mlsplitter 1.0.0

Convenient helpers for splitting DataFrames into features/target and creating train/dev/test splits.

mlsplitter

A small, well-tested Python package that provides two conveniences on top of scikit-learn:

1. x_y_mlsplitter – split a DataFrame into feature matrix X and target vector y by column name or index.
2. train_test_mlsplitter – thin validated wrapper around sklearn.model_selection.train_test_split.
3. train_dev_test_mlsplitter – split data into three sets (train / dev / test) with sizes expressed as fractions of the full dataset.

Installation

pip install mlsplitter

Or from source:

git clone https://github.com/Fares-Ayman-1/mlsplitter.git
cd mlsplitter
pip install -e ".[dev]"

Quick start

import pandas as pd
from mlsplitter import x_y_splitter, train_test_splitter, train_dev_test_splitter

df = pd.read_csv("my_data.csv")

# Split features from target (by name or by position)
X, y = x_y_splitter(df, column_name="price")
X, y = x_y_splitter(df, column_index=-1)

# Train / test split
x_train, x_test, y_train, y_test = train_test_splitter(X, y, test_size=0.2)

# Train / dev / test split
x_train, x_dev, x_test, y_train, y_dev, y_test = train_dev_test_splitter(
    X, y, dev_size=0.1, test_size=0.2
)

Running tests

pytest --cov=mlsplitter

License

MIT