foxhunt/crates/ml/tests/gpu_backtest_validation.rs

#![allow(
    clippy::assertions_on_constants,
    clippy::assertions_on_result_states,
    clippy::clone_on_copy,
    clippy::decimal_literal_representation,
    clippy::doc_markdown,
    clippy::empty_line_after_doc_comments,
    clippy::field_reassign_with_default,
    clippy::get_unwrap,
    clippy::identity_op,
    clippy::inconsistent_digit_grouping,
    clippy::indexing_slicing,
    clippy::integer_division,
    clippy::len_zero,
    clippy::let_underscore_must_use,
    clippy::manual_div_ceil,
    clippy::manual_let_else,
    clippy::manual_range_contains,
    clippy::modulo_arithmetic,
    clippy::needless_range_loop,
    clippy::non_ascii_literal,
    clippy::redundant_clone,
    clippy::shadow_reuse,
    clippy::shadow_same,
    clippy::shadow_unrelated,
    clippy::single_match_else,
    clippy::str_to_string,
    clippy::string_slice,
    clippy::tests_outside_test_module,
    clippy::too_many_lines,
    clippy::unnecessary_wraps,
    clippy::unseparated_literal_suffix,
    clippy::use_debug,
    clippy::useless_vec,
    clippy::wildcard_enum_match_arm,
    clippy::else_if_without_else,
    clippy::expect_used,
    clippy::missing_const_for_fn,
    clippy::similar_names,
    clippy::type_complexity,
    clippy::collapsible_else_if,
    clippy::doc_lazy_continuation,
    clippy::items_after_test_module,
    clippy::map_clone,
    clippy::multiple_unsafe_ops_per_block,
    clippy::unwrap_or_default,
    clippy::assign_op_pattern,
    clippy::needless_borrow,
    clippy::println_empty_string,
    clippy::unnecessary_cast,
    clippy::used_underscore_binding,
    clippy::create_dir,
    clippy::implicit_saturating_sub,
    clippy::exit,
    clippy::expect_fun_call,
    clippy::too_many_arguments,
    clippy::unnecessary_map_or,
    clippy::unwrap_used,
    dead_code,
    unused_imports,
    unused_variables,
    clippy::cloned_ref_to_slice_refs,
    clippy::neg_multiply,
    clippy::while_let_loop,
    clippy::bool_assert_comparison,
    clippy::excessive_precision,
    clippy::trivially_copy_pass_by_ref,
    clippy::op_ref,
    clippy::redundant_closure,
    clippy::unnecessary_lazy_evaluations,
    clippy::if_then_some_else_none,
    clippy::unnecessary_to_owned,
    clippy::single_component_path_imports,
)]
//! Validates GPU backtest evaluator produces reasonable metrics
//! using synthetic data and deterministic action models.
//!
//! These tests require a CUDA GPU and are skipped gracefully when none is available.
//! Run with: `cargo test -p ml --test gpu_backtest_validation -- --ignored`

use std::sync::Arc;

/// Generate deterministic synthetic price data (random walk with drift) using LCG.
fn generate_prices(n_bars: usize, seed: u64, drift: f32) -> Vec<[f32; 4]> {
    let mut rng_state = seed;
    let mut prices = Vec::with_capacity(n_bars);
    let mut price = 100.0_f32;

    for _ in 0..n_bars {
        // Simple LCG for determinism
        rng_state = rng_state
            .wrapping_mul(6_364_136_223_846_793_005)
            .wrapping_add(1_442_695_040_888_963_407);
        let rand_f = ((rng_state >> 33) as f32) / (u32::MAX as f32) - 0.5;
        let ret = drift + rand_f * 0.02;
        price *= 1.0 + ret;
        let ohlc = [price * 0.999, price * 1.001, price * 0.998, price];
        prices.push(ohlc);
    }
    prices
}

