riskfolio-lib

name	riskfolio-lib
description	Skill for using the Riskfolio-Lib Python library for portfolio optimization and quantitative strategic asset allocation.
version	0.1
author	Hermes Agent
tags	["finance","portfolio","optimization","python"]
category	data-science

Overview

Riskfolio‑Lib is a Python package that provides advanced tools for portfolio construction, risk measurement, and optimization. It implements classical mean‑variance models, Black‑Litterman, risk‑parity, hierarchical risk parity, and many modern approaches.

Installation

pip install riskfolio-lib

Requires Python ≥3.7 and the libraries numpy, pandas, scipy, matplotlib.

Quick Start

import pandas as pd
import riskfolio as rp

# 1️⃣ Load price data (e.g., daily close prices)
prices = pd.read_csv('prices.csv', index_col='Date', parse_dates=True)

# 2️⃣ Compute returns
returns = prices.pct_change().dropna()

# 3️⃣ Create a Portfolio object
port = rp.Portfolio(returns)

# 4️⃣ Estimate the risk‑return moments
port.assets_stats(method_mu='historical', method_cov='ledoit-wolf')

# 5️⃣ Choose an objective – here Minimum Variance (MV) with a risk‑budget constraint
def objective(w):
    return port.portfolio_performance(w)

# 6️⃣ Run the optimizer (different solvers are available)
weights = port.optimization(model='MV', rm='MV', obj='MinRisk', l=0, hist=True)

print('Optimized weights:', weights)

# 7️⃣ Plot the efficient frontier (optional)
port.plot_frontier()

Core Concepts

• Portfolio – central class handling assets, moments, constraints, and solvers.
• Risk Measures (rm) – MV (variance), CVaR, MAD, SemiStd, EVaR, etc.
• Objective Functions (obj) – MinRisk, MaxSharpe, Utility, ERC (risk parity), HRP (hierarchical), etc.
• Constraints – you can set bounds, cardinality, sector exposure, turnover, etc. via port.set_bounds(), port.set_cardinality(), port.set_sector_constraints().
• Solvers – default SLSQP; alternatives include ECOS, CVXOPT, MOSEK (if installed).

Example Use Cases

1️⃣ Classical Mean‑Variance (Markowitz)

port = rp.Portfolio(returns)
port.assets_stats(method_mu='mean', method_cov='ledoit-wolf')
weights = port.optimization(model='MV', rm='MV', obj='Sharpe', l=0.5)

2️⃣ Risk Parity (ERC)

weights = port.optimization(model='ERC', rm='MV')

3️⃣ Hierarchical Risk Parity (HRP)

weights = port.optimization(model='HRP', rm='MV')

4️⃣ Black‑Litterman Expected Returns

port = rp.Portfolio(returns)
port.black_litterman(tau=0.025, P=None, Q=None, pi='market')
weights = port.optimization(model='BL', rm='MV', obj='Sharpe')

Common Functions

Function	Purpose
rp.Portfolio(returns)	Initialise with a DataFrame of asset returns
assets_stats()	Estimate mean, covariance, higher‑order moments
set_bounds(lower, upper)	Impose weight limits
set_constraints()	Add linear constraints (e.g., sector, turnover)
optimization(model, rm, obj, **kwargs)	Run the optimizer – choose model (MV, ERC, HRP, BL, …) and risk measure
plot_frontier()	Visualise efficient frontier
plot_risk_contributions()	Show each asset’s contribution to portfolio risk

References

• Official repository: https://github.com/dcajasn/Riskfolio-Lib
• Documentation: https://riskfolio-lib.readthedocs.io/
• Academic paper: "Riskfolio‑Lib: A Python library for portfolio optimization" (arXiv:2109.00943)
• Official repository: https://github.com/dcajasn/Riskfolio-Lib
• Documentation: https://riskfolio-lib.readthedocs.io/
• Academic paper: "Riskfolio‑Lib: A Python library for portfolio optimization" (arXiv:2109.00943)

Tip for LLM Integration

When a downstream LLM needs to call a Riskfolio function, reference this skill and request the specific snippet. Example prompt to the model:

Use the `riskfolio-lib` skill to construct a minimum‑variance portfolio for the assets `AAPL`, `MSFT`, `GOOG` based on the CSV file `prices.csv`.

The skill will provide the exact code block and explain any required parameters.