PyPortfolioOpt

1.6.0 verified Mon Apr 27 auth: no python

PyPortfolioOpt is a financial portfolio optimization library for Python, providing methods for mean-variance optimization, Black-Litterman allocation, and risk parity. Version 1.6.0 supports CVXPY-based solvers and offers both classical and objective-based optimization approaches. Release cadence is irregular, with contributions from the community.

pip install pyportfolioopt

Common errors

error pypfopt.exceptions.OptimizationError: No feasible solution found ↓

cause The optimization constraints (e.g., weight bounds, target return) are too tight, or the covariance matrix is singular.

fix

Relax bounds, check for duplicate assets, or use a non-singular risk model like Ledoit-Wolf shrinkage.

error ImportError: cannot import name 'EfficientFrontier' from 'pypfopt' ↓

cause Installed pyportfolioopt via pip but tried to import from 'pyportfolioopt' instead of 'pypfopt'.

fix

Change import to 'from pypfopt import EfficientFrontier'

error cvxpy.error.SolverError: Solver 'ECOS' failed. Try another solver. ↓

cause The default solver ECOS may not be installed or fails on some problems (e.g., integer constraints).

fix

Install an alternative solver like SCS (pip install scs) or CVXOPT (pip install cvxopt) and specify solver='SCS' or solver='CVXOPT'.

Warnings

gotcha Import package as pypfopt, not pyportfolioopt. The pip install name and the import name differ. ↓

fix Use 'from pypfopt import ...' instead of 'from pyportfolioopt import ...'

gotcha CVXPY may require a separate solver installation (e.g., 'pip install cvxpy') on some systems, especially for advanced solvers like ECOS or SCS. ↓

fix Install CVXPY with pip install cvxpy or use the bundled solver (may not work for large problems).

breaking In version 1.5+, the EfficientFrontier object no longer accepts a target return as a parameter to some methods; use objective functions instead. ↓

fix Use methods like `ef.efficient_return(target_return)` directly, not via a constructor parameter.

deprecated The `pypfopt.plotting` module is deprecated as of version 1.6.0 and will be removed in a future release. ↓

fix Use matplotlib directly to plot efficient frontier or weights.

Imports

EfficientFrontier

wrong

from pyportfolioopt import EfficientFrontier

correct

from pypfopt import EfficientFrontier

Actual package name is pypfopt, not pyportfolioopt

risk_models

wrong

from pyportfolioopt import risk_models

correct

from pypfopt import risk_models

risk_models is a submodule of pypfopt

expected_returns

wrong

from pyportfolioopt import expected_returns

correct

from pypfopt import expected_returns

expected_returns is a submodule of pypfopt

Quickstart

Basic mean-variance optimization using historical returns and sample covariance.

import pandas as pd
from pypfopt import EfficientFrontier, risk_models, expected_returns

# Sample data: prices of 3 assets
data = pd.DataFrame({
    'AAPL': [1.0, 0.9, 0.8, 0.85],
    'GOOG': [1.0, 1.1, 1.0, 1.05],
    'MSFT': [1.0, 0.95, 1.1, 1.0]
})

# Calculate expected returns and covariance matrix
mu = expected_returns.mean_historical_return(data)
Sigma = risk_models.sample_cov(data)

# Optimize for maximum Sharpe ratio
ef = EfficientFrontier(mu, Sigma)
weights = ef.max_sharpe()

print(ef.clean_weights())