/// Generate minimal synthetic features (just enough for the evaluator).
fn generate_features(n_bars: usize, feature_dim: usize) -> Vec<Vec<f32>> {
    (0..n_bars)
        .map(|i| {
            let mut fv = vec![0.0_f32; feature_dim];
            // Put some variation in features so they are not all zero
            if let Some(f) = fv.get_mut(0) {
                *f = (i as f32) * 0.001;
            }
            fv
        })
        .collect()
}

mod gpu_tests {
    use super::*;
    use cudarc::driver::{CudaContext, CudaSlice, CudaStream};
    use ml::cuda_pipeline::gpu_backtest_evaluator::{
        DqnBacktestConfig, GpuBacktestConfig, GpuBacktestEvaluator,
    };
    use ml::cuda_pipeline::lob_bar::LobBar;
    use ml::MLError;
    use tracing::warn;

    /// SP15 Wave 3b — build a deterministic `[Vec<LobBar>; 1]` shaped to
    /// match the single-window OHLC `prices` slice passed to
    /// `GpuBacktestEvaluator::new`. Uses the close price for `LobBar.price`
    /// and zeroes `spread`/`ofi` (these tests don't exercise cost-net
    /// sharpe — they assert raw PnL/drawdown semantics — so the LobBar
    /// streams are inert by construction).
    fn lob_bars_from_prices(prices: &[[f32; 4]]) -> Vec<LobBar> {
        prices
            .iter()
            .map(|ohlc| LobBar::new(ohlc[3], 0.0, 0.0))
            .collect()
    }

    /// Skip test gracefully if no CUDA device is available.
    /// Returns the stream (owned Arc) on success.
    fn try_cuda_stream() -> Option<Arc<CudaStream>> {
        match CudaContext::new(0) {
            Ok(ctx) => Some(ctx.default_stream()),
            Err(_) => {
                warn!("CUDA not available, skipping GPU backtest test");
                None
            }
        }
    }

    /// Build a closure that always returns action indices for the given `action`.
    ///
    /// The closure receives `(&CudaSlice<f32>, n_windows, state_dim)` and
    /// returns `CudaSlice<i32>` of length `n_windows` filled with `action`.
    fn constant_action_model(
        action: i32,
        stream: &Arc<CudaStream>,
    ) -> impl Fn(&CudaSlice<f32>, usize, usize) -> Result<CudaSlice<i32>, MLError> {
        let stream = Arc::clone(stream);
        move |_states: &CudaSlice<f32>, n_windows: usize, _state_dim: usize| {
            let host_actions = vec![action; n_windows];
            let mut gpu_actions = stream
                .alloc_zeros::<i32>(n_windows)
                .map_err(|e| MLError::ModelError(format!("constant_action_model alloc: {e}")))?;
            stream
                .memcpy_htod(&host_actions, &mut gpu_actions)
                .map_err(|e| MLError::ModelError(format!("constant_action_model HtoD: {e}")))?;
            Ok(gpu_actions)
        }
    }

    // -- Individual test cases --

    /// Always-long model on upward-trending data should produce positive PnL.
    #[test]
    fn test_always_long_on_uptrend() {
        let stream = match try_cuda_stream() {
            Some(s) => s,
            None => return,
        };

        // FEATURE_DIM=42 mirrors production eval-baseline.rs (Market 42 from
        // extract_ml_features). The gather kernel computes `market_dim =
        // feat_dim - SL_OFI_DIM (32)` so anything < 32 produces a negative
        // market_dim → kernel OOB. 42 is the canonical production value.
        const FEATURE_DIM: usize = 42;
        const N_BARS: usize = 500;

        let prices = generate_prices(N_BARS, 42, 0.001); // positive drift
        let features = generate_features(N_BARS, FEATURE_DIM);

        let config = GpuBacktestConfig {
            max_position: 1.0,
            tx_cost_bps: 0.0, // zero costs for a clean signal
            spread_cost: 0.0,
            initial_capital: 100_000.0,
            ..Default::default()
        };

        let bars = lob_bars_from_prices(&prices);
        let mut evaluator = GpuBacktestEvaluator::new(
            &[prices],
            &[features],
            &[bars],
            FEATURE_DIM,
            config,
            &stream,
        )
        .expect("evaluator creation should succeed");

