mirror of
https://github.com/mihakralj/QuanTAlib.git
synced 2026-08-25 05:48:06 +00:00
Enhance validation tests for various indicators with external library comparisons
- Added detailed comments explaining the validation limitations for MMA and ZLEMA due to differences in algorithm implementations. - Implemented validation tests for True Range against TALib and Tulip, ensuring directional agreement. - Updated Ulcer Index validation to clarify differences in algorithmic approaches between QuanTAlib and Skender. - Enhanced Ease of Movement tests to verify directional agreement with Tulip's EMV, noting differences in volume scaling. - Expanded Klinger Volume Oscillator tests to validate against Skender and Tulip, focusing on directional agreement across multiple period configurations. - Improved Negative Volume Index tests to compare percentage changes with Tulip, addressing differences in starting values. - Updated Positive Volume Index tests to validate against Tulip, emphasizing percentage change comparisons. - Enhanced Williams Accumulation/Distribution tests to verify directional agreement with Tulip, highlighting formula differences.
This commit is contained in:
Miha Kralj
@@ -1,272 +1,259 @@
|
||||
using Skender.Stock.Indicators;
|
||||
using Xunit;
|
||||
using Xunit.Abstractions;
|
||||
|
||||
namespace QuanTAlib.Tests;
|
||||
|
||||
/// <summary>
|
||||
/// Validation tests for CMO against external libraries.
|
||||
/// Validation tests for CMO (Chande Momentum Oscillator) against external libraries.
|
||||
/// CMO = 100 × (SumUp - SumDown) / (SumUp + SumDown)
|
||||
///
|
||||
/// Note: TALib CMO uses Wilder's exponential smoothing internally, which produces
|
||||
/// fundamentally different results than the standard simple-sum CMO formula.
|
||||
/// QuanTAlib, Tulip, and Skender all use the standard simple-sum approach.
|
||||
/// </summary>
|
||||
public class CmoValidationTests
|
||||
public sealed class CmoValidationTests(ITestOutputHelper output) : IDisposable
|
||||
{
|
||||
private const double Epsilon = 1e-9;
|
||||
private readonly ValidationTestData _testData = new();
|
||||
private readonly ITestOutputHelper _output = output;
|
||||
private bool _disposed;
|
||||
|
||||
// ═══════════════════════════════════════════════════════════════════════════
|
||||
// Tulip Indicators Validation
|
||||
// ═══════════════════════════════════════════════════════════════════════════
|
||||
private const int TestPeriod = 14;
|
||||
|
||||
public void Dispose()
|
||||
{
|
||||
Dispose(disposing: true);
|
||||
}
|
||||
|
||||
private void Dispose(bool disposing)
|
||||
{
|
||||
if (_disposed) { return; }
|
||||
_disposed = true;
|
||||
if (disposing) { _testData?.Dispose(); }
|
||||
}
|
||||
|
||||
#region Tulip Validation
|
||||
|
||||
[Fact]
|
||||
public void Cmo_MatchesTulip_StandardData()
|
||||
public void Cmo_MatchesTulip_Batch()
|
||||
{
|
||||
// Generate test data
|
||||
double[] prices = new double[50];
|
||||
for (int i = 0; i < prices.Length; i++)
|
||||
{
|
||||
prices[i] = 100 + Math.Sin(i * 0.3) * 10 + i * 0.1;
|
||||
}
|
||||
double[] tData = _testData.RawData.ToArray();
|
||||
|
||||
int period = 14;
|
||||
double[] qOutput = new double[tData.Length];
|
||||
Cmo.Batch(tData.AsSpan(), qOutput.AsSpan(), TestPeriod);
|
||||
|
||||
// Calculate using Tulip
|
||||
// Tulip cmo
|
||||
var cmoIndicator = Tulip.Indicators.cmo;
