Files
QuanTAlib/lib/trends/butter/Butter.cs
T
Miha Kralj 16a21a5b65 feat: Add Cumulative Moving Average (CMA) implementation with detailed documentation
- Introduced Cma class for calculating the Cumulative Moving Average using Welford's algorithm with FMA for precision.
- Added methods for batch processing and streaming updates.
- Implemented a comprehensive markdown documentation for CMA, covering its mathematical foundation, performance profile, and use cases.
- Enhanced existing trend indicators (Bessel, Butter, Htit, Jma, Mama, Ssf, Vidya) with FMA for improved numerical stability and precision.
- Updated Adosc to utilize a single-pass algorithm for performance optimization.
- Fixed date initialization in benchmarks to ensure UTC consistency.
2025-12-29 09:34:37 -08:00

211 lines
5.9 KiB
C#

using System.Runtime.CompilerServices;
using System.Runtime.InteropServices;
namespace QuanTAlib;
public sealed class Butter : AbstractBase
{
private readonly int _period;
private double _a1, _a2, _b0, _b1, _b2;
private double _invA0;
private readonly TValuePublishedHandler _handler;
private State _state;
private State _p_state;
[StructLayout(LayoutKind.Auto)]
private record struct State
{
public double X1, X2;
public double Y1, Y2;
public int Count;
}
public override bool IsHot => _state.Count >= 2;
public Butter(int period)
{
if (period < 2)
{
throw new ArgumentOutOfRangeException(nameof(period), "Period must be greater than or equal to 2.");
}
_period = period;
CalculateCoefficients();
Name = $"Butter({_period})";
WarmupPeriod = 2;
_handler = Handle;
Init();
}
public Butter(ITValuePublisher source, int period) : this(period)
{
source.Pub += _handler;
}
private void Handle(object? sender, in TValueEventArgs args)
{
Update(args.Value, args.IsNew);
}
[MethodImpl(MethodImplOptions.AggressiveInlining)]
private static void ComputeCoefficients(int period, out double a1, out double a2, out double b0, out double b1, out double b2, out double invA0)
{
double omega = 2.0 * Math.PI / period;
double sinOmega = Math.Sin(omega);
double cosOmega = Math.Cos(omega);
double alpha = sinOmega / Math.Sqrt(2.0);
double a0 = 1.0 + alpha;
a1 = -2.0 * cosOmega;
a2 = 1.0 - alpha;
b0 = (1.0 - cosOmega) / 2.0;
b1 = 1.0 - cosOmega;
b2 = (1.0 - cosOmega) / 2.0;
invA0 = 1.0 / a0;
}
[MethodImpl(MethodImplOptions.AggressiveInlining)]
private void CalculateCoefficients()
{
ComputeCoefficients(_period, out _a1, out _a2, out _b0, out _b1, out _b2, out _invA0);
}
[MethodImpl(MethodImplOptions.AggressiveInlining)]
private void Init()
{
_state = new State();
_p_state = new State();
Last = new TValue(0, double.NaN);
}
public override void Reset()
{
Init();
}
public override void Prime(ReadOnlySpan<double> source, TimeSpan? step = null)
{
TimeSpan interval = step ?? TimeSpan.FromSeconds(1);
DateTime baseTime = DateTime.UtcNow;
for (int i = 0; i < source.Length; i++)
{
Update(new TValue(baseTime + interval * i, source[i]));
}
}
[MethodImpl(MethodImplOptions.AggressiveInlining)]
public override TValue Update(TValue input, bool isNew = true)
{
if (isNew)
{
_p_state = _state;
}
else
{
_state = _p_state;
}
if (double.IsNaN(input.Value) || double.IsInfinity(input.Value))
{
// Return Last (initialized to NaN) if no valid input has been seen yet
return Last;
}
double x = input.Value;
// IIR: y = (b0*x + b1*x1 + b2*x2 - a1*y1 - a2*y2) * invA0
// Using chained FMA for precision
double y = _state.Count < 2
? x
: Math.FusedMultiplyAdd(-_a2, _state.Y2,
Math.FusedMultiplyAdd(-_a1, _state.Y1,
Math.FusedMultiplyAdd(_b2, _state.X2,
Math.FusedMultiplyAdd(_b1, _state.X1, _b0 * x)))) * _invA0;
// Update state
_state.X2 = _state.X1;
_state.X1 = x;
_state.Y2 = _state.Y1;
_state.Y1 = y;
if (_state.Count < 2)
{
_state.Count++;
}
var tValue = new TValue(input.Time, y);
Last = tValue;
PubEvent(tValue, isNew);
return tValue;
}
public override TSeries Update(TSeries source)
{
var result = new TSeries();
Span<double> output = new double[source.Count];
Calculate(source.Values, output, _period, double.NaN);
for (int i = 0; i < source.Count; i++)
{
result.Add(new TValue(source[i].Time, output[i]));
}
// Restore state
Reset();
// Replay a reasonable amount (e.g. 4*period) for convergence of IIR state.
int replayCount = Math.Min(source.Count, 4 * _period);
int start = source.Count - replayCount;
for (int i = start; i < source.Count; i++)
{
Update(source[i]);
}
return result;
}
public static void Calculate(ReadOnlySpan<double> source, Span<double> destination, int period, double initialLast)
{
if (period < 2)
{
throw new ArgumentOutOfRangeException(nameof(period), "Period must be greater than or equal to 2.");
}
if (destination.Length < source.Length)
{
throw new ArgumentOutOfRangeException(nameof(destination), "Destination span must have length >= source length.");
}
ComputeCoefficients(period, out double a1, out double a2, out double b0, out double b1, out double b2, out double invA0);
double x1 = 0, x2 = 0;
double y1 = 0, y2 = 0;
for (int i = 0; i < source.Length; i++)
{
double x = source[i];
if (double.IsNaN(x) || double.IsInfinity(x))
{
destination[i] = i > 0 ? destination[i - 1] : initialLast;
continue;
}
// IIR: y = (b0*x + b1*x1 + b2*x2 - a1*y1 - a2*y2) * invA0
// Using chained FMA for precision
double y = i < 2
? x
: Math.FusedMultiplyAdd(-a2, y2,
Math.FusedMultiplyAdd(-a1, y1,
Math.FusedMultiplyAdd(b2, x2,
Math.FusedMultiplyAdd(b1, x1, b0 * x)))) * invA0;
x2 = x1;
x1 = x;
y2 = y1;
y1 = y;
destination[i] = y;
}
}
}