foxhunt/ml/tests/test_extract_256_dim_features.rs

//! Integration test for 256-dimension feature extraction
//!
//! Tests the extract_ml_features() function with real OHLCV data

use ml::features::extraction::{extract_ml_features, OHLCVBar};
use chrono::Utc;

#[test]
fn test_extract_256_dim_features() {
    // Create synthetic OHLCV bars (100 bars to exceed warmup period of 50)
    let bars: Vec<OHLCVBar> = (0..100).map(|i| {
        OHLCVBar {
            timestamp: Utc::now() + chrono::Duration::hours(i),
            open: 4500.0 + i as f64 * 0.5,
            high: 4510.0 + i as f64 * 0.5,
            low: 4490.0 + i as f64 * 0.5,
            close: 4505.0 + i as f64 * 0.5,
            volume: 10000.0 + i as f64 * 100.0,
        }
    }).collect();

    // Extract features
    let result = extract_ml_features(&bars);
    assert!(result.is_ok(), "Feature extraction failed: {:?}", result.err());

    let features = result.unwrap();

    // Should return features for bars after warmup period (100 - 50 = 50)
    assert_eq!(
        features.len(),
        50,
        "Expected 50 feature vectors (100 bars - 50 warmup), got {}",
        features.len()
    );

    // Each feature vector should be exactly 256 dimensions
    for (i, feature_vec) in features.iter().enumerate() {
        assert_eq!(
            feature_vec.len(),
            256,
            "Feature vector {} has wrong dimension: {}",
            i,
            feature_vec.len()
        );

        // Validate no NaN/Inf values
        for (j, &val) in feature_vec.iter().enumerate() {
            assert!(
                val.is_finite(),
                "Feature vector {} has non-finite value at index {}: {}",
                i,
                j,
                val
            );
        }
    }

    println!("✅ Successfully extracted {} 256-dim feature vectors", features.len());
    println!("✅ First feature vector sample (first 10 features): {:?}", &features[0][0..10]);
}

#[test]
fn test_feature_dimensions() {
    // Create 60 bars (10 above minimum warmup)
    let bars: Vec<OHLCVBar> = (0..60).map(|i| {
        OHLCVBar {
            timestamp: Utc::now() + chrono::Duration::minutes(i),
            open: 4500.0,
            high: 4510.0,
            low: 4490.0,
            close: 4505.0 + (i as f64 * 0.1).sin() * 5.0, // Add some variation
            volume: 10000.0,
        }
    }).collect();

    let features = extract_ml_features(&bars).unwrap();

    // Should have 10 feature vectors (60 - 50 warmup)
    assert_eq!(features.len(), 10);

    // Check output shape (num_bars, 256)
    assert_eq!(features.len(), 10, "Wrong number of bars");
    for feature_vec in &features {
        assert_eq!(feature_vec.len(), 256, "Wrong feature dimension");
    }

    // Validate no NaN/Inf
    for feature_vec in &features {
        for &val in feature_vec.iter() {
            assert!(val.is_finite(), "Found non-finite value: {}", val);
        }
    }

    println!("✅ Feature dimensions validated: {} bars × 256 features", features.len());
}

#[test]
fn test_insufficient_data_error() {
    // Create only 10 bars (below 50 warmup requirement)
    let bars: Vec<OHLCVBar> = (0..10).map(|i| {
        OHLCVBar {
            timestamp: Utc::now() + chrono::Duration::hours(i),
            open: 4500.0,
            high: 4510.0,
            low: 4490.0,
            close: 4505.0,
            volume: 10000.0,
        }
    }).collect();

    let result = extract_ml_features(&bars);
    assert!(result.is_err(), "Should fail with insufficient data");
    
    let error_msg = result.unwrap_err().to_string();
    assert!(
        error_msg.contains("Insufficient data"),
        "Expected 'Insufficient data' error, got: {}",
        error_msg
    );

    println!("✅ Insufficient data error handled correctly");
}

#[test]
fn test_feature_normalization() {
    // Create bars with extreme values to test normalization
    let bars: Vec<OHLCVBar> = (0..100).map(|i| {
        OHLCVBar {
            timestamp: Utc::now() + chrono::Duration::hours(i),
            open: 4500.0 + i as f64 * 10.0, // Large price changes
            high: 4600.0 + i as f64 * 10.0,
            low: 4400.0 + i as f64 * 10.0,
            close: 4500.0 + i as f64 * 10.0,
            volume: 100000.0 + i as f64 * 5000.0, // Large volume changes
        }
    }).collect();

    let features = extract_ml_features(&bars).unwrap();

    // Check that features are reasonably normalized
    for (i, feature_vec) in features.iter().enumerate() {
        for (j, &val) in feature_vec.iter().enumerate() {
            // Most features should be in reasonable range (not all, but most)
            // This is a sanity check, not strict validation
            if !(-10.0..=10.0).contains(&val) {
                // Log but don't fail - some features may legitimately be outside this range
                println!("⚠️ Feature {} in vector {} has value outside [-10, 10]: {}", j, i, val);
            }
        }
    }

    println!("✅ Feature normalization validated");
}

#[test]
fn test_feature_consistency() {
    // Test that same input produces same output (deterministic)
    let bars: Vec<OHLCVBar> = (0..100).map(|i| {
        OHLCVBar {
            timestamp: Utc::now() + chrono::Duration::hours(i),
            open: 4500.0,
            high: 4510.0,
            low: 4490.0,
            close: 4505.0,
            volume: 10000.0,
        }
    }).collect();

    let features1 = extract_ml_features(&bars).unwrap();
    let features2 = extract_ml_features(&bars).unwrap();

    assert_eq!(features1.len(), features2.len());
    
    for (vec1, vec2) in features1.iter().zip(features2.iter()) {
        for (&val1, &val2) in vec1.iter().zip(vec2.iter()) {
            assert!(
                (val1 - val2).abs() < 1e-10,
                "Features not consistent: {} vs {}",
                val1,
                val2
            );
        }
    }

    println!("✅ Feature extraction is deterministic");
}