|
||||
double[][] inputs = [prices];
|
||||
double[] options = [period];
|
||||
double[][] inputs = [tData];
|
||||
double[] options = [TestPeriod];
|
||||
int lookback = cmoIndicator.Start(options);
|
||||
double[][] outputs = [new double[prices.Length - lookback]];
|
||||
double[][] outputs = [new double[tData.Length - lookback]];
|
||||
|
||||
cmoIndicator.Run(inputs, options, outputs);
|
||||
double[] tulipOutput = outputs[0];
|
||||
double[] tulipResult = outputs[0];
|
||||
|
||||
// Calculate using our CMO
|
||||
double[] ourOutput = new double[prices.Length];
|
||||
Cmo.Batch(prices, ourOutput, period);
|
||||
ValidationHelper.VerifyData(qOutput, tulipResult, lookback);
|
||||
|
||||
// Compare results - Tulip outputs from index 0 corresponding to our index period
|
||||
for (int i = 0; i < tulipOutput.Length; i++)
|
||||
{
|
||||
Assert.Equal(tulipOutput[i], ourOutput[i + lookback], Epsilon);
|
||||
}
|
||||
_output.WriteLine("CMO Batch validated successfully against Tulip");
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public void Cmo_MatchesTulip_UpwardTrend()
|
||||
public void Cmo_MatchesTulip_Streaming()
|
||||
{
|
||||
// Steadily increasing prices
|
||||
double[] prices = new double[30];
|
||||
for (int i = 0; i < prices.Length; i++)
|
||||
double[] tData = _testData.RawData.ToArray();
|
||||
|
||||
// QuanTAlib CMO (streaming)
|
||||
var cmo = new Cmo(TestPeriod);
|
||||
var qResults = new List<double>();
|
||||
foreach (var item in _testData.Data)
|
||||
{
|
||||
prices[i] = 100 + i * 2;
|
||||
qResults.Add(cmo.Update(item).Value);
|
||||
}
|
||||
|
||||
int period = 10;
|
||||
// Tulip cmo
|
||||
var cmoIndicator = Tulip.Indicators.cmo;
|
||||
double[][] inputs = [tData];
|
||||
double[] options = [TestPeriod];
|
||||
int lookback = cmoIndicator.Start(options);
|
||||
double[][] outputs = [new double[tData.Length - lookback]];
|
||||
|
||||
cmoIndicator.Run(inputs, options, outputs);
|
||||
double[] tulipResult = outputs[0];
|
||||
|
||||
ValidationHelper.VerifyData(qResults, tulipResult, lookback);
|
||||
|
||||
_output.WriteLine("CMO Streaming validated successfully against Tulip");
|
||||
}
|
||||
|
||||
[Theory]
|
||||
[InlineData(5)]
|
||||
[InlineData(10)]
|
||||
[InlineData(20)]
|
||||
[InlineData(30)]
|
||||
public void Cmo_MatchesTulip_DifferentPeriods(int period)
|
||||
{
|
||||
double[] tData = _testData.RawData.ToArray();
|
||||
|
||||
double[] qOutput = new double[tData.Length];
|
||||
Cmo.Batch(tData.AsSpan(), qOutput.AsSpan(), period);
|
||||
|
||||
var cmoIndicator = Tulip.Indicators.cmo;
|
||||
double[][] inputs = [prices];
|
||||
double[][] inputs = [tData];
|
||||
double[] options = [period];
|
||||
int lookback = cmoIndicator.Start(options);
|
||||
double[][] outputs = [new double[prices.Length - lookback]];
|
||||
double[][] outputs = [new double[tData.Length - lookback]];
|
||||
|
||||
cmoIndicator.Run(inputs, options, outputs);
|
||||
double[] tulipOutput = outputs[0];
|
||||
double[] tulipResult = outputs[0];
|
||||
|
||||
double[] ourOutput = new double[prices.Length];
|
||||
Cmo.Batch(prices, ourOutput, period);
|
||||
ValidationHelper.VerifyData(qOutput, tulipResult, lookback);
|
||||
}
|
||||
|
||||
for (int i = 0; i < tulipOutput.Length; i++)
|
||||
{
|
||||
Assert.Equal(tulipOutput[i], ourOutput[i + lookback], Epsilon);
|
||||
}
|
||||
#endregion
|
||||
|
||||
#region Skender Validation
