foxhunt/ml/tests/wave_d_memory_stress_test.rs

//! Wave D Memory Stress Test - 100K+ Symbol Simulation
//!
//! Agent D27: Validates memory scalability and leak detection for production-scale deployments.
//!
//! ## Test Scenario
//! - Simulate 100,000 concurrent symbols (realistic for multi-exchange production)
//! - Each symbol maintains a FeatureExtractionPipeline instance
//! - Generate synthetic OHLCV data (1000 bars per symbol)
//! - Run for 10,000 update cycles
//!
//! ## Memory Targets
//! - Expected: 100K symbols × 4.6KB = 460MB
//! - Maximum allowed: 500MB
//! - No memory leaks: stable after initial allocation
//!
//! ## Success Criteria
//! - ✅ Memory usage <500MB for 100K symbols
//! - ✅ No memory leaks detected (stable RSS)
//! - ✅ Linear scaling O(n) with symbol count
//! - ✅ GC pressure minimal (<1% CPU)

use chrono::Utc;
use ml::features::extraction::OHLCVBar;
use ml::features::pipeline::{FeatureConfig, FeatureExtractionPipeline};
use std::collections::HashMap;
use std::time::{Duration, Instant};
use sysinfo::System;

/// Memory checkpoint for tracking allocations
#[derive(Debug, Clone)]
struct MemoryCheckpoint {
    timestamp: Instant,
    symbol_count: usize,
    rss_bytes: u64,
    virtual_bytes: u64,
    available_bytes: u64,
}

impl MemoryCheckpoint {
    fn capture(sys: &System, symbol_count: usize, start: Instant) -> Self {
        let pid = sysinfo::get_current_pid().expect("Failed to get PID");
        let process = sys.process(pid).expect("Process not found");

        Self {
            timestamp: start,
            symbol_count,
            rss_bytes: process.memory(),
            virtual_bytes: process.virtual_memory(),
            available_bytes: sys.available_memory(),
        }
    }

    fn memory_per_symbol(&self) -> f64 {
        if self.symbol_count == 0 {
            0.0
        } else {
            self.rss_bytes as f64 / self.symbol_count as f64
        }
    }
}

/// Generate synthetic OHLCV bar for testing
fn generate_synthetic_bar(base_price: f64, index: usize) -> OHLCVBar {
    let open = base_price + (index as f64 * 0.1) % 10.0;
    let high = open + 0.5;
    let low = open - 0.5;
    let close = open + (index as f64 * 0.05) % 1.0;
    let volume = 1000.0 + (index as f64 * 10.0) % 500.0;

    OHLCVBar {
        timestamp: Utc::now(),
        open,
        high,
        low,
        close,
        volume,
    }
}

/// Stress test metrics
#[derive(Debug)]
struct StressTestMetrics {
    start_time: Instant,
    end_time: Instant,
    total_symbols: usize,
    total_updates: u64,
    checkpoints: Vec<MemoryCheckpoint>,
    warmup_duration: Duration,
    stress_duration: Duration,
}

impl StressTestMetrics {
    fn new() -> Self {
        let now = Instant::now();
        Self {
            start_time: now,
            end_time: now,
            total_symbols: 0,
            total_updates: 0,
            checkpoints: Vec::new(),
            warmup_duration: Duration::ZERO,
            stress_duration: Duration::ZERO,
        }
    }

    fn memory_growth(&self) -> f64 {
        if self.checkpoints.len() < 2 {
            return 0.0;
        }
        let first = &self.checkpoints[0];
        let last = &self.checkpoints[self.checkpoints.len() - 1];
        ((last.rss_bytes as f64 - first.rss_bytes as f64) / first.rss_bytes as f64) * 100.0
    }

    fn memory_leak_detected(&self) -> bool {
        // Check if memory grew >5% after initial allocation stabilized
        if self.checkpoints.len() < 4 {
            return false;
        }

