ftstore 0.1.0

Advanced local dataset management for machine learning

ftstore - Dataset Management for Machine Learning

ftstore is a Python package for managing and loading local datasets in machine learning projects, inspired by scikit-learn's dataset API.

Features

• 🗂️ Organize datasets in a structured directory
• ⚡️ Fast loading of CSV, Parquet, Feather, HDF5 and other formats
• 🔄 Automatic dataset caching for faster reloads
• 🌐 Auto-download datasets from remote sources
• 📊 Metadata management with JSON files

Installation

pip install ftstore

For full format support:

pip install ftstore[full]

Quick Start

from ftstore import load_data

# Load a dataset
iris = load_data("iris")

# Access features and target
print("Features:", iris.feature_names)
print("Target:", iris.target_name)
print("Data shape:", iris.data.shape)

# Load as DataFrame
df = load_data("iris", as_frame=True)

# Load as NumPy arrays
X, y = load_data("breast_cancer", return_X_y=True)

Documentation

See the Getting Started Guide for detailed usage instructions.