1. 程式人生 > >揭開DRF序列化技術的神祕面紗

揭開DRF序列化技術的神祕面紗

在RESTful API中,介面返回的是JSON,JSON的內容對應的是資料庫中的資料,DRF是通過序列化(Serialization)的技術,把資料模型轉換為JSON的,反之,叫做反序列化(deserialization)。本文就來揭開DRF序列化技術的神祕面紗。 # 建立虛擬環境 虛擬環境是獨立的Python環境,可以和系統環境分離,只安裝需要的包即可,使用以下命令建立並激活: ```shell # 建立Python虛擬環境 python -m venv env # 啟用虛擬環境 env\Scripts\activate.bat # Mac中使用`source env/bin/activate` ``` > 退出虛擬環境用deactivate。 接著在虛擬環境中安裝本文用到的包: ```python pip install django pip install djangorestframework pip install pygments # 程式碼高亮 ``` # 建立專案 建立project: ```shell django-admin startproject tutorial ``` 建立app: ```shell cd tutorial python manage.py startapp snippets ``` 在`tutorial/settings.py`中的`INSTALLED_APPS`新增`snippets`和`rest_framework`: ```python INSTALLED_APPS = [ ... 'rest_framework', 'snippets.apps.SnippetsConfig', ] ``` > 只有新增後,app才會生效哦。 # 建立model 編輯`snippets/models.py`,建立Snippet模型: ```python from django.db import models from pygments.lexers import get_all_lexers from pygments.styles import get_all_styles LEXERS = [item for item in get_all_lexers() if item[1]] LANGUAGE_CHOICES = sorted([(item[1][0], item[0]) for item in LEXERS]) STYLE_CHOICES = sorted([(item, item) for item in get_all_styles()]) class Snippet(models.Model): created = models.DateTimeField(auto_now_add=True) title = models.CharField(max_length=100, blank=True, default='') code = models.TextField() linenos = models.BooleanField(default=False) language = models.CharField(choices=LANGUAGE_CHOICES, default='python', max_length=100) style = models.CharField(choices=STYLE_CHOICES, default='friendly', max_length=100) class Meta: ordering = ['created'] ``` Snippet有6個欄位,created、title、code、linenos、language、style。 接著同步資料庫: ```python python manage.py makemigrations snippets python manage.py migrate ``` # 建立Serializer rest_framework.serializers的結構如下:
> 檢視Python模組中的成員,需要勾選: > >
之前我們用過HyperlinkedModelSerializer,是封裝後的符合RESTful良好設計的序列化器,這裡直接使用更底層的Serializer。在`snippets`目錄下建立`serializers.py`,新增以下程式碼: ```python from rest_framework import serializers from snippets.models import Snippet, LANGUAGE_CHOICES, STYLE_CHOICES class SnippetSerializer(serializers.Serializer): id = serializers.IntegerField(read_only=True) title = serializers.CharField(required=False, allow_blank=True, max_length=100) code = serializers.CharField(style={'base_template': 'textarea.html'}) linenos = serializers.BooleanField(required=False) language = serializers.ChoiceField(choices=LANGUAGE_CHOICES, default='python') style = serializers.ChoiceField(choices=STYLE_CHOICES, default='friendly') def create(self, validated_data): """ Create and return a new `Snippet` instance, given the validated data. """ return Snippet.objects.create(**validated_data) def update(self, instance, validated_data): """ Update and return an existing `Snippet` instance, given the validated data. """ instance.title = validated_data.get('title', instance.title) instance.code = validated_data.get('code', instance.code) instance.linenos = validated_data.get('linenos', instance.linenos) instance.language = validated_data.get('language', instance.language) instance.style = validated_data.get('style', instance.style) instance.save() return instance ``` 程式碼第一部分定義了序列化/反序列化的欄位,欄位定義方式類似於Form。第二部分的`create()`和`update()`方法定義了呼叫`serializer.save()`時,建立和更新例項的處理方式。 # 玩玩Serializer 既然已經建立好了Serializer,何不玩玩它?只有玩了才知道序列化的技術細節。這裡用到了Django shell,可以在命令列實時看到輸出結果: ```shell python manage.py shell ``` 首先通過模型新增2條測試資料: ```python from snippets.models import Snippet from snippets.serializers import SnippetSerializer from rest_framework.renderers import JSONRenderer from rest_framework.parsers import JSONParser snippet = Snippet(code='foo = "bar"\n') snippet.save() snippet = Snippet(code='print("hello, world")\n') snippet.save() ```
