py-alpha-lib 0.2.1__tar.gz → 0.2.3__tar.gz

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/CHANGELOG.md

@@ -1,5 +1,17 @@
 # ChangeLog
 
+## [0.2.3] - 2026-04-13
+
+### Added
+
+- ExecContext: TS_QUANTILE, SLOPE, INTERCEPT, DMA, SUMBARS, BARSLAST, BARSSINCE, BINS, FRET, RCROSS, RLONGCROSS
+
+## [0.2.2] - 2026-04-11
+
+### Added
+
+- QUANTILE
+
 ## [0.2.1] - 2026-04-06
 
 ### Fix

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/Cargo.lock

@@ -4,7 +4,7 @@ version = 4
 
 [[package]]
 name = "alpha"
-version = "0.2.1"
+version = "0.2.3"
 dependencies = [
  "anyhow",
  "log",

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "alpha"
-version = "0.2.1"
+version = "0.2.3"
 edition = "2024"
 authors = ["ElseJJ"]
 readme = "README.md"

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: py-alpha-lib
-Version: 0.2.1
+Version: 0.2.3
 Classifier: Programming Language :: Rust
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy
@@ -203,6 +203,7 @@ Naming Rules:
 | MOMENT | Calculate rolling k-th central moment over a moving window |
 | NEUTRALIZE | Neutralize the effect of a categorical variable on a numeric variable |
 | PRODUCT | Calculate product of values in preceding `periods` window |
+| QUANTILE | Calculate rolling quantile over a moving window |
 | RANK | Calculate rank in a sliding window with size `periods` |
 | RCROSS | For 2 arrays A and B, return true if A[i-1] > B[i-1] and A[i] <= B[i] alias: death_cross, cross_le |
 | REF | Right shift input array by `periods`, r[i] = input[i - periods] |

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/README.md

@@ -188,6 +188,7 @@ Naming Rules:
 | MOMENT | Calculate rolling k-th central moment over a moving window |
 | NEUTRALIZE | Neutralize the effect of a categorical variable on a numeric variable |
 | PRODUCT | Calculate product of values in preceding `periods` window |
+| QUANTILE | Calculate rolling quantile over a moving window |
 | RANK | Calculate rank in a sliding window with size `periods` |
 | RCROSS | For 2 arrays A and B, return true if A[i-1] > B[i-1] and A[i] <= B[i] alias: death_cross, cross_le |
 | REF | Right shift input array by `periods`, r[i] = input[i - periods] |

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/python/alpha/algo/algo_gen.py

@@ -623,6 +623,30 @@ def PRODUCT(
     _algo.product(r, input, periods)
     return r
 
+def QUANTILE(
+  input: np.ndarray | list[np.ndarray], periods: int, q: float
+) -> np.ndarray | list[np.ndarray]:
+  """
+  Calculate rolling quantile over a moving window
+  
+  QUANTILE(x, d, q) returns the q-th quantile (0 <= q <= 1) of values
+  in the preceding d periods. Uses linear interpolation between data points
+  (matching numpy/pandas percentile with interpolation='linear').
+  NaN values are excluded from the computation. Requires at least 1 valid value.
+  
+  Ref: https://numpy.org/doc/stable/reference/generated/numpy.quantile.html
+  """
+  if isinstance(input, list):
+    input = [_to_f64(x) for x in input]
+    r = [np.empty_like(x) for x in input]
+    _algo.quantile(r, input, periods, q)
+    return r
+  else:
+    input = _to_f64(input)
+    r = np.empty_like(input)
+    _algo.quantile(r, input, periods, q)
+    return r
+
 def RANK(
   input: np.ndarray | list[np.ndarray], periods: int
 ) -> np.ndarray | list[np.ndarray]:

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/python/alpha/algo.md

@@ -33,6 +33,7 @@ the `np.ndarray` is `ndarray` type in `numpy` package
 - MOMENT(input: np.ndarray[float], periods: int, k: int): Calculate rolling k-th central moment over a moving window  MOMENT(x, d, k) = mean((x - mean)^k) over window of d periods. This is the raw (non-adjusted) sample moment. k=2 gives variance (population), k=3 gives raw third moment, etc.
 - NEUTRALIZE(category: np.ndarray[float], input: np.ndarray[float]): Neutralize the effect of a categorical variable on a numeric variable
 - PRODUCT(input: np.ndarray[float], periods: int): Calculate product of values in preceding `periods` window  If periods is 0, it calculates the cumulative product from the first valid value.
+- QUANTILE(input: np.ndarray[float], periods: int, q: float): Calculate rolling quantile over a moving window  QUANTILE(x, d, q) returns the q-th quantile (0 <= q <= 1) of values in the preceding d periods. Uses linear interpolation between data points (matching numpy/pandas percentile with interpolation='linear'). NaN values are excluded from the computation. Requires at least 1 valid value.
 - RANK(input: np.ndarray[float], periods: int): Calculate rank in a sliding window with size `periods`  Uses min-rank method for ties (same as pandas rankdata method='min'). NaN values are treated as larger than all non-NaN values.
 - RCROSS(a: np.ndarray[float], b: np.ndarray[float]): For 2 arrays A and B, return true if A[i-1] > B[i-1] and A[i] <= B[i] alias: death_cross, cross_le
 - REF(input: np.ndarray[float], periods: int): Right shift input array by `periods`, r[i] = input[i - periods]

