//! Unit Tests for Microstructure Features (Roll Measure & Amihud Illiquidity)
//!
//! TDD Implementation: Tests written FIRST, then implementation
//!
//! ## Test Coverage
//! - Roll Measure: Serial correlation, zero covariance, negative handling
//! - Amihud Illiquidity: Normal case, high volume, zero volume
//! - Performance: <5μs latency, 72 bytes memory per symbol
//! - Integration: 256-feature pipeline compatibility

use ml::features::microstructure::{AmihudIlliquidity, RollMeasure};

// ============================================================================
// Roll Measure Tests (Agent A9)
// ============================================================================

#[test]
fn test_roll_measure_positive_serial_correlation() {
    // Roll spread = 2 * sqrt(-cov(Δp_t, Δp_{t-1}))
    // With positive serial correlation, cov < 0, so sqrt should work

    let mut roll = RollMeasure::new();

    // Simulate mean-reverting prices (negative serial correlation)
    let prices = vec![100.0, 101.0, 100.0, 101.0, 100.0, 101.0];

    for price in prices {
        roll.update(price);
    }

    let spread = roll.compute();

    // Should produce positive spread estimate
    assert!(spread > 0.0, "Roll spread should be positive: {}", spread);
    assert!(
        spread < 10.0,
        "Roll spread should be reasonable: {}",
        spread
    );
}

#[test]
fn test_roll_measure_negative_serial_correlation() {
    // With negative serial correlation (mean reversion), cov > 0
    // Formula: 2 * sqrt(-cov) requires taking sqrt of negative value
    // Implementation should handle this by taking sqrt(abs(cov))

    let mut roll = RollMeasure::new();

    // Simulate trending prices (positive serial correlation)
    let prices = vec![100.0, 100.5, 101.0, 101.5, 102.0, 102.5];

    for price in prices {
        roll.update(price);
    }

    let spread = roll.compute();

    // Should still produce valid spread estimate (non-negative)
    assert!(
        spread >= 0.0,
        "Roll spread should be non-negative: {}",
        spread
    );
}

#[test]
fn test_roll_measure_zero_covariance() {
    // Random walk (no serial correlation) => cov ≈ 0
    // Roll spread should be close to zero

    let mut roll = RollMeasure::new();

    // Simulate random walk with alternating changes
    let prices = vec![100.0, 100.1, 100.0, 100.2, 100.1, 100.3];

    for price in prices {
        roll.update(price);
    }

    let spread = roll.compute();

    // Should be small (close to zero)
    assert!(spread >= 0.0, "Roll spread should be non-negative");
    assert!(
        spread < 1.0,
        "Roll spread should be small for random walk: {}",
        spread
    );
}

#[test]
fn test_roll_measure_insufficient_data() {
    let mut roll = RollMeasure::new();

    // Need at least 2 price changes (3 prices) for covariance
    roll.update(100.0);
    roll.update(101.0);

    let spread = roll.compute();

    // Should return 0.0 or handle gracefully
    assert!(
        spread >= 0.0,
        "Roll spread should be non-negative with insufficient data"
    );
}

#[test]
fn test_roll_measure_latency_requirement() {
    use std::time::Instant;

    let mut roll = RollMeasure::new();

    // Warm up with 20 prices
    for i in 0..20 {
        roll.update(100.0 + (i as f64) * 0.1);
    }

    // Measure update + compute latency
    let start = Instant::now();
    for _ in 0..100 {
        roll.update(105.0);
        let _ = roll.compute();
    }
    let elapsed = start.elapsed();

    let avg_latency_us = elapsed.as_micros() / 100;

    // Requirement: <5μs per update+compute
    assert!(
        avg_latency_us < 5,
        "Roll measure latency {}μs exceeds 5μs requirement",
        avg_latency_us
    );
}

#[test]
fn test_roll_measure_memory_footprint() {
    use std::mem::size_of;

    let roll = RollMeasure::new();
    let size = size_of::<RollMeasure>();

    // Requirement: 72 bytes per symbol
    assert!(
        size <= 72,
        "Roll measure memory {}B exceeds 72B requirement",
        size
    );
}

#[test]
fn test_roll_measure_real_market_data() {
    // Test with ES.FUT-like price movements
    let mut roll = RollMeasure::new();

    let prices = vec![
        4500.25, 4500.50, 4500.25, 4500.75, 4500.50, 4500.25, 4501.00, 4500.75, 4500.50, 4501.25,
    ];

    for price in prices {
        roll.update(price);
    }

    let spread = roll.compute();

    // Typical bid-ask spread for ES futures: 0.25-1.0 points
    assert!(spread >= 0.0, "Roll spread should be non-negative");
    assert!(
        spread < 5.0,
        "Roll spread should be realistic for ES.FUT: {}",
        spread
    );
}

#[test]
fn test_roll_measure_extreme_volatility() {
    let mut roll = RollMeasure::new();

    // Simulate flash crash scenario
    let prices = vec![100.0, 100.5, 101.0, 95.0, 90.0, 92.0, 95.0, 98.0, 100.0];

    for price in prices {
        roll.update(price);
    }

    let spread = roll.compute();

    // Should handle extreme volatility without panicking
    assert!(spread.is_finite(), "Roll spread should be finite");
    assert!(spread >= 0.0, "Roll spread should be non-negative");
}

// ============================================================================
// Amihud Illiquidity Tests (Agent A8)
// ============================================================================

