Installation

Quick Install

pip install fractime

This installs FracTime with all required dependencies.

---

From Source

For the latest development version:

git clone https://github.com/wayy-research/fracTime.git
cd fracTime
pip install -e .

For development with testing tools:

pip install -e ".[dev]"

---

Dependencies

FracTime automatically installs these dependencies:

Package	Purpose
numpy	Numerical computing
polars	Fast DataFrames
numba	JIT compilation for speed
wrchart	Interactive visualizations
scikit-learn	Clustering and preprocessing

Optional Dependencies

Package	Purpose	Install
pymc	Bayesian forecasting	pip install pymc

---

Verify Installation

import fractime as ft

# Check version
print(f"FracTime version: {ft.__version__}")

# Quick test
import numpy as np
prices = 100 * np.cumprod(1 + np.random.randn(100) * 0.02)
analyzer = ft.Analyzer(prices)
print(f"Hurst exponent: {analyzer.hurst}")

Expected output:

FracTime version: 0.3.0
Hurst exponent: hurst=0.5234

---

Platform Support

FracTime supports:

• Python: 3.10, 3.11, 3.12
• OS: Linux, macOS, Windows
• Architecture: x86_64, ARM64

---

Troubleshooting

Numba Compilation

On first use, Numba compiles optimized functions. This causes a one-time delay of a few seconds. Subsequent runs are fast.

Memory Issues

For very large datasets (>100,000 points), consider:

# Reduce bootstrap samples
analyzer = ft.Analyzer(prices, n_samples=500)

# Use smaller rolling windows
analyzer = ft.Analyzer(prices, window=30)

Import Errors

If you see import errors, ensure you have the latest version:

pip install --upgrade fractime