pyfolioanalytics 0.1.0

Add your description here

PyFolioAnalytics

Python implementation of the R package PortfolioAnalytics.

Features

• Portfolio Specification: Support for Box, Group, Turnover, Transaction Costs, and Position Limit constraints.
• Optimization Engines:
  • CVXPY: Linear, Quadratic (MVO), and Mixed-Integer programming.
  • SciPy (SLSQP): Non-linear optimization for Equal Risk Contribution (ERC).
  • Differential Evolution: Global heuristic search for non-convex problems.
• Risk Modeling:
  • Gaussian and Modified (Cornish-Fisher) VaR and ES.
  • Path-dependent measures: MaxDrawdown and AverageDrawdown.
• Statistical Models:
  • Black-Litterman posterior estimation.
  • Statistical Factor Models (PCA).
  • Meucci Entropy Pooling for view integration.
• Backtesting: Rolling-window and expanding-window rebalancing with flexible frequencies.
• Hierarchical Structures: Support for Regime Switching and Multi-layer portfolio architectures.

Installation

uv sync

Testing & Validation

This library has been rigorously cross-validated against the original R PortfolioAnalytics and PerformanceAnalytics libraries using:

1. EDHEC Dataset: Benchmark hedge fund index data.
2. Real Stock Data: AAPL, MSFT, GOOGL, AMZN, META (2020-2026).
3. Macro Asset Data: SPY, QQQ, GLD, TLT, BRK.B (2020-2026).

To run the parity tests:

uv run pytest

Structure

• src/pyfolioanalytics/: Core package source.
• data/: Sample datasets (EDHEC, Real Stock returns).
• tests/: Comprehensive test suite including multi-dataset cross-validation.
• third_party/PortfolioAnalytics/: Original R source for reference.