Skip to content

Release 0.5.2 — Research Platform

Date: 2026-05-31 (engine); docs/DX polish through 2026-06-26
Install: pip install quantwave or pip install "quantwave[polars]"

Headline

QuantWave 0.5.2 is the first release where indicators, plugins, backtest, and Python DX form a coherent research platform — not a collection of crates.

What's new

Indicators & PA

  • Full plugin parity (~219 .ta expression methods) with quantwave-polars
  • Price Action foundation: Market Structure, S/R monitor (ATR-relative), geometric patterns with H&S neckline breakout, PA confluence
  • ML features: .ta.features.* (Hurst, CyberCycle, ITL, trendflex, regime probs)

Backtest

  • Research-complete engine: sweep, walk-forward, WFO-optimize, cross-sectional, Monte Carlo (Rust)
  • Polars .bt namespace in Python

Python DX

  • qw.indicators(), qw.metadata(), qw.categories(), qw.boundary_info()
  • qw.assert_parity() — batch vs streaming verification
  • quantwave.talib with list_functions()
  • QuantwaveError hierarchy
  • Linux arm64 (manylinux_2_34_aarch64) + macOS arm64 wheels

Metadata pipeline

  • Rust *_METADATA → metadata_registry.rs → _metadata_generated.py (217 entries)
  • TrackedNext / qw.track_streaming() for warmup-aware live use

Documentation

Upgrade notes

  • Options helpers moved to quantwave.options (top-level deprecated with warning)
  • Polars layer is optional: import quantwave works without polars installed
  • Use pip install "quantwave[polars]" for .ta and .bt

Quality gates

./scripts/quantwave_verify.sh

What's next (v0.6)

See Roadmap and epic quantwave-motd: CI metadata gate, feature matrix helper, Python MC wrapper.