|
||||
|
||||
[Fact]
|
||||
public void Cmo_MatchesSkender_Batch()
|
||||
{
|
||||
// QuanTAlib CMO (batch)
|
||||
var qResult = Cmo.Batch(_testData.Data, TestPeriod);
|
||||
|
||||
// Skender CMO
|
||||
var sResult = _testData.SkenderQuotes.GetCmo(TestPeriod).ToList();
|
||||
|
||||
// Compare last 100 records
|
||||
ValidationHelper.VerifyData(qResult, sResult, (s) => s.Cmo);
|
||||
|
||||
_output.WriteLine("CMO Batch validated successfully against Skender");
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public void Cmo_MatchesTulip_DownwardTrend()
|
||||
public void Cmo_MatchesSkender_Streaming()
|
||||
{
|
||||
// Steadily decreasing prices
|
||||
double[] prices = new double[30];
|
||||
for (int i = 0; i < prices.Length; i++)
|
||||
// QuanTAlib CMO (streaming)
|
||||
var cmo = new Cmo(TestPeriod);
|
||||
var qResults = new List<double>();
|
||||
foreach (var item in _testData.Data)
|
||||
{
|
||||
prices[i] = 200 - i * 2;
|
||||
qResults.Add(cmo.Update(item).Value);
|
||||
}
|
||||
|
||||
int period = 10;
|
||||
// Skender CMO
|
||||
var sResult = _testData.SkenderQuotes.GetCmo(TestPeriod).ToList();
|
||||
|
||||
var cmoIndicator = Tulip.Indicators.cmo;
|
||||
double[][] inputs = [prices];
|
||||
double[] options = [period];
|
||||
int lookback = cmoIndicator.Start(options);
|
||||
double[][] outputs = [new double[prices.Length - lookback]];
|
||||
cmoIndicator.Run(inputs, options, outputs);
|
||||
double[] tulipOutput = outputs[0];
|
||||
int count = qResults.Count;
|
||||
int start = Math.Max(0, count - ValidationHelper.DefaultVerificationCount);
|
||||
|
||||
double[] ourOutput = new double[prices.Length];
|
||||
Cmo.Batch(prices, ourOutput, period);
|
||||
|
||||
for (int i = 0; i < tulipOutput.Length; i++)
|
||||
for (int i = start; i < count; i++)
|
||||
{
|
||||
Assert.Equal(tulipOutput[i], ourOutput[i + lookback], Epsilon);
|
||||
if (sResult[i].Cmo is null) { continue; }
|
||||
Assert.True(
|
||||
Math.Abs(qResults[i] - sResult[i].Cmo!.Value) <= ValidationHelper.SkenderTolerance,
|
||||
$"Mismatch at index {i}: QuanTAlib={qResults[i]:G17}, Skender={sResult[i].Cmo:G17}");
|
||||
}
|
||||
|
||||
_output.WriteLine("CMO Streaming validated successfully against Skender");
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public void Cmo_MatchesTulip_MultiplePeriods()
|
||||
[Theory]
|
||||
[InlineData(5)]
|
||||
[InlineData(10)]
|
||||
[InlineData(20)]
|
||||
[InlineData(30)]
|
||||
public void Cmo_MatchesSkender_DifferentPeriods(int period)
|
||||
{
|
||||
double[] prices = new double[100];
|
||||
var random = new Random(42);
|
||||
for (int i = 0; i < prices.Length; i++)
|
||||
{
|
||||
prices[i] = 100 + (random.NextDouble() - 0.5) * 20 + i * 0.05;
|
||||
}
|
||||
var qResult = Cmo.Batch(_testData.Data, period);
|
||||
|
||||
int[] periods = [5, 10, 14, 20, 30];
|
||||
var sResult = _testData.SkenderQuotes.GetCmo(period).ToList();
|
||||
|
||||
foreach (int period in periods)
|
||||
{
|
||||
var cmoIndicator = Tulip.Indicators.cmo;
|
||||
double[][] inputs = [prices];
|
||||
double[] options = [period];
|
||||
int lookback = cmoIndicator.Start(options);
|
||||
double[][] outputs = [new double[prices.Length - lookback]];
|
||||
cmoIndicator.Run(inputs, options, outputs);
|
||||
double[] tulipOutput = outputs[0];
|
||||
|
||||
double[] ourOutput = new double[prices.Length];
|
||||
Cmo.Batch(prices, ourOutput, period);
