QuantWave (Python)

Polars-native quantitative finance — 221 Rust indicators, Ehlers DSP, price action, regimes, and a built-in backtest engine with batch ↔ streaming parity.

pip install "quantwave[polars]"
quantwave doctor

Documentation · vs TA-Lib & pandas-ta · FAQ · llms.txt

Install

pip install "quantwave[polars]"

The wheel bundles quantwave._quantwave (indicators), quantwave_plugins (expression .ta), and quantwave._backtest. The [polars] extra adds Polars for .ta / .bt namespaces.

Example 1 — Polars batch (.ta)

import polars as pl
import quantwave  # registers pl.col().ta and LazyFrame.bt

df = pl.DataFrame({
    "high":  [101.0, 102.0, 103.0, 102.0, 104.0],
    "low":   [99.0, 100.0, 101.0, 100.0, 102.0],
    "close": [100.0, 101.0, 102.0, 101.0, 103.0],
})

out = (
    df.lazy()
    .with_columns(
        pl.col("close").ta.rsi(timeperiod=14).alias("rsi"),
        pl.col("close").ta.supertrend("high", "low", period=10, multiplier=3.0).alias("st"),
    )
    .collect()
)
print(out.tail())

Example 2 — Live streaming (parity with batch)

import quantwave as qw

cls = qw.streaming_class("rsi")
rsi = qw.wrap_streaming(cls(period=14), name="rsi")

for price in closes:
    val = rsi.next(price)
    if rsi.is_ready:
        print(price, val)

# Optional: qw.assert_parity("rsi", {"period": 14}, closes)

Example 3 — Backtest (.bt in the same package)

import polars as pl
import quantwave

bars = pl.DataFrame({
    "timestamp": range(20),
    "close": [100 + i * 0.5 for i in range(20)],
    "signal": [0.0, 0.0, 1.0, 1.0, 1.0, 0.0, 0.0, -1.0, -1.0, 0.0] * 2,
})

report = (
    bars.lazy()
    .bt.backtest_with_report(signal="signal", close_col="close", timestamp_col="timestamp")
)
print(report.metrics())

See examples/03_backtest_quick.py in this directory.

Architecture

pip install quantwave[polars]
        │
        ├── quantwave._quantwave   # Rust indicators (Next<T> core)
        ├── quantwave_plugins    # pl.col("close").ta.* expression plugins
        ├── quantwave._backtest  # Rust simulation engine
        │
        ├── import quantwave
        │     ├── pl.col().ta.*        batch indicators (plugins)
        │     ├── df.lazy().bt.*       backtest namespace (bt_polars)
        │     ├── qw.streaming_class   live bar-by-bar
        │     ├── qw.metadata / assert_parity / CLI
        │     └── quantwave.backtest   BacktestEngine direct API
        │
        └── Docs: lavs9.github.io/quantwave/api/ (mkdocstrings)

One Rust math path powers batch, streaming, and backtest signal evaluation — validated by gold-standard tests.

Discovery & CLI

import quantwave as qw

print(len(qw.indicators()), "names")
meta = qw.metadata("rsi")
print(meta.warmup_bars, meta.category)

quantwave list --category "Classic"
quantwave info supertrend
quantwave doctor

Examples in this repo

Learn more

• Getting Started funnel
• Python guide
• Backtest quickstart
• Indicator catalog
• Changelog

MIT licensed.