Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

added support for inverse of 3x3 matrix #7355

Merged
merged 6 commits into from Oct 26, 2022
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
137 changes: 122 additions & 15 deletions matrix/inverse_of_matrix.py
Expand Up @@ -2,47 +2,154 @@

from decimal import Decimal

from numpy import array


def inverse_of_matrix(matrix: list[list[float]]) -> list[list[float]]:
"""
A matrix multiplied with its inverse gives the identity matrix.
This function finds the inverse of a 2x2 matrix.
This function finds the inverse of a 2x2 and 3x3 matrix.
If the determinant of a matrix is 0, its inverse does not exist.

Sources for fixing inaccurate float arithmetic:
https://stackoverflow.com/questions/6563058/how-do-i-use-accurate-float-arithmetic-in-python
https://docs.python.org/3/library/decimal.html

Doctests for 2x2
>>> inverse_of_matrix([[2, 5], [2, 0]])
[[0.0, 0.5], [0.2, -0.2]]
>>> inverse_of_matrix([[2.5, 5], [1, 2]])
Traceback (most recent call last):
...
...
ValueError: This matrix has no inverse.
>>> inverse_of_matrix([[12, -16], [-9, 0]])
[[0.0, -0.1111111111111111], [-0.0625, -0.08333333333333333]]
>>> inverse_of_matrix([[12, 3], [16, 8]])
[[0.16666666666666666, -0.0625], [-0.3333333333333333, 0.25]]
>>> inverse_of_matrix([[10, 5], [3, 2.5]])
[[0.25, -0.5], [-0.3, 1.0]]

Doctests for 3x3
>>> inverse_of_matrix([[2, 5, 7], [2, 0, 1], [1, 2, 3]])
[[2.0, 5.0, -4.0], [1.0, 1.0, -1.0], [-5.0, -12.0, 10.0]]
>>> inverse_of_matrix([[1, 2, 2], [1, 2, 2], [3, 2, -1]])
Traceback (most recent call last):
...
ValueError: This matrix has no inverse.

>>> inverse_of_matrix([[],[]])
Traceback (most recent call last):
...
ValueError: Please provide a matrix of size 2x2 or 3x3.

>>> inverse_of_matrix([[1, 2], [3, 4], [5, 6]])
Traceback (most recent call last):
...
ValueError: Please provide a matrix of size 2x2 or 3x3.

>>> inverse_of_matrix([[1, 2, 1], [0,3, 4]])
Traceback (most recent call last):
...
ValueError: Please provide a matrix of size 2x2 or 3x3.

>>> inverse_of_matrix([[1, 2, 3], [7, 8, 9], [7, 8, 9]])
Traceback (most recent call last):
...
ValueError: This matrix has no inverse.

>>> inverse_of_matrix([[1, 0, 0], [0, 1, 0], [0, 0, 1]])
[[1.0, 0.0, 0.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0]]
"""

d = Decimal # An abbreviation for conciseness
d = Decimal

# Check if the provided matrix has 2 rows and 2 columns
# since this implementation only works for 2x2 matrices
if len(matrix) != 2 or len(matrix[0]) != 2 or len(matrix[1]) != 2:
raise ValueError("Please provide a matrix of size 2x2.")
if len(matrix) == 2 and len(matrix[0]) == 2 and len(matrix[1]) == 2:
# Calculate the determinant of the matrix
determinant = float(
d(matrix[0][0]) * d(matrix[1][1]) - d(matrix[1][0]) * d(matrix[0][1])
)
if determinant == 0:
raise ValueError("This matrix has no inverse.")

