NumPy: Shape, ndim and size

Every array is described by its shape, its number of dimensions and its size. Work them out from the data.

What to build

A program that describes an array the way NumPy's attributes would.

Requirements

  • The first line is 1d or 2d. For 1d, one line of numbers follows; for 2d, a line with the number of rows r follows, then r lines of numbers.
  • Print ndim: D, shape: S and size: N, with the shape written as NumPy writes it: (5,) for 1-D and (2, 3) for 2-D.
  • If the rows of a 2-D array have different lengths, print only ValueError: inhomogeneous shape.

Input

As described.

Output

Three lines, or the error.

Constraints

1 ≤ r ≤ 20; each row has 1 to 20 numbers.

Example

Input:

1d
4 8 15 16 23

Output:

ndim: 1
shape: (5,)
size: 5

Why it works

An array needs every row the same length, because its shape describes one regular block; the size is the product of the shape.

Examples

Example 1 · Input
1d
4 8 15 16 23
Expected output
ndim: 1
shape: (5,)
size: 5
Example 2 · Input
2d
2
1 2 3
4 5 6
Expected output
ndim: 2
shape: (2, 3)
size: 6

Prompts inside input() are ignored when checking, so input("n: ") is fine.

Hints

Hint 1

All rows must have the same length; otherwise NumPy cannot make a regular array.

Hint 2

A 1-D shape is written with a trailing comma: (5,).

Solution

Show one correct solution (try it yourself first)
kind = input()
if kind == "1d":
    values = input().split()
    print("ndim: 1")
    print(f"shape: ({len(values)},)")
    print(f"size: {len(values)}")
else:
    r = int(input())
    rows = [input().split() for _ in range(r)]
    lengths = {len(row) for row in rows}
    if len(lengths) != 1:
        print("ValueError: inhomogeneous shape")
    else:
        c = len(rows[0])
        print("ndim: 2")
        print(f"shape: ({r}, {c})")
        print(f"size: {r * c}")

Lesson: Why NumPy →