python | April 04, 2020
1) limit : 몇 순위까지 보여줄 것인지 선택
2) time_delta : 최근 얼마간의 data를 보여줄 것인지 선택
3) pizza_id : 피자별로 선택
4) order_by : 어떤 기준으로 순위를 보여줄 것인지 선택
- total_count => 누적 판수 기준
- average_time => 평균 시간 기준
- shortest_time => 최단 시간 기준
- completion_score => 완성도 기준
- total_score => 위의 4개를 합친 총점 기준import pandas as pd
from sklearn.preprocessing import MinMaxScaler
def get_ranking(ranking_list, order_by, limit):
completion_importance, time_importance, count_importance = 50, 30, 20
scores = pd.DataFrame({
"completion_score" : [rank.completion_score for rank in ranking_list],
"time_score" : [-rank.average_time for rank in ranking_list],
"count_score" : [rank.total_count for rank in ranking_list],
})
scores[ : ] = MinMaxScaler().fit_transform(scores[ : ])
total_score = (
scores.completion_score * completion_importance
+ scores.time_score * time_importance
+ scores.count_score * count_importance
)
rank_table = pd.DataFrame(ranking_list.values())
rank_table['total_score'] = total_score
ascending = (
True
if order_by == 'average_time' or order_by == 'shortest_time'
else False
)
ordered_table = rank_table.sort_values(order_by, ascending = ascending)[:limit]
return ordered_tableUser별로 가진 score 정보들의 평균, 총합 등을 계산해서 return하기 위해 annotate 사용
User.objects.annotate(total_count = Count('score'))dataframe의 id 필드의 값이 1이다
dataframe['id'] == 1dataframe에서 위의 조건을 적용한 row의 total_score 가져오기
dataframe[dataframe['id'] == 1]['total_score']위의 방법으로 return하면 dataframe 형태로 return이 되어서, 해당 값(total_score)만 가져오도록 float 사용
float(dataframe[dataframe['id'] == 1]['total_score'])original 방법
dictionary에서 값을 적을 때, 전체 변수명을 다 적어줘야 한다.
[{"key" : ranking_list.get(id = id_number)}
for ranking_list.get(id = id_number) in dataframe]변수명 바꾸기 (if :=)
ranking_list.get(id = id_number)라는 변수를 user에 새로 할당해줘서 변수명 ‘user’를 사용한다.
[{"key" : user.name}
for id_number in dataframe
if (user := ranking_list.get(id = id_number))]from django.db.models import Sum, Avg, Min, Count
class UserRankView(View): #Thinking_2
def get(self, request):
limit = request.GET.get('limit', User.objects.count())
pizza_id = request.GET.get('pizza_id')
order_by = request.GET.get('order_by', 'total_score')
time_delta = request.GET.get('time_delta')
ranking_list = (
User
.objects
.filter(**get_filter_condition(pizza_id, time_delta))
.select_related('store')
.prefetch_related('score_set')
.annotate(
total_count = Count('score'),
average_time = Avg('score__time'),
shortest_time = Min('score__time'),
average_quality = Avg('score__quality'),
average_sauce = Avg('score__sauce'),
average_cheese = Avg('score__cheese'),
average_topping = Avg('score__topping'),
completion_score = sum(
[
Avg('score__quality'),
Avg('score__sauce'),
Avg('score__cheese'),
Avg('score__topping')
] #Thinking_1
)
)
)
ordered_table = get_ranking(ranking_list, order_by, int(limit))
user_ranking = [
{
"id" : user.id,
"name" : user.name,
"image" : user.image,
"store_id" : user.store_id,
"store_name" : user.store.name,
"count" : user.total_count,
"average_time" : round(user.average_time),
"shortest_time" : user.shortest_time,
"quality" : round(user.average_quality),
"sauce" : round(user.average_sauce),
"cheese" : round(user.average_cheese),
"topping" : round(user.average_topping),
"completion_score" : round(user.completion_score),
"total_score" : round(float(ordered_table[ordered_table['id'] == user.id]['total_score']))
} for id_number in ordered_table['id'] if (user := ranking_list.get(id = id_number))]
return JsonResponse({"ranking" : user_ranking}, status = 200)