Skip to content

v0.4.0 — Portfolio Risk and Exact Stress

Latest
Latest

Choose a tag to compare

View all tags
@github-actions github-actions released this 16 Jul 00:15
v0.4.0
2ad88e9
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

v0.4.0 — Portfolio Risk and Exact Stress

v0.4.0 makes verified portfolio risk and exact user-supplied stress the project's first-class production story while preserving the v0.3.7 cross-platform packaging baseline and the broader analytic, Monte Carlo, QMC, PDE, Heston, SSVI, and numerical evidence.

New portfolio surfaces

  • bs_portfolio_risk returns position price, value, and Greeks plus quantity-weighted portfolio value, delta, gamma, vega, theta, and rho totals in one native batch.
  • bs_portfolio_scenarios exact-reprices spot, volatility, rate, dividend, and elapsed-time shocks, with compact aggregate-only output by default.

Artifact-bound proof

  • Independent QuantLib comparison: worst price error 3.91e-14, worst Greek error 3.40e-12, position scenario P&L error 2.06e-13, and portfolio scenario P&L error 2.66e-13.
  • Determinism: exact zero-shock identity and 32/32 concurrent replays bitwise identical.
  • Recorded Apple M3 Pro evaluator: 20.18x risk-batch speedup / 20.25M positions per second; 27.92x aggregate-scenario speedup / 32.13M cells per second.
  • Verification covers 60 mixed position cases, 72 scenario cells, native and installed-wheel tests, ASan/UBSan, package/data policy, and exact artifact hashes.

Release payload

  • 20 provider-built CPython 3.8–3.12 wheels across manylinux x86_64, macOS universal2, Windows x86, and Windows x86_64. Installed-wheel contracts passed wherever the runner could execute the slice; cibuildwheel's documented limitation prevented executing the CPython 3.8 arm64 slice of the repaired universal2 wheel, while its x86_64 slice passed.
  • One verified source distribution, a machine-readable release manifest with SHA-256 digests, the committed artifact manifest, and the validation payload.
  • PyPI/TestPyPI publication is not claimed by this GitHub release.

Honest boundary

This release is deterministic Black–Scholes European portfolio valuation and user-supplied stress infrastructure. It is not trading alpha, a forecast, probabilistic market-risk validation, volatility-surface dynamics, a hedge/return result, or live P&L.

Start with the README, the runnable portfolio_risk.py example, the product and evidence hub, and the full release notes.

Assets 26
Loading