让对象序列化自身

Question

使用 JSON 或 Pickle，我可以实例化一个对象并保存该对象，如下所示：

import pickle 
thing = Thingy(some_data) # takes a while...
# ... do stuff with thing then save it since it mostly is the same
with open(filename, 'wb') as f: 
    pickle.dump(thing, f)

如果序列化文件存在则读回：

if thingy_file.exists():
    with open(filename, 'rb') as f:
        thing=pickle.load(f)  # Thingy(some_data) restored
    thing.updateyourself()  # pretty fast
else: 
    thing=Thingy(some_data) # takes a while...

但是，此序列化是在 Thingy 对象的实例化外部完成的...

现在假设我有一些带有内部数据的自定义对象，需要花费很长时间来创建：

class Thingy:
   def __init__(self, stuff...)
      # this take a while the first time, but could be instant if read back
      # pseudo'ish:
        if self.file.exists():
            print(f'Existing Thingy {self.file} found...')
            # What I am trying to do here is read from a serialization file the
            # data comprising the last time THIS INSTANCE of Thingy existed
             with open(self.file, 'rb') as f:
                self=pickle.load(f)
        else:
           # bummer this is gonna take a while...
        
        # the rest of the object code...
        # HERE we either have calculated a Thingy or have read it back into itself
        # The instance may be modified -- now save its own self to be re-read in future:
     
    def save_my_own_instance(self):
       # now self is this instance of Thingy that will be serialized to disk
       # magic I do not know...
       with open(self.file, 'wb') as f:
            pickle.dump(my_own_instance_data, f, pickle.HIGHEST_PROTOCOL)

那么问题：对象使用序列化将自己的数据恢复到自己的实例中的方法是什么？将自身保存在自己的实例中怎么样？

---

什么是

Thingy

的一些细节：

from pathlib import Path

p=Path(a_HUGE_tree_of_data_files) # 750k plus large image files

tree_data=Thingy(p)
# initial run on p is hour+
# the tree changes very little, so an update of Thingy(p) 
# only takes seconds

# I don't want to educate the user to check for a previous run
# It should be automatic

Answer 1

解决此问题的一种方法是在

Thingy

中创建一个静态方法。

class Thingy:
    @staticmethod
    def load_instance_from_file_or_create_new(filename, stuff):
        if filename.exists():# check if file exists
            with open(filename, 'rb') as f:
                instance=pickle.load(f)
            return instance
        return Thingy(stuff)

    def __init__(self, stuff):
        # create the instance normally

    def __del__(self):
        # the instance is being deleted
        with open(self.file, 'wb') as f:
            pickle.dump(my_own_instance_data, f, pickle.HIGHEST_PROTOCOL)

可以用作

thing = Thingy.load_instance_from_file_or_create_new("xyz",stuff)

静态方法从文件加载

Thingy

实例（如果文件存在），否则使用提供的

stuff

创建一个新实例。
您必须将

filename

存储在

Thingy

之外，除非您有默认的

filename

。
每当实例被销毁时，都会自动调用

__del__

方法，因此您可以使用它将实例存储到文件中。