ImageDataSpliter 0.1.4

This utility splits an image dataset into train, validation, and test subsets while preserving the class folder structure.

DatasetSpliter

A simple Python utility for splitting an image dataset into train, validation, and test subsets while preserving class folder structure.

Overview

DatasetSpliter provides a function that takes a list of class folders (each containing image files) and splits the dataset into training, validation, and testing subsets. The images are copied into a new output directory while keeping the original class names.

Resulting directory structure

output_folder/
├── train/
│   └── class_name/
├── val/
│   └── class_name/
└── test/
    └── class_name/

The split process is random but reproducible using a fixed random seed.

Features

• Supports common image formats: .jpg, .jpeg, .png, .bmp, .gif, .tiff, .webp
• Preserves class directory names
• Prints a summary for each class
• Uses only the Python standard library

Installation

pip install ImageDataSpliter

Usage

from imagedataspliter import split_dataset

source_folders = [
    "data/cats",
    "data/dogs",
    "data/birds"
]

output_folder = "dataset_split"

split_dataset(
    source_folders=source_folders,
    output_folder=output_folder,
    train_ratio=0.7,
    val_ratio=0.15,
    test_ratio=0.15,
    seed=42
)

Function Signature

split_dataset(
    source_folders: list[str],
    output_folder: str,
    train_ratio: float = 0.7,
    val_ratio: float = 0.15,
    test_ratio: float = 0.15,
    seed: int = 42
)

Parameters

• source_folders List of paths to class directories containing images.

• output_folder Path to the directory where the split dataset will be saved.

• train_ratio Fraction of images used for training.

• val_ratio Fraction of images used for validation.

• test_ratio Fraction of images used for testing.

• seed Random seed for reproducible splits.

The sum of train_ratio, val_ratio, and test_ratio must equal 1.0.

Output

• Image files are copied, not moved

• Original data remains unchanged

• Class folder structure is preserved

• Progress information is printed to the console

Example Output

Processing class: cats
  Total images: 100
  Train: 70, Val: 15, Test: 15

Processing class: dogs
  Total images: 120
  Train: 84, Val: 18, Test: 18

Dataset saved to: dataset_split

Requirements

Python 3.8 or newer

No external dependencies

Contributing

Pull requests and issues are welcome. Feel free to suggest improvements or report bugs.