使用请求中的参数绑定缓存基于 Django 类的视图响应

Question

我想缓存我的 API 响应并将其与 GET 请求中存在的参数绑定。 请求如下所示：

GET /products?producent=some_company

这是我的简化课程：

class ProductsListCreate(generics.ListCreateAPIView):

    def list(self, request, *args, **kwargs):
        producent = request.query_params.get("producent")
        cached_response = cache.get(f"response-products-{producent}", None)
        if not cached_response:
            queryset = self.filter_queryset(self.get_queryset())
            page = self.paginate_queryset(queryset)
            serializer = self.get_serializer(page)
            response = self.get_paginated_response(serializer.data)

            cache.set(f"response-products-{producent}", response , timeout=20)
            return response

        return cached_response 

但是当我尝试缓存响应时，我收到错误：

django.template.response.ContentNotRenderedError: The response content must be rendered before it can be pickled.

你有什么建议给我吗？当我试图弄清楚这一点时，我在这里搜索https://docs.djangoproject.com/en/3.2/topics/cache/。起初我尝试了使用

@cache_page

的方法，但它不允许我使用请求中的参数，所以我想要走的路是低级缓存 API。

Answer 1

错误消息实际上非常具有描述性。在尝试将其 pickle 为可缓存格式之前，您必须对响应调用

render()

。问题是：这将导致您陷入解构响应并从缓存数据重建的困境。 IMO，你最好缓存查询集。尽管如此，这就是它的样子：

class ProductsListCreate(generics.ListCreateAPIView):

    def list(self, request, *args, **kwargs):
        producent = request.query_params.get("producent")
        response_triple = cache.get(f"response-products-{producent}", None)
        if not response_triple:
            queryset = self.filter_queryset(self.get_queryset())
            page = self.paginate_queryset(queryset)
            serializer = self.get_serializer(page)
            response = self.get_paginated_response(serializer.data)

            response.render()

            if not response.status_code >= 400:
                # django 3.0 has no .items() method, django 3.2 has no ._headers
                if hasattr(response, '_headers'):
                    headers = response._headers.copy()
                else:
                    headers = {k: (k, v) for k, v in response.items()}
                response_triple = (
                    response.rendered_content,
                    response.status_code,
                    headers
                )
                cache.set(f"response-products-{producent}", response_triple, timeout=20)
        else:
            # build smaller Django HttpResponse
            content, status, headers = response_triple
            response = HttpResponse(content=content, status=status)
            for k, v in headers.values():
                response[k] = v
        if not hasattr(response, '_closable_objects'):
            response._closable_objects = []

        return response

取自：https://github.com/chibisov/drf-extensions/blob/ecdf3a95d7f18ccf9cffa55809635c3715179605/rest_framework_extensions/cache/decorators.py#L63

Answer 2

您可以在基于类的视图上使用此自定义函数：

def generate_cache_key(key_prefix, request):
    """ Generate cache key """
    query_params = request.GET.dict()
    query_string = urlencode(sorted(query_params.items()))
    base_key = f"{key_prefix}:{request.path}:{query_string}"
    return hashlib.md5(base_key.encode('utf-8')).hexdigest()

def cache_page_with_query_params(timeout, key_prefix=''):
  def decorator(view_func):
    @wraps(view_func)
    def _wrapped_view(request, *args, **kwargs):
        # Generate the custom cache key
        cache_key = generate_cache_key(key_prefix, request)

        # Try to get the cached response
        cached_response = cache.get(cache_key)
        if cached_response:
            return JsonResponse(cached_response)

        # Call the view function to get the response
        response = view_func(request, *args, **kwargs)
        data = response.data

        # Cache the response
        if response.status_code == 200:
            cache.set(cache_key, data, timeout)
        return JsonResponse(data)
    return _wrapped_view
return decorator

现在您可以像这样使用它：

@method_decorator(cache_page_with_query_params(60 * 15, key_prefix="add_your_prefix"))
def dispatch(self, request, *args, **kwargs):
    return super().dispatch(request, *args, **kwargs)