        // Phase 8.5 (2026-05-12) — wire factored-action branch sizes for
        // env_step decoder. Without this, `evaluate()` bails on the
        // zero-b-size defensive guard. Match production eval-baseline.rs.
        evaluator.set_branch_sizes(&DqnBacktestConfig::from_network_dims((256, 256, 128, 128)));

        // Action 4 = Long100
        let model = constant_action_model(4, &stream);
        let metrics = evaluator
            .evaluate(&model, 24)
            .expect("evaluation should succeed");

        assert_eq!(metrics.len(), 1, "expected exactly one window result");
        let m = metrics
            .first()
            .expect("metrics vec must have at least one element");

        // With positive drift and always-long, should be profitable
        assert!(
            m.total_pnl > 0.0,
            "expected positive PnL for long on uptrend, got {}",
            m.total_pnl
        );
        assert!(
            m.max_drawdown >= 0.0,
            "drawdown should be non-negative, got {}",
            m.max_drawdown
        );
        assert!(
            (0.0..=1.0).contains(&m.win_rate),
            "win_rate {} is out of [0, 1] range",
            m.win_rate
        );
    }

    /// Always-long model on downward-trending data should produce negative PnL.
    #[test]
    fn test_always_long_on_downtrend() {
        let stream = match try_cuda_stream() {
            Some(s) => s,
            None => return,
        };

        // FEATURE_DIM=42 mirrors production eval-baseline.rs (Market 42 from
        // extract_ml_features). The gather kernel computes `market_dim =
        // feat_dim - SL_OFI_DIM (32)` so anything < 32 produces a negative
        // market_dim → kernel OOB. 42 is the canonical production value.
        const FEATURE_DIM: usize = 42;
        const N_BARS: usize = 500;

        let prices = generate_prices(N_BARS, 77, -0.001); // negative drift
        let features = generate_features(N_BARS, FEATURE_DIM);

        let config = GpuBacktestConfig {
            max_position: 1.0,
            tx_cost_bps: 0.0,
            spread_cost: 0.0,
            initial_capital: 100_000.0,
            ..Default::default()
        };

        let bars = lob_bars_from_prices(&prices);
        let mut evaluator = GpuBacktestEvaluator::new(
            &[prices],
            &[features],
            &[bars],
            FEATURE_DIM,
            config,
            &stream,
        )
        .expect("evaluator creation should succeed");

        // Phase 8.5 (2026-05-12) — wire factored-action branch sizes for
        // env_step decoder. Match production eval-baseline.rs.
        evaluator.set_branch_sizes(&DqnBacktestConfig::from_network_dims((256, 256, 128, 128)));

        let model = constant_action_model(4, &stream); // Always Long100
        let metrics = evaluator
            .evaluate(&model, 24)
            .expect("evaluation should succeed");

        assert_eq!(metrics.len(), 1);
        let m = metrics
            .first()
            .expect("metrics vec must have at least one element");

        assert!(
            m.total_pnl < 0.0,
            "expected negative PnL for long on downtrend, got {}",
            m.total_pnl
        );
        assert!(
            m.max_drawdown >= 0.0,
            "drawdown should be non-negative, got {}",
            m.max_drawdown
        );
    }

    /// Always-flat model should produce ~zero PnL and minimal trades.
    #[test]
    fn test_always_flat_produces_no_pnl() {
        let stream = match try_cuda_stream() {
            Some(s) => s,
            None => return,
        };

        // FEATURE_DIM=42 mirrors production eval-baseline.rs (Market 42 from
        // extract_ml_features). The gather kernel computes `market_dim =
        // feat_dim - SL_OFI_DIM (32)` so anything < 32 produces a negative
        // market_dim → kernel OOB. 42 is the canonical production value.
        const FEATURE_DIM: usize = 42;
        const N_BARS: usize = 200;

        let prices = generate_prices(N_BARS, 99, 0.0);
        let features = generate_features(N_BARS, FEATURE_DIM);

        let config = GpuBacktestConfig::default();

        let bars = lob_bars_from_prices(&prices);
        let mut evaluator = GpuBacktestEvaluator::new(
            &[prices],
            &[features],
            &[bars],
            FEATURE_DIM,
            config,
            &stream,
        )
        .expect("evaluator creation should succeed");