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/python/alpha/context.py

@@ -249,6 +249,9 @@ class ExecContext:
     else:
       return alpha.SMA(a, int(args[0]), int(args[1]))
 
+  def DMA(self, a: np.ndarray, weight: float) -> np.ndarray:
+    return alpha.DMA(a, float(weight))
+
   # ── TS: Std Dev / Variance ─────────────────────────────────────────
   #   BRAIN: ts_std_dev    GTJA: STD               wq101: STDDEV
 
@@ -315,6 +318,11 @@ class ExecContext:
 
   TSRANK = TS_RANK  # gtja191
 
+  # ── TS: Quantile ────────────────────────────────────────────────────
+
+  def TS_QUANTILE(self, a: np.ndarray, w: int, q: float) -> np.ndarray:
+    return alpha.QUANTILE(a, int(w), q)
+
   # ── TS: Delay / Delta ──────────────────────────────────────────────
   #   BRAIN: ts_delay, ts_delta
   #   wq101: DELAY, DELTA
@@ -364,6 +372,14 @@ class ExecContext:
 
   REGRESI = TS_REGRESI  # gtja191
 
+  # ── TS: Slope / Intercept ───────────────────────────────────────
+
+  def SLOPE(self, a: np.ndarray, w: int) -> np.ndarray:
+    return alpha.SLOPE(a, int(w))
+
+  def INTERCEPT(self, a: np.ndarray, w: int) -> np.ndarray:
+    return alpha.INTERCEPT(a, int(w))
+
   # ── TS: Counting / Conditional ─────────────────────────────────────
   #   GTJA: COUNT, SUMIF
 
@@ -379,6 +395,17 @@ class ExecContext:
 
   SUMIF = TS_SUMIF  # gtja191
 
+  def SUMBARS(self, a: np.ndarray, amount: float) -> np.ndarray:
+    return alpha.SUMBARS(a, amount)
+
+  # ── TS: Bar Counting ──────────────────────────────────────────
+
+  def BARSLAST(self, cond: np.ndarray) -> np.ndarray:
+    return alpha.BARSLAST(np.asarray(cond, dtype=bool))
+
+  def BARSSINCE(self, cond: np.ndarray) -> np.ndarray:
+    return alpha.BARSSINCE(np.asarray(cond, dtype=bool))
+
   # ── TS: Conditional Scan (SELF recursion) ───────────────────────────
   #   GTJA: SELF-referencing patterns
 
@@ -440,6 +467,16 @@ class ExecContext:
   def TS_MOMENT(self, a: np.ndarray, w: int, k: int = 2) -> np.ndarray:
     return alpha.MOMENT(a, int(w), int(k))
 
+  # ── TS: Binning ────────────────────────────────────────────────
+
+  def BINS(self, a: np.ndarray, n: int) -> np.ndarray:
+    return alpha.BINS(a, int(n))
+
+  # ── TS: Future Return ──────────────────────────────────────────
+
+  def FRET(self, open: np.ndarray, close: np.ndarray, is_calc: np.ndarray, delay: int, periods: int) -> np.ndarray:
+    return alpha.FRET(open, close, is_calc, int(delay), int(periods))
+
   # ── TS: Cross Detection ────────────────────────────────────────────
   #   AmiBroker/GTJA: CROSS, LONGCROSS
 
@@ -449,6 +486,12 @@ class ExecContext:
   def LONGCROSS(self, a: np.ndarray, b: np.ndarray, n: int) -> np.ndarray:
     return alpha.LONGCROSS(a, b, int(n))
 
+  def RCROSS(self, a: np.ndarray, b: np.ndarray) -> np.ndarray:
+    return alpha.RCROSS(a, b)
+
+  def RLONGCROSS(self, a: np.ndarray, b: np.ndarray, n: int) -> np.ndarray:
+    return alpha.RLONGCROSS(a, b, int(n))
+
   # ====================================================================
   #  Cross-Sectional Operators (no prefix)
   #