|
||||
|
||||
for (int i = 0; i < tulipOutput.Length; i++)
|
||||
{
|
||||
Assert.Equal(tulipOutput[i], ourOutput[i + lookback], Epsilon);
|
||||
}
|
||||
}
|
||||
ValidationHelper.VerifyData(qResult, sResult, (s) => s.Cmo);
|
||||
}
|
||||
|
||||
// ═══════════════════════════════════════════════════════════════════════════
|
||||
// Manual Calculation Validation
|
||||
// ═══════════════════════════════════════════════════════════════════════════
|
||||
#endregion
|
||||
|
||||
#region Mathematical Validation
|
||||
|
||||
[Fact]
|
||||
public void Cmo_ManualCalculation_AllUpMoves()
|
||||
public void Cmo_AllUpMoves_Returns100()
|
||||
{
|
||||
// All upward moves
|
||||
double[] prices = [100, 101, 102, 103, 104, 105];
|
||||
int period = 5;
|
||||
|
||||
double[] output = new double[prices.Length];
|
||||
Cmo.Batch(prices, output, period);
|
||||
double[] result = new double[prices.Length];
|
||||
Cmo.Batch(prices, result, period);
|
||||
|
||||
// After 5 periods: SumUp = 5, SumDown = 0
|
||||
// CMO = 100 * (5-0)/(5+0) = 100
|
||||
Assert.Equal(100.0, output[5], Epsilon);
|
||||
// After 5 periods: SumUp = 5, SumDown = 0 → CMO = 100
|
||||
Assert.Equal(100.0, result[5], 1e-9);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public void Cmo_ManualCalculation_AllDownMoves()
|
||||
public void Cmo_AllDownMoves_ReturnsNegative100()
|
||||
{
|
||||
// All downward moves
|
||||
double[] prices = [105, 104, 103, 102, 101, 100];
|
||||
int period = 5;
|
||||
|
||||
double[] output = new double[prices.Length];
|
||||
Cmo.Batch(prices, output, period);
|
||||
double[] result = new double[prices.Length];
|
||||
Cmo.Batch(prices, result, period);
|
||||
|
||||
// After 5 periods: SumUp = 0, SumDown = 5
|
||||
// CMO = 100 * (0-5)/(0+5) = -100
|
||||
Assert.Equal(-100.0, output[5], Epsilon);
|
||||
// After 5 periods: SumUp = 0, SumDown = 5 → CMO = -100
|
||||
Assert.Equal(-100.0, result[5], 1e-9);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public void Cmo_ManualCalculation_EqualMoves()
|
||||
public void Cmo_EqualMoves_ReturnsZero()
|
||||
{
|
||||
// Equal up and down moves
|
||||
double[] prices = [100, 102, 100, 102, 100]; // up 2, down 2, up 2, down 2
|
||||
int period = 4;
|
||||
|
||||
double[] output = new double[prices.Length];
|
||||
Cmo.Batch(prices, output, period);
|
||||
double[] result = new double[prices.Length];
|
||||
Cmo.Batch(prices, result, period);
|
||||
|
||||
// SumUp = 4, SumDown = 4
|
||||
// CMO = 100 * (4-4)/(4+4) = 0
|
||||
Assert.Equal(0.0, output[4], Epsilon);
|
||||
}
|
||||
|
||||
// ═══════════════════════════════════════════════════════════════════════════
|
||||
// Streaming vs Batch Validation
|
||||
// ═══════════════════════════════════════════════════════════════════════════
|
||||
|
||||
[Fact]
|
||||
public void Cmo_StreamingMatchesBatch()
|
||||
{
|
||||
double[] prices = new double[100];
|
||||
var random = new Random(12345);
|
||||
for (int i = 0; i < prices.Length; i++)
|
||||
{
|
||||
prices[i] = 100 + (random.NextDouble() - 0.5) * 30 + Math.Sin(i * 0.2) * 5;
|
||||
}
|
||||
|
||||
int period = 14;
|
||||
|
||||
// Batch calculation
|
||||
double[] batchOutput = new double[prices.Length];
|
||||
Cmo.Batch(prices, batchOutput, period);
|
||||
|
||||
// Streaming calculation
|
||||
var cmo = new Cmo(period);
|
||||
for (int i = 0; i < prices.Length; i++)
|
||||
{
|
||||
