REST API와 GraphQL API 비교

Django 프로젝트에서 REST API를 구현하고, GraphQL과 비교해봤다.

작업 브랜치: restapi

REST API 구현

1. Recipe, Ingredient, RecipeIngredient에 대한 Serializer 생성

from rest_framework import serializers
from .models import Recipe, Ingredient, RecipeIngredient 


class RecipeSerializer(serializers.ModelSerializer):
    class Meta:
        model = Recipe
        fields = [
            'id',
            'user',
            'title',
            'description',
            'preparation_time',
            'cooking_time',
            'difficulty_level'
        ]


class IngredientSerializer(serializers.ModelSerializer):
    class Meta:
        model = Ingredient
        fields = [
            'id',
            'name'
        ]


class RecipeIngredientSerializer(serializers.ModelSerializer):
    recipe = RecipeSerializer
    Ingredient =IngredientSerializer

    class Meta:
        model = RecipeIngredient
        fields = [
            'recipe',
            'ingredient',
            'quantity',
            'unit'
        ]

1. views.py에 ListCreateAPIView 생성

from rest_framework import generics
from .serializers import RecipeSerializer, IngredientSerializer, RecipeIngredientSerializer


class RecipeListCreate(generics.ListCreateAPIView):
    queryset = Recipe.objects.all()
    serializer_class = RecipeSerializer

class IngredientListCreate(generics.ListCreateAPIView):
    queryset = Ingredient.objects.all()
    serializer_class = IngredientSerializer

class RecipeIngredientListCreate(generics.ListCreateAPIView):
    queryset = RecipeIngredient.objects.all()
    serializer_class = RecipeIngredientSerializer

1. urls.py 설정

    path('rest-api/recipe/', views.RecipeListCreate.as_view()),
    path('rest-api/ingredient/', views.IngredientListCreate.as_view()),
    path('rest-api/recipe-ingredient/', views.RecipeIngredientListCreate.as_view())

REST API와 GraphQL API 비교

환경 설정

import requests
import pandas as pd

Over fetching

모든 레시피의 id, title, difficultyLevel를 조회하는 경우를 살펴보자.

REST API

recipe_list_url = 'http://localhost:8000/rest-api/recipe/'

response = requests.get(recipe_list_url)

recipe_list = response.json()
df = pd.DataFrame(recipe_list)
df.head()

id	user	title	description	preparation_time	cooking_time	difficulty_level
1	2	Korean Bibimbap zz	Bibimbap is a signature Korean…	30	25	Moderate
2	2	Kimchijeon	Kimchijeon is a savory Korean …	15	15	Easy
3	2	Dubu Jorim	Dubu Jorim is a flavorful Kore…	15	20	Easy
4	3	계란 후라이	egg fry	0	3	쉬움
5	2	Scramble Egg	Egg	0	3	Easy

레시피의 user, description, preparation_time, cooking_time까지 조회되게 된다.

GraphQL

recipe_query = """query MyQuery {
  allRecipes {
    id
    title
    difficultyLevel
  }
}"""

graphql_url = 'http://localhost:8000/graphql/'  

response = requests.get(graphql_url, json={'query': recipe_query})

recipe_list = response.json()['data']['allRecipes']
df = pd.DataFrame(recipe_list)
df.head()

id	title	difficultyLevel
1	Korean Bibimbap zz	Moderate
2	Kimchijeon	Easy
3	Dubu Jorim	Easy
4	계란 후라이	쉬움
5	Scramble Egg	Easy

Under fetching

모든 레시피의 id, title, difficultyLevel와 레시피에 들어가는 재료의 정보를 조회하는 경우를 살펴보자.

REST API

recipe_list_url = 'http://localhost:8000/rest-api/recipe/'
response = requests.get(recipe_list_url)

recipe_list = response.json()
recipe_df = pd.DataFrame(recipe_list)
recipe_df.head()

id	user	title	description	preparation_time	cooking_time	difficulty_level
1	2	Korean Bibimbap zz	Bibimbap is a signature Korean…	30	25	Moderate
2	2	Kimchijeon	Kimchijeon is a savory Korean …	15	15	Easy
3	2	Dubu Jorim	Dubu Jorim is a flavorful Kore…	15	20	Easy
4	3	계란 후라이	egg fry	0	3	쉬움
5	2	Scramble Egg	Egg	0	3	Easy

recipe_ingredient_list_url = 'http://localhost:8000/rest-api/recipe-ingredient/'
response = requests.get(recipe_ingredient_list_url)

recipe_ingredient_list = response.json()
recipe_ingredient_df = pd.DataFrame(recipe_ingredient_list)
recipe_ingredient_df.head()

recipe	ingredient	quantity	unit
1	1	4	cups
1	2	200	grams
1	3	3	tablespoons
1	4	2	tablespoons
1	5	1	tablespoons

ingredient_list_url = 'http://localhost:8000/rest-api/ingredient/'
response = requests.get(ingredient_list_url)
ingredient_list = response.json()
ingredient_df = pd.DataFrame(ingredient_list)
ingredient_df.head()

id	name
1	Cooked rice
2	Beef
3	Soy sauce
4	Sesame oil
5	Sugar

cols = ['id', 'title', 'difficulty_level']
recipe_df = recipe_df[cols]

df = (
    recipe_ingredient_df
    .merge(recipe_df, left_on='recipe', right_on='id', how='left')
    .drop(columns=['id'])
    .merge(ingredient_df, left_on='ingredient', right_on='id', how='left')
    .drop(columns=['id', 'ingredient'])
    .rename(columns={'name': 'ingredient_name', 'recipe': 'id'})
)
df.head()

id	quantity	unit	title	difficulty_level	ingredient_name
1	4	cups	Korean Bibimbap zz	Moderate	Cooked rice
1	200	grams	Korean Bibimbap zz	Moderate	Beef
1	3	tablespoons	Korean Bibimbap zz	Moderate	Soy sauce
1	2	tablespoons	Korean Bibimbap zz	Moderate	Sesame oil
1	1	tablespoons	Korean Bibimbap zz	Moderate	Sugar

필요한 정보를 전부 가져오기 위해 여러 번의 요청을 보내야 한다.

GraphQL

recipe_ingredient_query = """query MyQuery {
  allRecipes {
    id
    title
    difficultyLevel
    recipeingredientSet {
      ingredient {
        name
      }
      quantity
      unit
    }
  }
}"""

response = requests.get(graphql_url, json={'query': recipe_ingredient_query})
recipe_list = response.json()['data']['allRecipes']

df = pd.json_normalize(
    recipe_list,
    record_path='recipeingredientSet',
    meta=['id', 'title', 'difficultyLevel']
)
df.head()

quantity	unit	ingredient.name	id	title	difficultyLevel
4	cups	Cooked rice	1	Korean Bibimbap zz	Moderate
200	grams	Beef	1	Korean Bibimbap zz	Moderate
3	tablespoons	Soy sauce	1	Korean Bibimbap zz	Moderate
2	tablespoons	Sesame oil	1	Korean Bibimbap zz	Moderate
1	tablespoons	Sugar	1	Korean Bibimbap zz	Moderate

한번의 요청으로 필요한 데이터를 전부 가져온다.