#[test]
fn test_amihud_normal_case() {
    // Amihud = |return| / dollar_volume

    let mut amihud = AmihudIlliquidity::new(0.05);

    amihud.update(100.0, 1_000_000.0); // price, volume
    amihud.update(101.0, 1_000_000.0);

    let illiquidity = amihud.compute();

    // Expected: abs(log(101/100)) / 1_000_000 ≈ 0.00995 / 1M ≈ 1e-8
    assert!(illiquidity > 0.0, "Amihud should be positive");
    assert!(
        illiquidity < 1e-5,
        "Amihud should be small for liquid market: {}",
        illiquidity
    );
}

#[test]
fn test_amihud_high_volume_low_illiquidity() {
    let mut amihud = AmihudIlliquidity::new(0.05);

    // High volume => low illiquidity
    amihud.update(100.0, 10_000_000.0);
    amihud.update(101.0, 10_000_000.0);

    let high_vol_illiquidity = amihud.compute();

    // Compare with low volume
    let mut amihud2 = AmihudIlliquidity::new(0.05);
    amihud2.update(100.0, 1_000_000.0);
    amihud2.update(101.0, 1_000_000.0);

    let low_vol_illiquidity = amihud2.compute();

    assert!(
        high_vol_illiquidity < low_vol_illiquidity,
        "High volume should have lower illiquidity"
    );
}

#[test]
fn test_amihud_zero_volume() {
    let mut amihud = AmihudIlliquidity::new(0.05);

    // Zero volume should be handled gracefully
    amihud.update(100.0, 0.0);
    amihud.update(101.0, 0.0);

    let illiquidity = amihud.compute();

    // Should return max illiquidity or capped value
    assert!(illiquidity.is_finite(), "Amihud should handle zero volume");
}

#[test]
fn test_amihud_latency_requirement() {
    use std::time::Instant;

    let mut amihud = AmihudIlliquidity::new(0.05);

    // Warm up
    for i in 0..20 {
        amihud.update(100.0 + (i as f64) * 0.1, 1_000_000.0);
    }

    // Measure latency
    let start = Instant::now();
    for _ in 0..100 {
        amihud.update(105.0, 1_000_000.0);
        let _ = amihud.compute();
    }
    let elapsed = start.elapsed();

    let avg_latency_us = elapsed.as_micros() / 100;

    // Requirement: <5μs
    assert!(
        avg_latency_us < 5,
        "Amihud latency {}μs exceeds 5μs requirement",
        avg_latency_us
    );
}

#[test]
fn test_amihud_memory_footprint() {
    use std::mem::size_of;

    let amihud = AmihudIlliquidity::new(0.05);
    let size = size_of::<AmihudIlliquidity>();

    // Requirement: 72 bytes per symbol
    assert!(
        size <= 72,
        "Amihud memory {}B exceeds 72B requirement",
        size
    );
}

// ============================================================================
// Integration Tests
// ============================================================================

#[test]
fn test_microstructure_integration_256_features() {
    // Verify microstructure features fit within 256-dim feature vector
    // Features 115-164 are allocated for microstructure (50 features)

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

    let bars: Vec<OHLCVBar> = (0..100)
        .map(|i| OHLCVBar {
            timestamp: Utc::now() + chrono::Duration::hours(i),
            open: 100.0 + (i as f64) * 0.1,
            high: 101.0 + (i as f64) * 0.1,
            low: 99.0 + (i as f64) * 0.1,
            close: 100.5 + (i as f64) * 0.1,
            volume: 1_000_000.0 + (i as f64) * 10_000.0,
        })
        .collect();

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

    // Should extract 225-dim features
    assert_eq!(features.len(), 50); // 100 bars - 50 warmup
    assert_eq!(features[0].len(), 225);

    // Verify all features are finite
    for feature_vec in &features {
        for (i, &val) in feature_vec.iter().enumerate() {
            assert!(val.is_finite(), "Feature {} is not finite: {}", i, val);
        }
    }
}

#[test]
fn test_microstructure_features_non_negative() {
    // Roll and Amihud should produce non-negative values

    let mut roll = RollMeasure::new();
    let mut amihud = AmihudIlliquidity::new(0.05);

    // Feed price/volume data
    for i in 0..20 {
        let price = 100.0 + (i as f64) * 0.1;
        let volume = 1_000_000.0 + (i as f64) * 10_000.0;

        roll.update(price);
        amihud.update(price, volume);
    }

    let roll_spread = roll.compute();
    let amihud_illiq = amihud.compute();

    assert!(roll_spread >= 0.0, "Roll spread should be non-negative");
    assert!(
        amihud_illiq >= 0.0,
        "Amihud illiquidity should be non-negative"
    );
}

#[test]
fn test_microstructure_features_normalization() {
    // Features should be normalized for ML training

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

    let bars: Vec<OHLCVBar> = (0..100)
        .map(|i| OHLCVBar {
            timestamp: Utc::now() + chrono::Duration::hours(i),
            open: 100.0,
            high: 101.0,
            low: 99.0,
            close: 100.5,
            volume: 1_000_000.0,
        })
        .collect();

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

    // Microstructure features (115-164) should be normalized
    for feature_vec in &features {
        for i in 115..165 {
            let val = feature_vec[i];

            // Check if normalized (0-1 range or standardized)
            // Most features should be in reasonable range
            assert!(
                val.abs() < 10.0,
                "Feature {} has unreasonable value: {}",
                i,
                val
            );
        }
    }
}