{py_alpha_lib-0.2.1 → py_alpha_lib-0.2.3}/src/algo/mod.rs

@@ -13,6 +13,7 @@ mod ma;
 mod misc;
 mod moments;
 mod neutralize;
+mod quantile;
 mod rank;
 mod returns;
 mod scan;
@@ -36,6 +37,7 @@ pub use ma::*;
 pub use misc::*;
 pub use moments::*;
 pub use neutralize::*;
+pub use quantile::*;
 pub use rank::*;
 pub use returns::*;
 pub use scan::*;

py_alpha_lib-0.2.3/src/algo/quantile.rs

@@ -0,0 +1,224 @@
+// Copyright 2026 MSD-RS Project LiJia
+// SPDX-License-Identifier: BSD-2-Clause
+
+use num_traits::Float;
+use rayon::prelude::*;
+
+use crate::algo::{Context, Error, is_normal};
+
+/// Calculate rolling quantile over a moving window
+///
+/// QUANTILE(x, d, q) returns the q-th quantile (0 <= q <= 1) of values
+/// in the preceding d periods. Uses linear interpolation between data points
+/// (matching numpy/pandas percentile with interpolation='linear').
+/// NaN values are excluded from the computation. Requires at least 1 valid value.
+///
+/// Ref: https://numpy.org/doc/stable/reference/generated/numpy.quantile.html
+pub fn ta_quantile<NumT: Float + Send + Sync>(
+  ctx: &Context,
+  r: &mut [NumT],
+  input: &[NumT],
+  periods: usize,
+  q: NumT,
+) -> Result<(), Error> {
+  if r.len() != input.len() {
+    return Err(Error::LengthMismatch(r.len(), input.len()));
+  }
+
+  if periods == 0 {
+    r.fill(NumT::nan());
+    return Ok(());
+  }
+
+  if q < NumT::zero() || q > NumT::one() || q.is_nan() {
+    r.fill(NumT::nan());
+    return Ok(());
+  }
+
+  r.par_chunks_mut(ctx.chunk_size(r.len()))
+    .zip(input.par_chunks(ctx.chunk_size(input.len())))
+    .for_each(|(r, x)| {
+      let start = ctx.start(r.len());
+      let end = ctx.end(r.len());
+      r.fill(NumT::nan());
+
+      // Allocate a buffer for sorting within the window
+      let mut buf = vec![NumT::zero(); periods];
+
+      for i in start..end {
+        let val = x[i];
+        if !is_normal(&val) {
+          continue;
+        }
+
+        // Determine window boundaries
+        let win_start = if i >= periods { i - periods + 1 } else { 0 };
+        let win_end = i + 1;
+
+        // Collect valid (non-NaN) values into buffer
+        let mut count = 0;
+        for k in win_start..win_end {
+          let v = x[k];
+          if is_normal(&v) {
+            buf[count] = v;
+            count += 1;
+          }
+        }
+
+        if count == 0 {
+          continue;
+        }
+
+        // Check strict cycle: require full window of `periods` valid values
+        if ctx.is_strictly_cycle() {
+          if i < periods - 1 || count != periods || (win_end - win_start) != periods {
+            continue;
+          }
+        }
+
+        // Sort the buffer (insertion sort for small windows, efficient for typical d=5..50)
+        let slice = &mut buf[..count];
+        for j in 1..count {
+          let key = slice[j];
+          let mut k = j;
+          while k > 0 && slice[k - 1] > key {
+            slice[k] = slice[k - 1];
+            k -= 1;
+          }
+          slice[k] = key;
+        }
+
+        // Linear interpolation quantile (matching numpy/pandas)
+        let pos = q * NumT::from(count - 1).unwrap();
+        let lo = pos.floor();
+        let hi = lo + NumT::one();
+        let lo_idx = lo.to_usize().unwrap_or(0).min(count - 1);
+        let hi_idx = hi.to_usize().unwrap_or(0).min(count - 1);
+        let frac = pos - lo;
+        r[i] = slice[lo_idx] * (NumT::one() - frac) + slice[hi_idx] * frac;
+      }
+    });
