Traditional Class Definition

To illustrate, let's consider a simplified example related to currency trading, focusing on five key attributes: id, symbol, price, is_success, and addrs.

class CoinTrans:

def __init__(self, id: str, symbol: str, price: float, is_success: bool, addrs: list) -> None:

self.id = id

self.symbol = symbol

self.price = price

self.addrs = addrs

self.is_success = is_success

In the conventional approach, attributes are initialized within the __init__ method. To create an instance and display it, you would do the following:

if __name__ == "__main__":

coin_trans = CoinTrans("id01", "BTC/USDT", 71000, True, ["0x1111", "0x2222"])

print(coin_trans)

However, this code only prints the object's memory address, not its attribute values. To achieve a more readable output, we need to implement the __str__ method:

def __str__(self) -> str:

return f"Transaction Information: {self.id}, {self.symbol}, {self.price}, {self.addrs}, {self.is_success}"

After running the code again, the output will now include the transaction details.

Using the Dataclass Decorator

Now, let's see how the same class can be defined more succinctly using the @dataclass decorator:

from dataclasses import dataclass

@dataclass class CoinTrans:

id: str

symbol: str

price: float

is_success: bool

addrs: list

Executing the following code:

if __name__ == "__main__":

coin_trans = CoinTrans("id01", "BTC/USDT", 71000, True, ["0x1111", "0x2222"])

print(coin_trans)

Now produces a clear output of the object attributes, eliminating the need for custom __init__ and __str__ methods.

Default Values with Dataclasses

Utilizing the dataclass decorator allows for easy assignment of default values directly in the class definition:

@dataclass class CoinTrans:

id: str = "id01"

symbol: str = "BTC/USDT"

price: float = 71000.8

is_success: bool = True

addrs: list[str] = field(default_factory=list)

However, using mutable types like lists directly as default values will lead to errors. Instead, you can define a factory function to generate default values:

def gen_list():

return ["0x1111", "0x2222"]

@dataclass class CoinTrans:

id: str = "id01"

symbol: str = "BTC/USDT"

price: float = 71000.8

is_success: bool = True

addrs: list[str] = field(default_factory=gen_list)

Now, running the code will work seamlessly.

Hiding Sensitive Information

To prevent sensitive information from being printed, such as is_success and addrs, you can set repr=False:

@dataclass class CoinTrans:

id: str = "id01"

symbol: str = "BTC/USDT"

price: float = 71000.8

is_success: bool = field(default=True, repr=False)

addrs: list[str] = field(default_factory=gen_list, repr=False)

Running the code again will display only the non-sensitive attributes.

Making Objects Immutable

To ensure data integrity, you can set a dataclass as read-only using the frozen attribute:

@dataclass(frozen=True) class CoinTrans:

id: str = "id01"

#... other attributes ...

Attempting to modify an attribute will raise a FrozenInstanceError, safeguarding the data from inadvertent changes.

Converting to Tuples and Dictionaries

The dataclasses module also provides convenient methods to convert data classes into tuples and dictionaries, which is useful for interoperability with other systems:

from dataclasses import astuple, asdict

if __name__ == "__main__":

coin_trans = CoinTrans()

print(astuple(coin_trans))

print(asdict(coin_trans))

This results in outputs that can easily be used in various applications.