接著把它序列化看看: ```python serializer = SnippetSerializer(snippet) serializer.data # {'id': 2, 'title': '', 'code': 'print("hello, world")\n', 'linenos': False, 'language': 'python', 'style': 'friendly'} ``` > 因為`snippet`變數最後引用的是第2個數據模型的例項,所以這裡只顯示第2條資料。 serializer.data返回的是個字典,最後把字典轉換為JSON: ```python content = JSONRenderer().render(serializer.data) content # b'{"id": 2, "title": "", "code": "print(\\"hello, world\\")\\n", "linenos": false, "language": "python", "style": "friendly"}' ``` 這就是序列化的過程,模型例項→Python字典→JSON。 反序列化的過程是類似的,先把JSON轉換為Python字典: ```python import io stream = io.BytesIO(content) data = JSONParser().parse(stream) ``` 再把字典轉換為模型例項: ```python serializer = SnippetSerializer(data=data) serializer.is_valid() # True serializer.validated_data # OrderedDict([('title', ''), ('code', 'print("hello, world")\n'), ('linenos', False), ('language', 'python'), ('style', 'friendly')]) serializer.save() # ``` > 因為SnippetSerializer的create()方法程式碼是`return Snippet.objects.create(**validated_data)`,所以`serializer.save()`會在資料庫中新增1條資料,並返回Snippet模型例項。 > 另外,序列化物件不只有模型例項,也可以是資料集: ```python serializer = SnippetSerializer(Snippet.objects.all(), many=True) serializer.data # [OrderedDict([('id', 1), ('title', ''), ('code', 'foo = "bar"\n'), ('linenos', False), ('language', 'python'), ('style', 'friendly')]), OrderedDict([('id', 2), ('title', ''), ('code', 'print("hello, world")\n'), ('linenos', False), ('language', 'python'), ('style', 'friendly')]), OrderedDict([('id', 3), ('title', ''), ('code', 'print("hello, world")'), ('linenos', False), ('language', 'python'), ('style', 'friendly')])] ``` # 使用ModelSerializers rest_framework.serializers.Serializers比較底層,需要重複定義model中的欄位,DRF的作者當然也想到了這個問題,並且給出瞭解決辦法:把Serializers封裝成ModelSerializers。在`snippets/serializers.py`中替換`SnippetSerializer`為: ```python class SnippetSerializer(serializers.ModelSerializer): class Meta: model = Snippet fields = ['id', 'title', 'code', 'linenos', 'language', 'style'] ``` 寫法果然簡潔了很多。使用Django shell打印出來看看: ```python from snippets.serializers import SnippetSerializer serializer = SnippetSerializer() print(repr(serializer)) # SnippetSerializer(): # id = IntegerField(label='ID', read_only=True) # title = CharField(allow_blank=True, max_length=100, required=False) # code = CharField(style={'base_template': 'textarea.html'}) # linenos = BooleanField(required=False) # language = ChoiceField(choices=[('Clipper', 'FoxPro'), ('Cucumber', 'Gherkin'), ('RobotFramework', 'RobotFramework'), ('abap', 'ABAP'), ('ada', 'Ada')... # style = ChoiceField(choices=[('autumn', 'autumn'), ('borland', 'borland'), ('bw', 'bw'), ('colorful', 'colorful')... ``` 和手動定義的欄位一模一樣。其實ModelSerializer並沒有做額外的處理,我們可以從它的部分程式碼片段看到: ```python serializer_field_mapping = { models.AutoField: IntegerField, models.BigIntegerField: IntegerField, models.BooleanField: BooleanField, # Default `create` and `update` behavior... def create(self, validated_data): def update(self, instance, validated_data): raise_errors_on_nested_writes('update', self, validated_data) ``` 建立欄位對映,定義`create()`和`update()`等方法,只做了最簡單的封裝。 # 編寫views 我們之前用的是已經封裝好的rest_framework.viewsets,這裡直接用Django原生的view。