mst-package 0.1.0

No project description provided

martstelm_DE2v2.1.5

A package containing utility functions for matrix operations including transpose, 1D windowing, and 2D convolution.

Installation

You can install the package using pip:

pip install martstelm_DE2v2.1.5

Usage

Importing the Functions

from package.module import transpose2d, window1d, convolution2d

Functions

1. transpose2d

Transpose a 2D matrix.

Arguments:

• input_matrix (List[List[float]]): A 2D list representing the matrix to be transposed.

Returns:

• List[List[float]]: The transposed 2D list.

Example:

matrix = [
    [1, 2, 3],
    [4, 5, 6]
]

transposed_matrix = transpose2d(matrix)
print(transposed_matrix)
# Output: [[1, 4], [2, 5], [3, 6]]

2. window1d

Generate a 1D sliding window view of the input array.

Arguments:

• input_array (Union[List, np.ndarray]): The input array.
• size (int): The size of each window.
• shift (int, optional): The shift between consecutive windows. Defaults to 1.
• stride (int, optional): The stride within each window. Defaults to 1.

Returns:

• List[Union[List, np.ndarray]]: A list of windows.

Example:

array = [1, 2, 3, 4, 5]
windows = window1d(array, size=3, shift=1)
print(windows)
# Output: [[1, 2, 3], [2, 3, 4], [3, 4, 5]]

3. convolution2d

Perform a 2D convolution on the input matrix using the given kernel.

Arguments:

• input_matrix (np.ndarray): The input matrix.
• kernel (np.ndarray): The convolution kernel.
• stride (int, optional): The stride of the convolution. Defaults to 1.

Returns:

• np.ndarray: The result of the convolution.

Example:

import numpy as np

input_matrix = np.array([
    [1, 2, 3],
    [4, 5, 6],
    [7, 8, 9]
])

kernel = np.array([
    [1, 0],
    [0, -1]
])

convolution_result = convolution2d(input_matrix, kernel)
print(convolution_result)
# Output: [[ -4.  -4.]
#          [-4. -4.]]