1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
//! Median Moving Average.
use std::collections::VecDeque;
use crate::error::{Error, Result};
use crate::traits::Indicator;
/// Median Moving Average — the rolling median of the last `period` inputs.
///
/// For an odd `period` the output is the middle order statistic of the window;
/// for an even `period` it is the average of the two central values. Because it
/// is a rank statistic rather than a sum, the median MA is far more robust to
/// single outliers than the [`Sma`](crate::Sma): a lone spike shifts the rank
/// by at most one position instead of dragging the whole average.
///
/// Each `update` slides the window and computes the median by sorting a copy of
/// the `period` buffered values — O(`period` · log `period`) per step, with the
/// period fixed and bounded.
///
/// # Example
///
/// ```
/// use wickra_core::{Indicator, MedianMa};
///
/// let mut indicator = MedianMa::new(5).unwrap();
/// let mut last = None;
/// for i in 0..80 {
/// last = indicator.update(100.0 + f64::from(i));
/// }
/// assert!(last.is_some());
/// ```
#[derive(Debug, Clone)]
pub struct MedianMa {
period: usize,
window: VecDeque<f64>,
/// Reusable scratch buffer to avoid allocating per `update`.
scratch: Vec<f64>,
/// Median of the current window, recomputed by `update`.
last: Option<f64>,
}
impl MedianMa {
/// Construct a new median moving average over `period` inputs.
///
/// # Errors
///
/// Returns [`Error::PeriodZero`] if `period == 0`.
pub fn new(period: usize) -> Result<Self> {
if period == 0 {
return Err(Error::PeriodZero);
}
if period > crate::error::MAX_PERIOD {
return Err(Error::InvalidPeriod {
message: crate::error::PERIOD_ABOVE_MAX,
});
}
Ok(Self {
period,
window: VecDeque::with_capacity(period),
scratch: Vec::with_capacity(period),
last: None,
})
}
/// Configured period.
pub const fn period(&self) -> usize {
self.period
}
/// Current value if the window is full.
///
/// Cheap: the median is computed once per `update` rather than on every
/// read, which also keeps the sort off the caller's path.
pub const fn value(&self) -> Option<f64> {
self.last
}
/// Recompute the median of the live window into `last`.
fn recompute(&mut self) {
if self.window.len() != self.period {
self.last = None;
return;
}
self.scratch.clear();
self.scratch.extend(self.window.iter().copied());
// Total ordering rather than `partial_cmp`: the window only ever holds
// finite values, but this needs no justification to stay correct.
self.scratch.sort_unstable_by(f64::total_cmp);
let mid = self.period / 2;
self.last = Some(if self.period % 2 == 1 {
self.scratch[mid]
} else {
f64::midpoint(self.scratch[mid - 1], self.scratch[mid])
});
}
}
impl Indicator for MedianMa {
type Input = f64;
type Output = f64;
#[inline]
fn update(&mut self, input: f64) -> Option<f64> {
if !input.is_finite() {
return None;
}
if self.window.len() == self.period {
self.window.pop_front();
}
self.window.push_back(input);
self.recompute();
self.last
}
fn reset(&mut self) {
self.window.clear();
self.scratch.clear();
self.last = None;
}
#[inline]
fn warmup_period(&self) -> usize {
self.period
}
#[inline]
fn is_ready(&self) -> bool {
self.window.len() == self.period
}
#[inline]
fn name(&self) -> &'static str {
"MedianMA"
}
}
#[cfg(test)]
mod tests {
use super::*;
use crate::traits::BatchExt;
use approx::assert_relative_eq;
#[test]
fn new_rejects_zero_period() {
assert!(matches!(MedianMa::new(0), Err(Error::PeriodZero)));
}
/// Cover the const accessor `period` and the Indicator-impl `warmup_period`
/// + `name`.
#[test]
fn accessors_and_metadata() {
let mma = MedianMa::new(7).unwrap();
assert_eq!(mma.period(), 7);
assert_eq!(mma.warmup_period(), 7);
assert_eq!(mma.name(), "MedianMA");
}
#[test]
fn warmup_returns_none_then_odd_median() {
let mut mma = MedianMa::new(3).unwrap();
assert_eq!(mma.update(5.0), None);
assert_eq!(mma.update(1.0), None);
// median of [5, 1, 3] = 3 (middle order statistic).
assert_relative_eq!(mma.update(3.0).unwrap(), 3.0, epsilon = 1e-12);
}
#[test]
fn even_period_averages_two_central_values() {
// median of [1, 2, 3, 4] = (2 + 3) / 2 = 2.5.
let mut mma = MedianMa::new(4).unwrap();
let v = mma.batch(&[1.0, 2.0, 3.0, 4.0]);
assert_relative_eq!(v[3].unwrap(), 2.5, epsilon = 1e-12);
}
#[test]
fn robust_to_single_outlier() {
// A lone spike does not move the median of an odd window the way it
// would move an SMA. median of [10, 11, 9999] = 11.
let mut mma = MedianMa::new(3).unwrap();
let v = mma.batch(&[10.0, 11.0, 9999.0]);
assert_relative_eq!(v[2].unwrap(), 11.0, epsilon = 1e-12);
}
#[test]
fn period_one_is_pass_through() {
let mut mma = MedianMa::new(1).unwrap();
assert_relative_eq!(mma.update(5.5).unwrap(), 5.5, epsilon = 1e-12);
assert_relative_eq!(mma.update(7.5).unwrap(), 7.5, epsilon = 1e-12);
}
#[test]
fn slides_window_correctly() {
// After [1,2,3] the window slides to [2,3,4] -> median 3, then [3,4,5] -> 4.
let mut mma = MedianMa::new(3).unwrap();
let v = mma.batch(&[1.0, 2.0, 3.0, 4.0, 5.0]);
assert_relative_eq!(v[2].unwrap(), 2.0, epsilon = 1e-12);
assert_relative_eq!(v[3].unwrap(), 3.0, epsilon = 1e-12);
assert_relative_eq!(v[4].unwrap(), 4.0, epsilon = 1e-12);
}
#[test]
fn reset_clears_state() {
let mut mma = MedianMa::new(4).unwrap();
mma.batch(&[1.0, 2.0, 3.0, 4.0, 5.0]);
assert!(mma.is_ready());
mma.reset();
assert!(!mma.is_ready());
assert_eq!(mma.update(10.0), None);
}
#[test]
fn batch_equals_streaming() {
let prices: Vec<f64> = (1..=20).map(|i| (f64::from(i) * 0.7).sin() * 5.0).collect();
let mut a = MedianMa::new(5).unwrap();
let mut b = MedianMa::new(5).unwrap();
assert_eq!(
a.batch(&prices),
prices.iter().map(|p| b.update(*p)).collect::<Vec<_>>()
);
}
#[test]
fn ignores_non_finite_input_but_keeps_state() {
let mut mma = MedianMa::new(3).unwrap();
mma.update(5.0);
mma.update(1.0);
let _ready = mma
.update(3.0)
.expect("MedianMA(3) ready after three inputs");
assert_eq!(mma.update(f64::NAN), None);
assert_eq!(mma.update(f64::INFINITY), None);
// Window still [5, 1, 3] -> next real input slides to [1, 3, 8] -> median 3.
assert_relative_eq!(mma.update(8.0).unwrap(), 3.0, epsilon = 1e-12);
}
}