        // Phase 8.5 (2026-05-12) — wire factored-action branch sizes for
        // env_step decoder. Without this, `evaluate()` bails on the
        // zero-b-size defensive guard. Match production eval-baseline.rs.
        evaluator.set_branch_sizes(&DqnBacktestConfig::from_network_dims((256, 256, 128, 128)));

        // Action 2 = Flat -- never enters a position
        let model = constant_action_model(2, &stream);
        let metrics = evaluator
            .evaluate(&model, 24)
            .expect("evaluation should succeed");

        assert_eq!(metrics.len(), 1);
        let m = metrics
            .first()
            .expect("metrics vec must have at least one element");

        // Flat action means no position changes, so PnL should be approximately zero
        assert!(
            m.total_pnl.abs() < 0.01,
            "expected ~zero PnL for flat model, got {}",
            m.total_pnl
        );
    }

    /// Multiple windows must produce one result per window with sensible ordering.
    #[test]
    fn test_multiple_windows_produce_results() {
        let stream = match try_cuda_stream() {
            Some(s) => s,
            None => return,
        };

        // FEATURE_DIM=42 mirrors production eval-baseline.rs (Market 42 from
        // extract_ml_features). The gather kernel computes `market_dim =
        // feat_dim - SL_OFI_DIM (32)` so anything < 32 produces a negative
        // market_dim → kernel OOB. 42 is the canonical production value.
        const FEATURE_DIM: usize = 42;
        const N_BARS: usize = 300;

        let prices_up = generate_prices(N_BARS, 42, 0.001);
        let prices_down = generate_prices(N_BARS, 123, -0.001);
        let features1 = generate_features(N_BARS, FEATURE_DIM);
        let features2 = generate_features(N_BARS, FEATURE_DIM);

        let config = GpuBacktestConfig {
            max_position: 1.0,
            tx_cost_bps: 0.0,
            spread_cost: 0.0,
            initial_capital: 100_000.0,
            ..Default::default()
        };

        let bars_up = lob_bars_from_prices(&prices_up);
        let bars_down = lob_bars_from_prices(&prices_down);
        let mut evaluator = GpuBacktestEvaluator::new(
            &[prices_up, prices_down],
            &[features1, features2],
            &[bars_up, bars_down],
            FEATURE_DIM,
            config,
            &stream,
        )
        .expect("evaluator creation should succeed");

        // Phase 8.5 (2026-05-12) — wire factored-action branch sizes for
        // env_step decoder. Match production eval-baseline.rs.
        evaluator.set_branch_sizes(&DqnBacktestConfig::from_network_dims((256, 256, 128, 128)));

        let model = constant_action_model(4, &stream); // Always Long100
        let metrics = evaluator
            .evaluate(&model, 24)
            .expect("evaluation should succeed");

        assert_eq!(metrics.len(), 2, "expected exactly 2 window results");

        let m0 = metrics.first().expect("window 0 result must exist");
        let m1 = metrics.get(1).expect("window 1 result must exist");

        // Uptrend window (0) should be more profitable than downtrend window (1)
        assert!(
            m0.total_pnl > m1.total_pnl,
            "expected uptrend window more profitable: {} vs {}",
            m0.total_pnl,
            m1.total_pnl
        );

        // Both drawdowns must be non-negative
        assert!(m0.max_drawdown >= 0.0, "window 0 drawdown negative");
        assert!(m1.max_drawdown >= 0.0, "window 1 drawdown negative");
    }

    /// Extended metrics (VaR, CVaR, Calmar, Omega) must be finite and self-consistent.
    #[test]
    fn test_extended_metrics_populated() {
        let stream = match try_cuda_stream() {
            Some(s) => s,
            None => return,
        };