在`snippets/views.py`中新增程式碼: ```python from django.http import HttpResponse, JsonResponse from django.views.decorators.csrf import csrf_exempt from rest_framework.parsers import JSONParser from snippets.models import Snippet from snippets.serializers import SnippetSerializer ``` 先建立snippet_list檢視,返回snippets列表或者新建1個snippet: ```python @csrf_exempt def snippet_list(request): """ List all code snippets, or create a new snippet. """ if request.method == 'GET': snippets = Snippet.objects.all() serializer = SnippetSerializer(snippets, many=True) return JsonResponse(serializer.data, safe=False) elif request.method == 'POST': data = JSONParser().parse(request) serializer = SnippetSerializer(data=data) if serializer.is_valid(): serializer.save() return JsonResponse(serializer.data, status=201) return JsonResponse(serializer.errors, status=400) ``` 本文示例中,客戶端發起POST請求是不帶CSRF token的,所以需要新增csrf_exempt(exempt翻譯為免除)裝飾器,允許跨域訪問。這裡只是演示,實際會使用`django-cors-headers`來解決跨域問題,而不是給每個view都加上`@csrf_exempt`。 > CSRF是指跨站請求偽造,攻擊者盜用你的身份,以你的名義傳送惡意請求。CSRF token是指伺服器通過token來認證,如果請求中沒有token或者token不匹配,那麼就認為可能是CSRF而拒絕該請求。 接著新增一個view,對單個snippet進行刪(delete)、改(update)、查(retrieve): ```python @csrf_exempt def snippet_detail(request, pk): """ Retrieve, update or delete a code snippet. """ try: snippet = Snippet.objects.get(pk=pk) except Snippet.DoesNotExist: return HttpResponse(status=404) if request.method == 'GET': serializer = SnippetSerializer(snippet) return JsonResponse(serializer.data) elif request.method == 'PUT': data = JSONParser().parse(request) serializer = SnippetSerializer(snippet, data=data) if serializer.is_valid(): serializer.save() return JsonResponse(serializer.data) return JsonResponse(serializer.errors, status=400) elif request.method == 'DELETE': snippet.delete() return HttpResponse(status=204) ``` > 注意,前面create返回的status為201,這裡retrieve返回的預設的200,附上幾個常見狀態碼: > > - 200 OK - GET/PUT/PATCH:伺服器成功返回使用者請求的資料,該操作是冪等的(Idempotent) > - 201 CREATED - POST:使用者新建資料成功 > - 204 NO content - DELETE:使用者刪除資料成功 > - 400 Invalid request - POST/PUT/PATCH:使用者發出的請求有錯誤,伺服器沒有進行新建或修改資料的操作,該操作是冪等的 > - 401 Unauthorized - *:表示使用者沒有許可權(令牌、使用者名稱、密碼錯誤) > - 403 Forbidden - *:表示使用者得到授權(與401錯誤相對),但是訪問是被禁止的 最後,新增路由,新建檔案`snippets/urls.py`: ```python from django.urls import path from snippets import views urlpatterns = [ path('snippets/', views.snippet_list), path('snippets//', views.snippet_detail), ] ``` 再配置到`tutorial/urls.py`: ```python from django.urls import path, include urlpatterns = [ path('', include('snippets.urls')), ] ``` # 測試API 啟動專案: ```shell python manage.py runserver ``` 新開一個Terminal,安裝HTTP命令列工具: ```shell pip install httpie ``` 請求snippets列表: ```shell http http://127.0.0.1:8000/snippets/ HTTP/1.1 200 OK ... [ { "id": 1, "title": "", "code": "foo = \"bar\"\n", "linenos": false, "language": "python", "style": "friendly" }, { "id": 2, "title": "", "code": "print(\"hello, world\")\n", "linenos": false, "language": "python", "style": "friendly" } ] ``` 請求id為2的snippet: ```shell http http://127.0.0.1:8000/snippets/2/ HTTP/1.1 200 OK ... { "id": 2, "title": "", "code": "print(\"hello, world\")\n", "linenos": false, "language": "python", "style": "friendly" } ``` # 東方說 不得不說DRF的Tutorial實踐性很強,基本上照著做都能敲出來,體驗比較好。序列化的過程,就是模型例項→Python字典→JSON的過程。通過本文也理解了在app中`models.py`→`serializers.py`→`views.py`→`urls.py`的程式碼編寫順序。 > 參考資料: > > https://www.django-rest-framework.org/tutorial/1-serialization/ > > https://blog.csdn.net/yexudengzhidao/article/details/