# Creates a copy of the matrix with swapped positions of the elements
swapped_matrix = [[0.0, 0.0], [0.0, 0.0]]
swapped_matrix[0][0], swapped_matrix[1][1] = matrix[1][1], matrix[0][0]
swapped_matrix[1][0], swapped_matrix[0][1] = -matrix[1][0], -matrix[0][1]

# Calculate the inverse of the matrix
return [
[(float(d(n)) / determinant) or 0.0 for n in row] for row in swapped_matrix
]
elif (
len(matrix) == 3
and len(matrix[0]) == 3
and len(matrix[1]) == 3
and len(matrix[2]) == 3
):
# Calculate the determinant of the matrix using Sarrus rule
determinant = float(
(
(d(matrix[0][0]) * d(matrix[1][1]) * d(matrix[2][2]))
+ (d(matrix[0][1]) * d(matrix[1][2]) * d(matrix[2][0]))
+ (d(matrix[0][2]) * d(matrix[1][0]) * d(matrix[2][1]))
)
- (
(d(matrix[0][2]) * d(matrix[1][1]) * d(matrix[2][0]))
+ (d(matrix[0][1]) * d(matrix[1][0]) * d(matrix[2][2]))
+ (d(matrix[0][0]) * d(matrix[1][2]) * d(matrix[2][1]))
)
)
if determinant == 0:
raise ValueError("This matrix has no inverse.")

# Creating cofactor matrix
cofactor_matrix = [
[d(0.0), d(0.0), d(0.0)],
[d(0.0), d(0.0), d(0.0)],
[d(0.0), d(0.0), d(0.0)],
]
cofactor_matrix[0][0] = (d(matrix[1][1]) * d(matrix[2][2])) - (
d(matrix[1][2]) * d(matrix[2][1])
)
cofactor_matrix[0][1] = -(
(d(matrix[1][0]) * d(matrix[2][2])) - (d(matrix[1][2]) * d(matrix[2][0]))
)
cofactor_matrix[0][2] = (d(matrix[1][0]) * d(matrix[2][1])) - (
d(matrix[1][1]) * d(matrix[2][0])
)
cofactor_matrix[1][0] = -(
(d(matrix[0][1]) * d(matrix[2][2])) - (d(matrix[0][2]) * d(matrix[2][1]))
)
cofactor_matrix[1][1] = (d(matrix[0][0]) * d(matrix[2][2])) - (
d(matrix[0][2]) * d(matrix[2][0])
)
cofactor_matrix[1][2] = -(
(d(matrix[0][0]) * d(matrix[2][1])) - (d(matrix[0][1]) * d(matrix[2][0]))
)
cofactor_matrix[2][0] = (d(matrix[0][1]) * d(matrix[1][2])) - (
d(matrix[0][2]) * d(matrix[1][1])
)
cofactor_matrix[2][1] = -(
(d(matrix[0][0]) * d(matrix[1][2])) - (d(matrix[0][2]) * d(matrix[1][0]))
)
cofactor_matrix[2][2] = (d(matrix[0][0]) * d(matrix[1][1])) - (
d(matrix[0][1]) * d(matrix[1][0])
)

# Calculate the determinant of the matrix
determinant = d(matrix[0][0]) * d(matrix[1][1]) - d(matrix[1][0]) * d(matrix[0][1])
if determinant == 0:
raise ValueError("This matrix has no inverse.")
# Transpose the cofactor matrix (Adjoint matrix)
adjoint_matrix = array(cofactor_matrix)
for i in range(3):
for j in range(3):
adjoint_matrix[i][j] = cofactor_matrix[j][i]

# Creates a copy of the matrix with swapped positions of the elements
swapped_matrix = [[0.0, 0.0], [0.0, 0.0]]
swapped_matrix[0][0], swapped_matrix[1][1] = matrix[1][1], matrix[0][0]
swapped_matrix[1][0], swapped_matrix[0][1] = -matrix[1][0], -matrix[0][1]
# Inverse of the matrix using the formula (1/determinant) * adjoint matrix
inverse_matrix = array(cofactor_matrix)
for i in range(3):
for j in range(3):
inverse_matrix[i][j] /= d(determinant)

# Calculate the inverse of the matrix
return [[float(d(n) / determinant) or 0.0 for n in row] for row in swapped_matrix]
# Calculate the inverse of the matrix
return [[float(d(n)) or 0.0 for n in row] for row in inverse_matrix]
raise ValueError("Please provide a matrix of size 2x2 or 3x3.")