filed = await rdb.get(f'{game}_event')
     all_filed = json.loads(all_filed)
 
-    data_attr = await crud.data_attr.find_many(db, {'game':game, 'cat':'event'})
+    data_attr = await crud.data_attr.find_many(db, {'game': game, 'cat': 'event'})
     data_attr = {item['name']: item for item in data_attr}
     event_props = []
     for item in event_prop_list:
@@ -288,7 +286,7 @@ async def load_filter_props(request: Request,
         }
         event_props.append(event_prop)
 
-    data_attr = await crud.data_attr.find_many(db, {'game':game, 'cat':'user'})
+    data_attr = await crud.data_attr.find_many(db, {'game': game, 'cat': 'user'})
     data_attr = {item['name']: item for item in data_attr}
     user_props = []
     for item in user_prop_list:

api/api_v1/endpoints/query.py

@@ -72,7 +72,7 @@ async def event_model(
         q['date_range'] = date_range
         df = await ckdb.query_dataframe(sql)
         if df.shape[0] == 0:
-            return schemas.Msg(code=0, msg='ok', data=q)
+            return schemas.Msg(code=0, msg='ok', data=[q])
 
         if groupby:
             # 有分组
@@ -141,31 +141,31 @@ async def retention_model(
     concat_data = []
     df = pd.concat([df, pd.DataFrame(concat_data, columns=df.columns)])
     # 计算整体
-    summary_df = df.groupby(['date', 'event_name'])[['values', 'amount']].sum()
+    summary_df = df.groupby('date')[['val_a', 'val_b', 'amount_a']].sum()
     summary_values = {}
     for i, d1 in enumerate(date_range):
-        a = set(summary_df.loc[(d1, event_a)]['values']) if (d1, event_a) in summary_df.index else set()
+        a = set(summary_df.loc[d1]['val_a']) if d1 in summary_df.index else set()
         if not a:
             continue
         key = d1.strftime('%Y-%m-%d')
         for j, d2 in enumerate(date_range[i:]):
             if j > unit_num:
                 break
-            b = set(summary_df.loc[(d2, event_b)]['values']) if (d2, event_b) in summary_df.index else set()
+            b = set(summary_df.loc[d2]['val_b']) if d2 in summary_df.index else set()
             tmp = summary_values.setdefault(key, {})
             tmp.setdefault('d0', len(a))
             tmp.setdefault('p', []).append(division(len(a & b) * 100, len(a)))
             tmp.setdefault('n', []).append(len(a & b))
-    groups = set([tuple(i) for i in df[res['groupby'][2:]].values])
+    groups = set([tuple(i) for i in df[res['groupby']].values])
     df.set_index(res['groupby'], inplace=True)
     df.sort_index(inplace=True)
     values = {}
     days = [i for i in range((date_range[-1] - date_range[0]).days + 1)][:unit_num + 1]
     for i, d1 in enumerate(date_range):
         for g in groups:
-            if g == tuple():
+            if len(g) == 1:
                 continue
-            a = set(df.loc[(d1, event_a, *g)]['values']) if (d1, event_a, *g) in df.index else set()
+            a = set(df.loc[g]['val_a']) if g in df.index else set()
             if not a:
                 continue
             key = d1.strftime("%Y-%m-%d")
@@ -173,8 +173,8 @@ async def retention_model(
             for j, d2 in enumerate(date_range[i:]):
                 if j > unit_num:
                     break
-                b = set(df.loc[(d2, event_b, *g)]['values']) if (d2, event_b, *g) in df.index else set()
+                b = set(df.loc[g]['val_b']) if g in df.index else set()
-                tmp = tmp_g.setdefault(','.join(g), {})
+                tmp = tmp_g.setdefault(','.join(g[1:]), {})
                 tmp.setdefault('d0', len(a))
                 tmp.setdefault('p', []).append(division(len(a & b) * 100, len(a)))
                 tmp.setdefault('n', []).append(len(a & b))

models/behavior_analysis.py

@@ -44,7 +44,7 @@ class BehaviorAnalysis:
             self.event_view = report['query']['eventView']
             self.events = report['query']['events']
             e_days, s_days = self.event_view['recentDay'].split('-')
-            self.event_view['endTime'] = arrow.get().shift(days=-int(e_days)+1).strftime('%Y-%m-%d 23:59:59')
+            self.event_view['endTime'] = arrow.get().shift(days=-int(e_days) + 1).strftime('%Y-%m-%d 23:59:59')
             self.event_view['startTime'] = arrow.get().shift(days=-int(s_days)).strftime('%Y-%m-%d 00:00:00')
 
 
@@ -144,27 +144,42 @@ class BehaviorAnalysis:
     def retention_model_sql(self):
         event_name_a = self.events[0]['eventName']
         event_name_b = self.events[1]['eventName']
+        visit_name = self.events[0].get('event_attr_id')
         event_time_col = getattr(self.event_tbl.c, '#event_time')
         event_name_col = getattr(self.event_tbl.c, '#event_name')
         e_account_id_col = getattr(self.event_tbl.c, '#account_id')
         u_account_id_col = getattr(self.user_tbl.c, '#account_id')
         date_col = sa.Column('date')
+        who_visit = e_account_id_col
+        if visit_name:
+            who_visit = getattr(self.event_tbl.c, visit_name)
 
         selectd = [func.toStartOfDay(func.addHours(event_time_col, self.zone_time)).label('date'),
-                   event_name_col.label('event_name'),
                    *self.groupby,
-                   func.arrayDistinct(func.groupArray(e_account_id_col)).label('values'),
+                   func.arrayDistinct(
-                   func.length(sa.Column('values')).label('amount')
+                       (func.groupArray(func.if_(event_name_col == event_name_a, who_visit, None)))).label('val_a'),
+
+                   func.length(sa.Column('val_a')).label('amount_a'),
+                   func.length(sa.Column('val_b')).label('amount_b'),
                    ]
+
+        if event_name_b == '*':
+            val_b = func.arrayDistinct(
+                (func.groupArray(func.if_(event_name_col != event_name_a, who_visit, None)))).label('val_b'),
+            selectd.insert(-2, *val_b)
+        else:
+            val_b = func.arrayDistinct(
+                (func.groupArray(func.if_(event_name_col == event_name_b, who_visit, None)))).label('val_b'),
+            selectd.insert(-2, *val_b)
+
         base_where = [
             func.addHours(event_time_col, self.zone_time) >= self.start_date,
             func.addHours(event_time_col, self.zone_time) <= self.end_date,
-            event_name_col.in_([event_name_a, event_name_b]),
         ]
 
         event_filter, user_filter = self.handler_filts()
 
-        groupby = [date_col, event_name_col] + self.groupby
+        groupby = [date_col] + self.groupby
         oredrby = [date_col]
         if user_filter:
             qry = sa.select(selectd).select_from(
@@ -177,7 +192,7 @@ class BehaviorAnalysis:
         sql = str(qry.compile(compile_kwargs={"literal_binds": True}))
         print(sql)
         return {'sql': sql,
-                'groupby': ['date', 'event_name'] + [i.key for i in self.groupby],
+                'groupby': ['date'] + [i.key for i in self.groupby],
                 'date_range': self.date_range,
                 'event_name': [event_name_a, event_name_b],
                 'unit_num': self.unit_num