var result = cmo.Update(new TValue(DateTime.Now.Ticks + i, prices[i]));
|
||||
Assert.Equal(batchOutput[i], result.Value, Epsilon);
|
||||
}
|
||||
}
|
||||
|
||||
// ═══════════════════════════════════════════════════════════════════════════
|
||||
// Edge Case Validation
|
||||
// ═══════════════════════════════════════════════════════════════════════════
|
||||
|
||||
[Fact]
|
||||
public void Cmo_NoChange_ReturnsZero()
|
||||
{
|
||||
double[] prices = [100, 100, 100, 100, 100, 100];
|
||||
int period = 5;
|
||||
|
||||
double[] output = new double[prices.Length];
|
||||
Cmo.Batch(prices, output, period);
|
||||
|
||||
// No movement = 0
|
||||
Assert.Equal(0.0, output[5]);
|
||||
// SumUp = 4, SumDown = 4 → CMO = 0
|
||||
Assert.Equal(0.0, result[4], 1e-9);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public void Cmo_RangeIsBounded()
|
||||
{
|
||||
double[] prices = new double[100];
|
||||
var random = new Random(54321);
|
||||
for (int i = 0; i < prices.Length; i++)
|
||||
{
|
||||
prices[i] = 100 + (random.NextDouble() - 0.5) * 50;
|
||||
}
|
||||
double[] tData = _testData.RawData.ToArray();
|
||||
|
||||
double[] output = new double[prices.Length];
|
||||
Cmo.Batch(prices, output, 14);
|
||||
double[] result = new double[tData.Length];
|
||||
Cmo.Batch(tData.AsSpan(), result.AsSpan(), TestPeriod);
|
||||
|
||||
// All values should be in [-100, 100] range
|
||||
for (int i = 14; i < output.Length; i++)
|
||||
// All values after warmup should be in [-100, 100]
|
||||
for (int i = TestPeriod; i < result.Length; i++)
|
||||
{
|
||||
Assert.True(output[i] >= -100.0 && output[i] <= 100.0,
|
||||
$"CMO at index {i} = {output[i]} is out of range [-100, 100]");
|
||||
Assert.True(result[i] >= -100.0 && result[i] <= 100.0,
|
||||
$"CMO at index {i} = {result[i]} is out of range [-100, 100]");
|
||||
}
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public void Cmo_AlternatingMoves_ConvergesToZero()
|
||||
public void Batch_MatchesStreaming_IdenticalResults()
|
||||
{
|
||||
// Alternating pattern with equal magnitude
|
||||
double[] prices = new double[50];
|
||||
for (int i = 0; i < prices.Length; i++)
|
||||
double[] tData = _testData.RawData.ToArray();
|
||||
|
||||
// Batch
|
||||
double[] batchOutput = new double[tData.Length];
|
||||
Cmo.Batch(tData.AsSpan(), batchOutput.AsSpan(), TestPeriod);
|
||||
|
||||
// Streaming
|
||||
var cmo = new Cmo(TestPeriod);
|
||||
var streamingResults = new double[tData.Length];
|
||||
for (int i = 0; i < tData.Length; i++)
|
||||
{
|
||||
prices[i] = 100 + (i % 2 == 0 ? 0 : 2); // 100, 102, 100, 102, ...
|
||||
streamingResults[i] = cmo.Update(new TValue(DateTime.UtcNow.Ticks + i, tData[i])).Value;
|
||||
}
|
||||
|
||||
double[] output = new double[prices.Length];
|
||||
Cmo.Batch(prices, output, 10);
|
||||
|
||||
// Result should be close to 0 for balanced oscillation
|
||||
Assert.True(Math.Abs(output[^1]) < 20,
|
||||
$"CMO for alternating pattern should be near zero, got {output[^1]}");
|
||||
int count = tData.Length;
|
||||
int start = Math.Max(0, count - ValidationHelper.DefaultVerificationCount);
|
||||
for (int i = start; i < count; i++)
|
||||
{
|
||||
Assert.Equal(batchOutput[i], streamingResults[i], 1e-9);
|
||||
}
|
||||
_output.WriteLine("CMO Batch vs Streaming consistency validated");
|
||||
}
|
||||
|
||||
#endregion
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user