rustockstats

A Rust crate that provides stockstats-like technical indicators on top of a Polars DataFrame.

Indicators/derived columns are computed lazily on-demand: when you call get("rsi"), the column is created if missing and then returned.

Installation

[dependencies]
rustockstats = { git = "https://github.com/jealous/rustockstats.git" }
# or, once published:
# rustockstats = "0.x"

Usage

use polars::prelude::*;
use rustockstats::StockDataFrame;

fn main() -> anyhow::Result<()> {
    let df = df!(
        "open" => &[1.0, 2.0, 3.0],
        "high" => &[1.2, 2.2, 3.2],
        "low"  => &[0.8, 1.8, 2.8],
        "close"=> &[1.1, 2.1, 3.1],
        "volume" => &[10.0, 11.0, 12.0],
    )?;

    let mut sdf = StockDataFrame::retype(df)?;

    // Base indicator (uses default windows)
    let _rsi = sdf.get("rsi")?;

    // Parameterized indicator column name: <col>_<window>_<indicator>
    let _ema12 = sdf.get("close_12_ema")?;

    // Access the underlying DataFrame
    let out = sdf.into_df();
    println!("{}", out);

    Ok(())
}

Supported indicators (high level)

This crate supports a broad subset of Python stockstats, including:

• Moving averages: sma, ema, smma, tema, vwma, linear_wma
• Momentum/oscillators: rsi, stochrsi, wr, roc, cmo, kst
• Volatility/trend: atr, cci, chop, supertrend
• MACD family: macd, plus related signal/hist columns
• Volume-based: mfi, vr, pvo, ppo
• Others: boll (and boll_ub/boll_lb), ao, aroon, ichimoku, rvgi, qqe, cr, bop, cti, pgo, psl, ftr, kama, ker

Many indicators support parameterization by encoding windows into the column name (e.g. close_12_ema, rsi_6).

Comparison with Python stockstats

• API style: similar “computed column” model, but backed by Rust + Polars.
• Performance: intended to avoid repeated scans by caching computed columns in the DataFrame.
• Type system: errors are explicit via Result and StockStatsError.

License

BSD (see LICENSE).