这就算是第一篇博客了，我老大陈如杰（一个牛逼的程序员）讲过，只有自己研究了的东西才能写到博客上，我谨遵教诲。
环境：Django==1.8.2、django-tastypie==0.12.2-dev、redis==2.10.3、django-redis==4.1.0
django配置了使用redis作为后台缓存
在ubuntu中安装了redis-server（sudo apt-get install redis-server），在django的settings中加入如下配置:（具体可以参考http://niwinz.github.io/django-redis/latest/）
# to set django's default cache add by minkedong
CACHES = {
"default": {
"BACKEND": "django_redis.cache.RedisCache",
"LOCATION": "redis://127.0.0.1:6379/1",
"OPTIONS": {
"CLIENT_CLASS": "django_redis.client.DefaultClient",
}
}
}
在写tastypie自己的ModelResource时，在Meta中加入：
cache = SimpleCache(timeout=8*60*60)
当我请求此ModelResource的时候，发出GET请求“/api/v1/author/1/”后，在ubuntu的bash中使用命令：redis-cli -n 1 进入redis，在redis中键入命令：keys *，查看到已经缓存了此API的请求结果。
但当我发出GET请求"/api/v1/author/"后，进入redis查看，发现并没有缓存我的结果。
果断进入tastypie的源码查看，在"tastypie.resources.ModelResource"中有如下两个方法：
def get_list(self, request, **kwargs):
"""
Returns a serialized list of resources.
Calls ``obj_get_list`` to provide the data, then handles that result
set and serializes it.
Should return a HttpResponse (200 OK).
"""
# TODO: Uncached for now. Invalidation that works for everyone may be
# impossible.
base_bundle = self.build_bundle(request=request)
objects = self.obj_get_list(bundle=base_bundle, **self.remove_api_resource_names(kwargs))
sorted_objects = self.apply_sorting(objects, options=request.GET)
paginator = self._meta.paginator_class(request.GET, sorted_objects, resource_uri=self.get_resource_uri(), limit=self._meta.limit, max_limit=self._meta.max_limit, collection_name=self._meta.collection_name)
to_be_serialized = paginator.page()
# Dehydrate the bundles in preparation for serialization.
bundles = []
for obj in to_be_serialized[self._meta.collection_name]:
bundle = self.build_bundle(obj=obj, request=request)
bundles.append(self.full_dehydrate(bundle, for_list=True))
to_be_serialized[self._meta.collection_name] = bundles
to_be_serialized = self.alter_list_data_to_serialize(request, to_be_serialized)
return self.create_response(request, to_be_serialized)
def get_detail(self, request, **kwargs):
"""
Returns a single serialized resource.
Calls ``cached_obj_get/obj_get`` to provide the data, then handles that result
set and serializes it.
Should return a HttpResponse (200 OK).
"""
basic_bundle = self.build_bundle(request=request)
try:
obj = self.cached_obj_get(bundle=basic_bundle, **self.remove_api_resource_names(kwargs))
except ObjectDoesNotExist:
return http.HttpNotFound()
except MultipleObjectsReturned:
return http.HttpMultipleChoices("More than one resource is found at this URI.")
bundle = self.build_bundle(obj=obj, request=request)
bundle = self.full_dehydrate(bundle)
bundle = self.alter_detail_data_to_serialize(request, bundle)
return self.create_response(request, bundle)
使用pdb打断点，发现“/api/v1/author/1/”请求调用get_detail方法，而"/api/v1/author/"请求调用get_list方法，从源码中不难看出，get_detail中调用了cached_obj_get方法，其从cache中拿数据，而get_list调用了obj_get_list方法（源码中有cached_obj_get_list方法，其并未调用），即未从cache中拿数据（其实get_list方法的注释中已经给出了解释
# TODO: Uncached for now. Invalidation that works for everyone may be impossible.），好吧，那解决方式就比较简单了，直接在自己的ModelResource中覆盖掉此方法，更改成：
def get_list(self, request, **kwargs): """ Returns a serialized list of resources. Calls ``obj_get_list`` to provide the data, then handles that result set and serializes it. Should return a HttpResponse (200 OK). """ # TODO: Uncached for now. Invalidation that works for everyone may be # impossible. base_bundle = self.build_bundle(request=request) objects = self.cached_obj_get_list(bundle=base_bundle, **self.remove_api_resource_names(kwargs)) sorted_objects = self.apply_sorting(objects, options=request.GET) paginator = self._meta.paginator_class(request.GET, sorted_objects, resource_uri=self.get_resource_uri(), limit=self._meta.limit, max_limit=self._meta.max_limit, collection_name=self._meta.collection_name) to_be_serialized = paginator.page() # Dehydrate the bundles in preparation for serialization. bundles = [] for obj in to_be_serialized[self._meta.collection_name]: bundle = self.build_bundle(obj=obj, request=request) bundles.append(self.full_dehydrate(bundle, for_list=True)) to_be_serialized[self._meta.collection_name] = bundles to_be_serialized = self.alter_list_data_to_serialize(request, to_be_serialized) return self.create_response(request, to_be_serialized)
再次请求"/api/v1/author/"后，发现redis中已经缓存了此结果！
但是，这里面后面的实际请求会有一个问题，那就是当你请求“/api/v1/author/”和请求“/api/v1/author/?format=json&name=jack”（假设有name筛选参数），其返回的json数据都是全量的数据，为什么呢？因为他在redis中生成的cache key都是一样的，我们要针对不同的请求参数生成不同的cache key，相当于不同的请求参数也cache一份数据。我的做法如下，直接在自己的ModelResource中覆盖掉cached_obj_get_list方法，更改成：
def cached_obj_get_list(self, bundle, **kwargs):
"""
更改get_list的cache key,使其对不同参数存不同的cache key
"""
if hasattr(bundle.request, 'GET'):
# Grab a mutable copy.
filters = bundle.request.GET.copy()
filters.update(kwargs)
applicable_filters = self.build_filters(filters=filters)
applicable_str = ':'.join(sorted(["%s=%s" % (key, ','.join(value) if type(value) == type([]) else value, ) for key, value in applicable_filters.items()]))
cache_key = self.generate_cache_key('list', applicable_str, **kwargs)
obj_list = self._meta.cache.get(cache_key)
if obj_list is None:
obj_list = self.obj_get_list(bundle=bundle, **kwargs)
self._meta.cache.set(cache_key, obj_list)
return obj_list
因为对redis的机制不太了解，后面自己将研究一下redis的东西，有什么错误，欢迎大家指正！