Understanding the python syntax

Question

Can somebody please help with the below code snippets?

1. (e, x), y = d
2. yhat = atom_count @ w + b
3. baseline_val_loss = [0.0 for _ in range(epochs)]

I am looking for verbal explanations of the above three lines as I never saw those syntaxes in shell scripting. The code I am learning is part of a book located at https://dmol.pub/applied/QM9.html

Answer 1

(1) (e, x), y = d

What you are doing here is essentially unpacking the values from the variable d into the tuple (e, x) and the variable y. This only works if d is in the same format of the left-hand side.

For instance, suppose d = ((3, 3), 4). Then:

d = ((3, 3), 4)
(e, x), y = d

print(e) # prints 3
print(x) # prints 3
print(y) # prints 4

Simpler examples would be:

a, b = 2, 3
print(a) # prints 2
print(b) # prints 3

(x, y, z) = (10, 20, 30)
print(x) # prints 10
print(y) # prints 20
print(z) # prints 30

The only difference is that in your example you are using a tuple, thus you need to unpack all values correctly.

(2) yhat = atom_count @ w + b

The @ operator is an operator used for matrix multiplication, added in Python 3.5.

Here you are essentially declaring a variable yhat that is equal to the matrix multiplication of the matrices atom_count and w+b. For instance:

import numpy as np

A = np.array([
    [1, 2],
    [3, 4]
])

B = np.array([
    [4, 5],
    [6, 7]
])

print(A @ B) # prints [[19 22]
             #         [43 50]]

(3) baseline_val_loss = [0.0 for _ in range(epochs)]

This is called list comprehension. It is a way to initialize lists instead of writing verbose for loops. Here you are initializing a list called baseline_val_loss with an amount of epochs zeros.

If you have epochs = 10, then you would have a python list with 10 zeros. The range function creates a list from 0 to epochs - 1. You can use _ as variable to iterate because you simply do not care about the current loop value; you won't do anything with it.