eimantas_data_transformations 0.1.1

A Python library for data transformations

Data Transformations Library

This is a Python library for performing common data transformations used in machine learning and data science workflows. The library includes functions for transposing matrices, creating time series windows, and performing 2D cross-correlation.

Features

• Transpose Function: Transposes a 2D matrix (list of lists).
• Time Series Windowing Function: Creates windows from a 1D array with specified size, shift, and stride.
• Cross-Correlation Function: Performs 2D cross-correlation on an input matrix using a given kernel.

Installation

To install the library, use pip3:

pip3 install eimantas_data_transformations

Or, if you are using Poetry, add it to your project with:

poetry add eimantas_data_transformations

Usage

Transpose Function

Transposes a 2D matrix (list of lists). This function takes a 2D matrix represented as a list of lists and returns a new matrix that is the transpose of the input matrix. Transposing a matrix means switching the rows and columns.

Example:

from eimantas_data_transformations.transformations import transpose2d
matrix = [[1, 2, 3], [4, 5, 6]]
transposed = transpose2d(matrix)
print(transposed)
# Output: [[1, 4], [2, 5], [3, 6]]

Time Series Windowing Function

Creates windows from a 1D array with specified size, shift, and stride.

Example:

from eimantas_data_transformations.transformations import window1d
input_array = [1, 2, 3, 4, 5, 6, 7, 8, 9]
windows = window1d(input_array, size=3, shift=2, stride=2)
print(windows)
# Output: [[1, 3, 5], [3, 5, 7], [5, 7, 9]]

Cross-Correlation Function

Performs 2D cross-correlation on an input matrix using a given kernel.

Example:

import numpy as np
from eimantas_data_transformations.transformations import convolution2d

input_matrix = np.array([
    [1, 2, 3, 4],
    [5, 6, 7, 8],
    [9, 10, 11, 12],
    [13, 14, 15, 16]
])
kernel = np.array([
    [1, 0],
    [0, 1]
])
result = convolution2d(input_matrix, kernel, stride=2)
print(result)
# Output: [[ 7. 11.]
#          [23. 27.]]

Author

Eimantas Venslovas