quantwave-core 0.1.14

use crate::indicators::metadata::{IndicatorMetadata, ParamDef};
use crate::traits::Next;
use std::collections::VecDeque;

/// Donchian Channels
/// Upper Band: Highest high over the last N periods.
/// Lower Band: Lowest low over the last N periods.
/// Middle Band: (Upper + Lower) / 2
#[derive(Debug, Clone)]
pub struct DonchianChannels {
    period: usize,
    highs: VecDeque<f64>,
    lows: VecDeque<f64>,
}

impl DonchianChannels {
    pub fn new(period: usize) -> Self {
        Self {
            period,
            highs: VecDeque::with_capacity(period),
            lows: VecDeque::with_capacity(period),
        }
    }
}

impl Next<(f64, f64)> for DonchianChannels {
    type Output = (f64, f64, f64);

    fn next(&mut self, (high, low): (f64, f64)) -> Self::Output {
        self.highs.push_back(high);
        self.lows.push_back(low);

        if self.highs.len() > self.period {
            self.highs.pop_front();
            self.lows.pop_front();
        }

        let mut max_high = f64::MIN;
        let mut min_low = f64::MAX;

        for &h in self.highs.iter() {
            if h > max_high {
                max_high = h;
            }
        }

        for &l in self.lows.iter() {
            if l < min_low {
                min_low = l;
            }
        }

        let middle = (max_high + min_low) / 2.0;

        (max_high, middle, min_low)
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use proptest::prelude::*;
    use serde::Deserialize;
    use std::fs;
    use std::path::Path;

    #[derive(Debug, Deserialize)]
    struct DonchianCase {
        highs: Vec<f64>,
        lows: Vec<f64>,
        expected_middle: Vec<f64>,
    }

    #[test]
    fn test_donchian_gold_standard() {
        let manifest_dir = std::env::var("CARGO_MANIFEST_DIR").unwrap();
        let manifest_path = Path::new(&manifest_dir);
        let path = manifest_path.join("tests/gold_standard/donchian_5.json");
        let path = if path.exists() {
            path
        } else {
            manifest_path
                .parent()
                .unwrap()
                .join("tests/gold_standard/donchian_5.json")
        };
        let content = fs::read_to_string(path).unwrap();
        let case: DonchianCase = serde_json::from_str(&content).unwrap();

        let mut dc = DonchianChannels::new(5);
        for i in 0..case.highs.len() {
            let (_, middle, _) = dc.next((case.highs[i], case.lows[i]));
            approx::assert_relative_eq!(middle, case.expected_middle[i]);
        }
    }

    #[test]
    fn test_donchian_basic() {
        let mut dc = DonchianChannels::new(3);

        // bar 1: H=10, L=8 -> U=10, M=9, L=8
        let (u1, m1, l1) = dc.next((10.0, 8.0));
        assert_eq!(u1, 10.0);
        assert_eq!(m1, 9.0);
        assert_eq!(l1, 8.0);

        // bar 2: H=12, L=7 -> U=12, M=9.5, L=7
        let (u2, m2, l2) = dc.next((12.0, 7.0));
        assert_eq!(u2, 12.0);
        assert_eq!(m2, 9.5);
        assert_eq!(l2, 7.0);

        // bar 3: H=11, L=9 -> U=12, M=9.5, L=7
        let (u3, m3, l3) = dc.next((11.0, 9.0));
        assert_eq!(u3, 12.0);
        assert_eq!(m3, 9.5);
        assert_eq!(l3, 7.0);

        // bar 4: H=13, L=10 -> U=13, M=10, L=7 (bar 1 is out)
        let (u4, m4, l4) = dc.next((13.0, 10.0));
        assert_eq!(u4, 13.0);
        assert_eq!(m4, 10.0);
        assert_eq!(l4, 7.0);
    }

    fn donchian_batch(data: Vec<(f64, f64)>, period: usize) -> Vec<f64> {
        let mut dc = DonchianChannels::new(period);
        // We'll just return the middle band for parity check to simplify,
        // or we could change the parity helper to handle tuples.
        data.into_iter().map(|x| dc.next(x).1).collect()
    }

    proptest! {
        #[test]
        fn test_donchian_parity(highs in prop::collection::vec(0.0..1000.0, 1..100), lows in prop::collection::vec(0.0..1000.0, 1..100)) {
            let len = highs.len().min(lows.len());
            let highs: Vec<f64> = highs[..len].to_vec();
            let lows: Vec<f64> = lows[..len].to_vec();
            let mut input = Vec::with_capacity(len);
            for i in 0..len {
                let h = highs[i];
                let l = lows[i].min(h); // Ensure low <= high
                input.push((h, l));
            }

            let period = 5;
            let mut dc = DonchianChannels::new(period);
            let mut streaming_results = Vec::with_capacity(len);
            for &val in &input {
                streaming_results.push(dc.next(val).1);
            }

            let batch_results = donchian_batch(input, period);

            for (s, b) in streaming_results.iter().zip(batch_results.iter()) {
                assert_eq!(s, b);
            }
        }
    }
}

pub const DONCHIAN_METADATA: IndicatorMetadata = IndicatorMetadata {
    name: "Donchian Channels",
    description: "Donchian Channels are volatility indicators formed by taking the highest high and the lowest low of the last N periods.",
    usage: "Use for breakout trading systems: a close above the N-period high signals a long entry; below the N-period low signals a short entry. The Turtle Traders famously used 20 and 55-day Donchian channels.",
    keywords: &["breakout", "volatility", "trend", "classic", "support-resistance"],
    ehlers_summary: "Developed by Richard Donchian in the 1970s, Donchian Channels plot the highest high and lowest low over N bars. They define the current trading range and signal breakouts when price escapes the channel. The Turtle Trading system of Richard Dennis built its entire entry and exit logic on 20 and 55-day Donchian channels. — TurtleTrader.com",
    params: &[ParamDef {
        name: "period",
        default: "20",
        description: "Channel period",
    }],
    formula_source: "https://www.investopedia.com/terms/d/donchianchannels.asp",
    formula_latex: r#"
\[
UC = \max(H_{t-n \dots t}) \\ LC = \min(L_{t-n \dots t})
\]
"#,
    gold_standard_file: "donchian.json",
    category: "Classic",
};