+
+  Ok(())
+}
+
+#[cfg(test)]
+mod tests {
+  use super::*;
+  use crate::algo::{assert_vec_eq_nan, context::FLAG_STRICTLY_CYCLE};
+
+  #[test]
+  fn test_quantile_median() {
+    // Median (q=0.5) of [1,2,3] = 2.0
+    let input = vec![1.0, 2.0, 3.0, 4.0, 5.0];
+    let periods = 3;
+    let mut r = vec![0.0; input.len()];
+    let ctx = Context::new(0, 0, 0);
+    ta_quantile(&ctx, &mut r, &input, periods, 0.5).unwrap();
+
+    // i=0: [1] -> median=1.0
+    // i=1: [1,2] -> median=1.5
+    // i=2: [1,2,3] -> median=2.0
+    // i=3: [2,3,4] -> median=3.0
+    // i=4: [3,4,5] -> median=4.0
+    assert_vec_eq_nan(&r, &vec![1.0, 1.5, 2.0, 3.0, 4.0]);
+  }
+
+  #[test]
+  fn test_quantile_strict() {
+    let input = vec![1.0, 2.0, 3.0, 4.0, 5.0];
+    let periods = 3;
+    let mut r = vec![0.0; input.len()];
+    let ctx = Context::new(0, 0, FLAG_STRICTLY_CYCLE);
+    ta_quantile(&ctx, &mut r, &input, periods, 0.5).unwrap();
+
+    assert_vec_eq_nan(
+      &r,
+      &vec![f64::NAN, f64::NAN, 2.0, 3.0, 4.0],
+    );
+  }
+
+  #[test]
+  fn test_quantile_q0_q1() {
+    let input = vec![3.0, 1.0, 4.0, 1.0, 5.0];
+    let periods = 3;
+    let mut r = vec![0.0; input.len()];
+    let ctx = Context::new(0, 0, FLAG_STRICTLY_CYCLE);
+
+    // q=0 -> min
+    ta_quantile(&ctx, &mut r, &input, periods, 0.0).unwrap();
+    assert_vec_eq_nan(
+      &r,
+      &vec![f64::NAN, f64::NAN, 1.0, 1.0, 1.0],
+    );
+
+    // q=1 -> max
+    ta_quantile(&ctx, &mut r, &input, periods, 1.0).unwrap();
+    assert_vec_eq_nan(
+      &r,
+      &vec![f64::NAN, f64::NAN, 4.0, 4.0, 5.0],
+    );
+  }
+
+  #[test]
+  fn test_quantile_with_nan() {
+    let input = vec![1.0, f64::NAN, 3.0, 4.0, 5.0];
+    let periods = 3;
+    let mut r = vec![0.0; input.len()];
+    let ctx = Context::new(0, 0, 0);
+    ta_quantile(&ctx, &mut r, &input, periods, 0.5).unwrap();
+
+    // i=0: [1] -> 1.0
+    // i=1: NaN -> NaN
+    // i=2: [1, 3] (NaN skipped) -> median of [1,3] = 2.0
+    // i=3: [3, 4] (NaN skipped) -> median of [3,4] = 3.5
+    // i=4: [3, 4, 5] -> median = 4.0
+    assert_vec_eq_nan(&r, &vec![1.0, f64::NAN, 2.0, 3.5, 4.0]);
+  }
+
+  #[test]
+  fn test_quantile_groups() {
+    // 2 groups of 3 each
+    let input = vec![1.0, 2.0, 3.0, 10.0, 20.0, 30.0];
+    let periods = 3;
+    let mut r = vec![0.0; input.len()];
+    let ctx = Context::new(0, 2, FLAG_STRICTLY_CYCLE);
+    ta_quantile(&ctx, &mut r, &input, periods, 0.5).unwrap();
+
+    // Group 1: [1,2,3] -> median=2.0
+    // Group 2: [10,20,30] -> median=20.0
+    assert_vec_eq_nan(
+      &r,
+      &vec![f64::NAN, f64::NAN, 2.0, f64::NAN, f64::NAN, 20.0],
+    );
+  }
+
+  #[test]
+  fn test_quantile_periods_zero() {
+    let input = vec![1.0, 2.0, 3.0];
+    let mut r = vec![0.0; input.len()];
+    let ctx = Context::new(0, 0, 0);
+    ta_quantile(&ctx, &mut r, &input, 0, 0.5).unwrap();
+    // periods=0 → all NaN
+    assert!(r.iter().all(|x| x.is_nan()));
+  }
+
+  #[test]
+  fn test_quantile_invalid_q() {
+    let input = vec![1.0, 2.0, 3.0];
+    let mut r = vec![0.0; input.len()];
+    let ctx = Context::new(0, 0, 0);
+
+    // q < 0
+    ta_quantile(&ctx, &mut r, &input, 3, -0.1).unwrap();
+    assert!(r.iter().all(|x| x.is_nan()));
+
+    // q > 1
+    ta_quantile(&ctx, &mut r, &input, 3, 1.5).unwrap();
+    assert!(r.iter().all(|x| x.is_nan()));
+
+    // q = NaN
+    ta_quantile(&ctx, &mut r, &input, 3, f64::NAN).unwrap();
+    assert!(r.iter().all(|x| x.is_nan()));
+  }
+}