Array protocol and __array__ in NumPy - Mini Project: Build & Apply

Using the Array Protocol and __array__ Method in NumPy

📖 Scenario: Imagine you have a custom class that stores data like a list, but you want to use NumPy functions on it easily. NumPy can convert objects to arrays if they follow the array protocol by implementing the __array__ method.

🎯 Goal: You will create a simple custom class with a __array__ method so that NumPy can convert it to an array. Then, you will use NumPy functions on your custom object.

📋 What You'll Learn

Create a custom class called MyData that stores a list of numbers.

Add a __array__ method to MyData that returns a NumPy array of the stored data.

Create an instance of MyData with specific numbers.

Use np.array() to convert the MyData instance to a NumPy array.

Print the resulting NumPy array.

💡 Why This Matters

🌍 Real World

Custom data classes can store data in special ways but still work with NumPy functions by implementing the array protocol.

💼 Career

Understanding the array protocol helps you integrate your own data structures with NumPy, a key skill in data science and scientific computing.

1

Create the MyData class with stored data

Create a class called MyData with an __init__ method that takes a list called data and stores it in self.data. Then create an instance called my_data with the list [10, 20, 30].

NumPy

# Your code here

Need a hint?

Define a class with __init__ to save the list in self.data. Then create an object my_data with the list [10, 20, 30].

2

Add the __array__ method to MyData

Inside the MyData class, add a method called __array__ that takes self, dtype=None and returns np.array(self.data, dtype=dtype). Make sure to import NumPy as np at the top.

NumPy

import numpy as np

class MyData:
    def __init__(self, data):
        self.data = data
    # Your code here

my_data = MyData([10, 20, 30])

Need a hint?

The __array__ method should take self, dtype=None and return np.array(self.data, dtype=dtype).

3

Convert my_data to a NumPy array using np.array()

Create a variable called array_version and set it to np.array(my_data) to convert the MyData instance to a NumPy array.

NumPy

import numpy as np

class MyData:
    def __init__(self, data):
        self.data = data
    def __array__(self, dtype=None):
        return np.array(self.data, dtype=dtype)

my_data = MyData([10, 20, 30])

# Your code here

Need a hint?

Use np.array(my_data) to get the NumPy array from your custom object.

4

Print the NumPy array array_version

Write a print statement to display the variable array_version.

NumPy

import numpy as np

class MyData:
    def __init__(self, data):
        self.data = data
    def __array__(self, dtype=None):
        return np.array(self.data, dtype=dtype)

my_data = MyData([10, 20, 30])

array_version = np.array(my_data)

# Your code here

Need a hint?

Use print(array_version) to show the NumPy array.