Installation¶

Basic Install¶

pip install -e .

This installs the optimizer library with all runtime dependencies: numpy, pandas, scipy, scikit-learn, skfolio, hmmlearn, and arch.

Development Install¶

pip install -e ".[dev]"

This includes test, lint, typecheck, and docs dependencies — everything needed for development and CI.

Optional Dependencies¶

Group	Install Command	Includes
`test`	`pip install -e ".[test]"`	pytest, pytest-cov, hypothesis
`lint`	`pip install -e ".[lint]"`	ruff, pip-audit
`typecheck`	`pip install -e ".[typecheck]"`	mypy
`docs`	`pip install -e ".[docs]"`	mkdocs-material, mkdocstrings
`dmm`	`pip install -e ".[dmm]"`	torch, pyro-ppl
`dev`	`pip install -e ".[dev]"`	All of the above + pre-commit

Requirements¶

Python >= 3.10
numpy, pandas, scipy, scikit-learn, skfolio, hmmlearn, arch

Verifying the Installation¶

After installing, verify the library loads correctly:

import optimizer
from optimizer.optimization import MeanRiskConfig, build_mean_risk
from optimizer.pipeline import run_full_pipeline

# Quick sanity check
config = MeanRiskConfig.for_max_sharpe()
optimizer = build_mean_risk(config)
print(f"Optimizer ready: {type(optimizer).__name__}")

Deep Markov Model (Optional)¶

The DMM module (optimizer.moments._dmm) requires PyTorch and Pyro, which are not declared in the standard dependencies due to their size. Install them separately:

pip install -e ".[dmm]"

The DMM module is imported conditionally — if torch/pyro are not installed, the rest of the library works normally and DMM-related imports are silently skipped.

Running Tests¶

# All tests
pytest tests/ -v

# Single module
pytest tests/rebalancing/ -v

# Single test
pytest -k "test_name"

Linting and Type Checking¶

# Lint
ruff check optimizer/ tests/

# Lint + auto-fix
ruff check . --fix

# Type check
mypy optimizer/