        // FEATURE_DIM=42 mirrors production eval-baseline.rs (Market 42 from
        // extract_ml_features). The gather kernel computes `market_dim =
        // feat_dim - SL_OFI_DIM (32)` so anything < 32 produces a negative
        // market_dim → kernel OOB. 42 is the canonical production value.
        const FEATURE_DIM: usize = 42;
        const N_BARS: usize = 500;

        let prices = generate_prices(N_BARS, 42, 0.001);
        let features = generate_features(N_BARS, FEATURE_DIM);
        let config = GpuBacktestConfig::default();

        let bars = lob_bars_from_prices(&prices);
        let mut evaluator = GpuBacktestEvaluator::new(
            &[prices],
            &[features],
            &[bars],
            FEATURE_DIM,
            config,
            &stream,
        )
        .expect("evaluator creation should succeed");

        // Phase 8.5 (2026-05-12) — wire factored-action branch sizes for
        // env_step decoder. Match production eval-baseline.rs.
        evaluator.set_branch_sizes(&DqnBacktestConfig::from_network_dims((256, 256, 128, 128)));

        let model = constant_action_model(4, &stream);
        let metrics = evaluator
            .evaluate(&model, 24)
            .expect("evaluation should succeed");

        let m = metrics
            .first()
            .expect("metrics vec must have at least one element");

        assert!(!m.var_95.is_nan(), "VaR should not be NaN");
        assert!(!m.cvar_95.is_nan(), "CVaR should not be NaN");
        assert!(!m.calmar.is_nan(), "Calmar should not be NaN");
        assert!(!m.omega_ratio.is_nan(), "Omega ratio should not be NaN");

        // CVaR (conditional VaR / expected shortfall) must be <= VaR because CVaR
        // averages the worst returns that are already worse than the VaR threshold.
        // Add a small tolerance for floating-point rounding.
        assert!(
            m.cvar_95 <= m.var_95 + 1e-3,
            "CVaR {} should be <= VaR {} (mean of tail should not exceed threshold)",
            m.cvar_95,
            m.var_95
        );
    }

    /// total_trades must be positive when the model takes an active position.
    #[test]
    fn test_active_model_records_trades() {
        let stream = match try_cuda_stream() {
            Some(s) => s,
            None => return,
        };

        // FEATURE_DIM=42 mirrors production eval-baseline.rs (Market 42 from
        // extract_ml_features). The gather kernel computes `market_dim =
        // feat_dim - SL_OFI_DIM (32)` so anything < 32 produces a negative
        // market_dim → kernel OOB. 42 is the canonical production value.
        const FEATURE_DIM: usize = 42;
        const N_BARS: usize = 300;

        let prices = generate_prices(N_BARS, 55, 0.001);
        let features = generate_features(N_BARS, FEATURE_DIM);
        let config = GpuBacktestConfig::default();

        let bars = lob_bars_from_prices(&prices);
        let mut evaluator = GpuBacktestEvaluator::new(
            &[prices],
            &[features],
            &[bars],
            FEATURE_DIM,
            config,
            &stream,
        )
        .expect("evaluator creation should succeed");

        // Phase 8.5 (2026-05-12) — wire factored-action branch sizes for
        // env_step decoder. Match production eval-baseline.rs.
        evaluator.set_branch_sizes(&DqnBacktestConfig::from_network_dims((256, 256, 128, 128)));

        let model = constant_action_model(4, &stream); // Always Long100
        let metrics = evaluator
            .evaluate(&model, 24)
            .expect("evaluation should succeed");

        let m = metrics
            .first()
            .expect("metrics vec must have at least one element");

        // An always-long model must execute at least the initial entry trade
        assert!(
            m.total_trades > 0.0,
            "expected at least one trade for active model, got {}",
            m.total_trades
        );
        assert!(
            (0.0..=1.0).contains(&m.win_rate),
            "win_rate {} out of [0, 1]",
            m.win_rate
        );
    }
}