techr

techr exposes technical indicators as Polars expression plugins.

Installation

uv add techr

Supported indicators

• Price/trend: sma, wma, ema, disparity, mom, roc, rsi, rsi_line, rsi_signal, psl
• Bands/channels: bband_middle, bband_lower, bband_upper, env_upper, env_middle, env_lower, pchan_upper, pchan_middle, pchan_lower
• Momentum/oscillators: macd, macd_line, macd_signal, macd_hist, macd_histogram, ppo_line, ppo_signal, ppo_histogram, pvo_line, pvo_signal, pvo_histogram, sonar_line, sonar_signal, trix_line, trix_signal, stochf_percent_k, stochf_percent_d, stoch_percent_k, stoch_percent_d, stochrsi_percent_k, stochrsi_percent_d
• High/low/volume indicators: ad, adx, adxr, aroon_up, aroon_down, aroonosc, atr, cci, cci_line, cci_signal, cmf, co, cv, dmi_plus, dmi_minus, efi, eom_line, eom_signal, erbear, erbull, massi_line, massi_signal, mfi, nvi_line, nvi_signal, obv_line, obv_signal, psar, pvi_line, pvi_signal, ultosc, vr, willr
• Ichimoku: ichimoku_base_line, ichimoku_conversion_line, ichimoku_leading_span_a, ichimoku_leading_span_b, ichimoku_lagging_span

Usage

import polars as pl
import techr as ta

df = pl.DataFrame(
    {
        "high": [11.0, 12.0, 13.0],
        "low": [9.0, 10.0, 11.0],
        "close": [10.0, 11.0, 12.0],
    }
)

result = df.select(
    ta.sma(pl.col("close"), period=2).alias("sma_2"),
    ta.macd(pl.col("close"), fast_period=12, slow_period=26).alias("macd"),
    ta.stochf_percent_k(
        pl.col("high"),
        pl.col("low"),
        pl.col("close"),
        fastk_period=14,
        fastd_period=3,
    ).alias("stochf_k"),
)

Null input values are accepted. Indicators preserve row alignment and emit nulls according to the core rolling-window and seed recovery rules.

Ichimoku Notes

• Standalone Ichimoku rolling-window lines such as ichimoku_base_line and ichimoku_conversion_line use period.
• ichimoku_leading_span_a uses base_line_period and conversion_line_period because it combines two lines.
• ichimoku_leading_span_b uses period for the rolling window and base_line_period for the forward displacement. The Python wrapper defaults base_line_period to 26.
• ichimoku_lagging_span uses base_line_period for its backward displacement.
• Polars plugins keep the output row-aligned with the input, so ichimoku_leading_span_a and ichimoku_leading_span_b truncate the forward-projected tail from the core result.

Development

cd polars
uv sync --group dev
uv run maturin develop --uv
uv run pytest

Build distributable artifacts locally with:

cd polars
uv run maturin build --release --sdist --out dist
uv run python scripts/check_artifacts.py dist

Release

1. Update the version in Cargo.toml.
2. Merge the version bump to main.
3. Polars Release builds the artifacts and publishes them to PyPI.
4. After the PyPI publish succeeds, the workflow creates the matching polars-vX.Y.Z GitHub Release.

If the matching GitHub Release already exists, the workflow exits without publishing a duplicate release.

Before the first release, configure a Trusted Publisher for PyPI on alphaprime-dev/techr.