Serialization

Serialization in Python

Serialization refers to the process of converting an object into a format that can be easily stored, transmitted, or reconstructed later. In Python, this involves converting complex data structures, such as objects or dictionaries, into a byte stream.

Why Do We Use Serialization?

Serialization allows data to be easily saved to disk or transmitted over a network, and later reconstructed back into its original form. It is important for tasks like saving game states, storing user preferences, or exchanging data between different systems.

Serialization Libraries in Python

Python offers several libraries for serialization, each with its own advantages. Here is a detailed overview of some commonly used serialization libraries in Python −

• Pickle − This is Python’s built-in module for serializing and deserializing Python objects. It is simple to use but specific to Python and may have security implications if used with untrusted data.
• JSON − JSON (JavaScript Object Notation) is a lightweight data interchange format that is human-readable and easy to parse. It is ideal for web APIs and cross-platform communication.
• YAML − YAML: YAML (YAML Ain’t Markup Language) is a human-readable data serialization standard that is also easy for both humans and machines to read and write. It supports complex data structures and is often used in configuration files.

Serialization Using Pickle Module

The pickle module in Python is used for serializing and deserializing objects. Serialization, also known as pickling, involves converting a Python object into a byte stream, which can then be stored in a file or transmitted over a network.

Deserialization, or unpickling, is the reverse process, converting the byte stream back into a Python object.

Serializing an Object

We can serialize an object using the dump() function and write it to a file. The file must be opened in binary write mode (‘wb’).

Example

In the following example, a dictionary is serialized and written to a file named “data.pkl” −

import pickle

data ={'name':'Alice','age':30,'city':'New York'}# Open a file in binary write modewithopen('data.pkl','wb')asfile:# Serialize the data and write it to the file
   pickle.dump(data,file)print("File created!!")

When above code is executed, the dictionary object’s byte representation will be stored in data.pkl file.

Deserializing an Object

To deserialize or unpickle the object, you can use the load() function. The file must be opened in binary read mode (‘rb’) as shown below −

import pickle

# Open the file in binary read modewithopen('data.pkl','rb')asfile:# Deserialize the data
   data = pickle.load(file)print(data)

This will read the byte stream from “data.pkl” and convert it back into the original dictionary as shown below −

{'name': 'Alice', 'age': 30, 'city': 'New York'}

Pickle Protocols

Protocols are the conventions used in constructing and deconstructing Python objects to/from binary data.

The pickle module supports different serialization protocols, with higher protocols generally offering more features and better performance. Currently pickle module defines 6 different protocols as listed below −

Sr.No.	Protocol & Description
1	Protocol version 0Original “human-readable” protocol backwards compatible with earlier versions.
2	Protocol version 1Old binary format also compatible with earlier versions of Python.
3	Protocol version 2Introduced in Python 2.3 provides efficient pickling of new-style classes.
4	Protocol version 3Added in Python 3.0. recommended when compatibility with other Python 3 versions is required.
5	Protocol version 4Introduced in Python 3.4. It adds support for very large objects.
6	Protocol version 5Introduced in Python 3.8. It adds support for out-of-band data.

You can specify the protocol by passing it as an argument to pickle.dump() function.

To know the highest and default protocol version of your Python installation, use the following constants defined in the pickle module −

>>>import pickle
>>> pickle.HIGHEST_PROTOCOL
5>>> pickle.DEFAULT_PROTOCOL
4

Pickler and Unpickler Classes

The pickle module in Python also defines Pickler and Unpickler classes for more detailed control over the serialization and deserialization processes. The “Pickler” class writes pickle data to a file, while the “Unpickler” class reads binary data from a file and reconstructs the original Python object.

Using the Pickler Class

To serialize a Python object using the Pickler class, you can follow these steps −

from pickle import Pickler

# Open a file in binary write modewithopen("data.txt","wb")as f:# Create a dictionary
   dct ={'name':'Ravi','age':23,'Gender':'M','marks':75}# Create a Pickler object and write the dictionary to the file
   Pickler(f).dump(dct)print("Success!!")

After executing the above code, the dictionary object’s byte representation will be stored in “data.txt” file.

Using the Unpickler Class

To deserialize the data from a binary file using the Unpickler class, you can do the following −

from pickle import Unpickler

# Open the file in binary read modewithopen("data.txt","rb")as f:# Create an Unpickler object and load the dictionary from the file
   dct = Unpickler(f).load()# Print the dictionaryprint(dct)

We get the output as follows −

{'name': 'Ravi', 'age': 23, 'Gender': 'M', 'marks': 75}

Pickling Custom Class Objects

The pickle module can also serialize and deserialize custom classes. The class definition must be available at both the time of pickling and unpickling.

