NumPy: A stable argsort

argsort tells you where each value of the sorted array came from. Compute np.argsort(a, kind="stable").

What to build

A program that prints the stable argsort of a list of numbers.

Requirements

  • Print the positions that sort the values in ascending order, keeping equal values in their original order.
  • Use NumPy's 1-D format: square brackets, numbers right-aligned to the widest one and separated by one space.

Input

One line of whole numbers.

Output

One line.

Constraints

1 to 1000 numbers; |x| ≤ 10⁹.

Example

Input:

30 10 20

Output:

[1 2 0]

Why it works

Sorting the positions by the values they point to gives exactly the order in which the values appear when sorted, and a stable sort keeps tied positions in increasing order.

Examples

Example 1 · Input
30 10 20
Expected output
[1 2 0]
Example 2 · Input
5 5 1
Expected output
[2 0 1]

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

Hints

Hint 1

sorted(range(n), key=lambda i: a[i]) gives the positions in sorted order; Python's sort is stable.

Hint 2

Print the positions in NumPy's 1-D format, right-aligned to the widest.

Solution

Show one correct solution (try it yourself first)
a = [int(x) for x in input().split()]
order = sorted(range(len(a)), key=lambda i: a[i])
w = max(len(str(i)) for i in order)
print("[" + " ".join(str(i).rjust(w) for i in order) + "]")

Lesson: Sorting, unique values and searching →