Skip to content

Release 0.6.0 — Product Guardrails & Research Loop

Date: 2026-06-28
Install: pip install quantwave or pip install "quantwave[polars]"

Headline

QuantWave 0.6.0 ships the v0.6 productization epic: CI guardrails, a single verify entrypoint, the full research loop (feature matrix + Monte Carlo), fractional differencing, HTML tear sheets, and documentation SOA completion.

What's new

PyPI packaging (unified wheel)

  • Single wheel bundles core (quantwave._quantwave), Polars plugins (quantwave_plugins), and backtest (quantwave._backtest)
  • pip install "quantwave[polars]" — Polars extra for batch .ta / .bt namespaces
  • quantwave CLI — list, info, doctor, version
  • PyPI smoke test — scripts/pypi_smoke_test.py validates install in a fresh venv (CI + verify)
  • Core-only import — import quantwave works without Polars (streaming, metadata, list API)

Product guardrails (Tier 1)

  • ./scripts/quantwave_verify.sh — one command for metadata drift, doc lint, nextest, and pytest smoke (local + CI parity)
  • Metadata drift gate in CI — generated Python/Rust metadata cannot drift silently
  • Streamlined GitHub Actions — two workflows: CI (every push) and Release (tag v* → crates.io + PyPI)
  • Plugin vs .ta guide — when to use expression plugins vs the Polars namespace
  • Regime user guide — expanded HMM/GMM/changepoint documentation

Research loop (Tier 2)

  • qw.build_feature_matrix() — batch ML feature matrix from OHLCV
  • lf.ta().features().recommended_matrix() (Rust Polars)
  • lf.bt.monte_carlo() — trade bootstrap + return-path VaR/CVaR (Python)
  • Rust Polars .bt — walk_forward_optimize, monte_carlo_trade_bootstrap

Indicators & ML

  • Fractional differencing (FracDiff) — Prado-style stationary features with streaming/batch parity
  • 55 custom Polars expression plugins — remaining custom indicators as native plugins
  • PA foundation — S/R Polars, confluence, geometric patterns with H&S neckline breakout
  • 98 auto-generated pyo3-polars bindings for standard indicators
  • 217 registered indicators with full IndicatorMetadata pipeline

Backtest & reporting

  • HTML tear sheets — BacktestReport.to_html() / save_html() with equity curve, drawdown, and trade table
  • Research-complete engine unchanged: sweep, WFO, cross-sectional, Monte Carlo (Rust)

Documentation

  • Indicator doc SOA complete — 217 indicators, 220+ native pages under DOCUMENTATION_STANDARDS.md
  • Slug redirect stubs — metadata slugs (sdo, rsi, adaptive_ema, …) resolve to canonical doc pages
  • Complete catalog — auto-generated index at guides/indicators/native/index.md
  • Landing page redesign; Roadmap moved under Community; legacy TA-Lib nav removed
  • Doc drift script — scripts/check_doc_drift.py + sync_indicator_docs.py wired into verify

Upgrade notes

  • No breaking API changes from 0.5.2 for core indicator/backtest surfaces
  • New optional APIs: frac_diff, build_feature_matrix, to_html tear sheets
  • Run ./scripts/quantwave_verify.sh before upgrading production pipelines

Quality gates

./scripts/quantwave_verify.sh

What's next

See Roadmap: portfolio-wide backtest (quantwave-8v4s), live Nautilus bridge (deferred).