Example

In this example, an instance of the “Person” class is serialized and then deserialized, maintaining the state of the object −

import pickle
classPerson:def__init__(self, name, age, city):
  self.name = name
  self.age = age
  self.city = city
# Create an instance of the Person class
person = Person('Alice',30,'New York')# Serialize the person objectwithopen('person.pkl','wb')asfile:
   pickle.dump(person,file)# Deserialize the person objectwithopen('person.pkl','rb')asfile:
   person = pickle.load(file)print(person.name, person.age, person.city)

After executing the above code, we get the following output −

Alice 30 New York

The Python standard library also includes the marshal module, which is used for internal serialization of Python objects. Unlike pickle, which is designed for general-purpose use, marshal is primarily intended for use by Python itself (e.g., for writing .pyc files).

It is generally not recommended for general-purpose serialization due to potential compatibility issues between Python versions.

Using JSON for Serialization

JSON (JavaScript Object Notation) is a popular format for data interchange. It is human-readable, easy to write, and language-independent, making it ideal for serialization.

Python provides built-in support for JSON through the json module, which allows you to serialize and deserialize data to and from JSON format.

Serialization

Serialization is the process of converting a Python object into a JSON string or writing it to a file.

Example: Serialize Data to a JSON String

In the example below, we use the json.dumps() function to convert a Python dictionary to a JSON string −

import json

# Create a dictionary
data ={"name":"Alice","age":25,"city":"San Francisco"}# Serialize the dictionary to a JSON string
json_string = json.dumps(data)print(json_string)

Following is the output of the above code −

{"name": "Alice", "age": 25, "city": "San Francisco"}

Example: Serialize Data and Write to a File

In here, we use the json.dump() function to write the serialized JSON data directly to a file −

import json

# Create a dictionary
data ={"name":"Alice","age":25,"city":"San Francisco"}# Serialize the dictionary and write it to a filewithopen("data.json","w")as f:
   json.dump(data, f)print("Success!!")

Deserialization

Deserialization is the process of converting a JSON string back into a Python object or reading it from a file.

Example: Deserialize a JSON String

In the following example, we use the json.loads() function to convert a JSON string back into a Python dictionary −

import json

# JSON string
json_string ='{"name": "Alice", "age": 25, "city": "San Francisco"}'# Deserialize the JSON string into a Python dictionary
loaded_data = json.loads(json_string)print(loaded_data)

It will produce the following output −

{'name': 'Alice', 'age': 25, 'city': 'San Francisco'}

Example: Deserialize Data from a File

Here, we use the json.load() function to read JSON data from a file and convert it to a Python dictionary−

import json

# Open the file and load the JSON data into a Python dictionarywithopen("data.json","r")as f:
   loaded_data = json.load(f)print(loaded_data)

The output obtained is as follows −

{'name': 'Alice', 'age': 25, 'city': 'San Francisco'}

Using YAML for Serialization

YAML (YAML Ain’t Markup Language) is a human-readable data serialization standard that is commonly used for configuration files and data interchange.

Python supports YAML serialization and deserialization through the pyyamlpackage, which needs to be installed first as shown below −

pip install pyyaml

Example: Serialize Data and Write to a YAML File

In the below example, yaml.dump() function converts the Python dictionary data into a YAML string and writes it to the file “data.yaml”.

The “default_flow_style” parameter ensures that the YAML output is more human-readable with expanded formatting −

import yaml

# Create a Python dictionary
data ={"name":"Emily","age":35,"city":"Seattle"}# Serialize the dictionary and write it to a YAML filewithopen("data.yaml","w")as f:
   yaml.dump(data, f, default_flow_style=False)print("Success!!")

Example: Deserialize Data from a YAML File

Here, yaml.safe_load() function is used to safely load the YAML data from “data.yaml” and convert it into a Python dictionary (loaded_data) −

Using safe_load() is preferred for security reasons as it only allows basic Python data types and avoids executing arbitrary code from YAML files.

import yaml

# Deserialize data from a YAML filewithopen("data.yaml","r")as f:
   loaded_data = yaml.safe_load(f)print(loaded_data)

The output produced is as shown below −

{'age': 35, 'city': 'Seattle', 'name': 'Emily'}