        // Compare middle checkpoint (after warmup) to final checkpoint
        let mid_idx = self.checkpoints.len() / 2;
        let mid = &self.checkpoints[mid_idx];
        let last = &self.checkpoints[self.checkpoints.len() - 1];

        let growth =
            ((last.rss_bytes as f64 - mid.rss_bytes as f64) / mid.rss_bytes as f64) * 100.0;
        growth > 5.0
    }

    fn print_summary(&self) {
        println!("\n{}", "=".repeat(80));
        println!("Wave D Memory Stress Test - Summary");
        println!("{}", "=".repeat(80));
        println!("Total Symbols:    {}", self.total_symbols);
        println!("Total Updates:    {}", self.total_updates);
        println!("Warmup Duration:  {:?}", self.warmup_duration);
        println!("Stress Duration:  {:?}", self.stress_duration);
        println!("Total Duration:   {:?}", self.end_time - self.start_time);
        println!("\nMemory Checkpoints:");
        println!("{}", "-".repeat(80));
        println!(
            "{:<15} {:<15} {:<15} {:<15}",
            "Symbols", "RSS (MB)", "Virtual (MB)", "Per Symbol (KB)"
        );
        println!("{}", "-".repeat(80));

        for checkpoint in &self.checkpoints {
            println!(
                "{:<15} {:<15.2} {:<15.2} {:<15.2}",
                checkpoint.symbol_count,
                checkpoint.rss_bytes as f64 / 1_048_576.0,
                checkpoint.virtual_bytes as f64 / 1_048_576.0,
                checkpoint.memory_per_symbol() / 1024.0
            );
        }

        println!("{}", "-".repeat(80));
        println!("\nMemory Analysis:");
        println!("  Memory Growth:    {:.2}%", self.memory_growth());
        println!(
            "  Leak Detected:    {}",
            if self.memory_leak_detected() {
                "❌ YES"
            } else {
                "✅ NO"
            }
        );

        if let Some(last) = self.checkpoints.last() {
            let final_mb = last.rss_bytes as f64 / 1_048_576.0;
            let target_mb = 500.0;
            println!("  Final RSS:        {:.2} MB", final_mb);
            println!("  Target:           {:.2} MB", target_mb);
            println!(
                "  Status:           {}",
                if final_mb <= target_mb {
                    "✅ PASS"
                } else {
                    "❌ FAIL"
                }
            );
        }
        println!("{}\n", "=".repeat(80));
    }
}

#[test]
#[ignore] // Expensive test - run explicitly with: cargo test wave_d_memory_stress_100k_symbols -- --ignored --nocapture
fn wave_d_memory_stress_100k_symbols() {
    println!("\n🚀 Starting Wave D Memory Stress Test - 100K Symbols");
    println!("Target: <500MB memory usage, no leaks, linear scaling\n");

    let mut metrics = StressTestMetrics::new();
    let mut sys = System::new_all();
    sys.refresh_all();

    // Capture baseline memory
    let baseline = MemoryCheckpoint::capture(&sys, 0, Instant::now());
    metrics.checkpoints.push(baseline.clone());
    println!(
        "📊 Baseline RSS: {:.2} MB",
        baseline.rss_bytes as f64 / 1_048_576.0
    );

    // Phase 1: Allocate 100K pipelines (simulating 100K symbols)
    const TOTAL_SYMBOLS: usize = 100_000;
    const CHECKPOINT_INTERVALS: [usize; 4] = [1_000, 10_000, 50_000, 100_000];
    const WARMUP_BARS: usize = 50;

    println!(
        "\n🔧 Phase 1: Allocating {} FeatureExtractionPipeline instances...",
        TOTAL_SYMBOLS
    );
    let phase1_start = Instant::now();

    let config = FeatureConfig {
        enable_price: true,
        enable_volume: true,
        enable_time: true,
        enable_indicators: true,
        enable_microstructure: true,
        enable_statistical: true,
        warmup_bars: WARMUP_BARS,
    };

    let mut pipelines: HashMap<String, FeatureExtractionPipeline> =
        HashMap::with_capacity(TOTAL_SYMBOLS);

    for i in 0..TOTAL_SYMBOLS {
        let symbol = format!("SYM{:06}", i);
        let pipeline = FeatureExtractionPipeline::with_config(config.clone());
        pipelines.insert(symbol, pipeline);

        // Memory checkpoints
        if CHECKPOINT_INTERVALS.contains(&(i + 1)) {
            sys.refresh_all();
            let checkpoint = MemoryCheckpoint::capture(&sys, i + 1, phase1_start);
            metrics.checkpoints.push(checkpoint.clone());
            println!(
                "  ✓ {} symbols allocated: RSS {:.2} MB ({:.2} KB/symbol)",
                checkpoint.symbol_count,
                checkpoint.rss_bytes as f64 / 1_048_576.0,
                checkpoint.memory_per_symbol() / 1024.0
            );
        }

        // Progress indicator every 10K
        if (i + 1) % 10_000 == 0 && !CHECKPOINT_INTERVALS.contains(&(i + 1)) {
            println!("  ... {} symbols allocated", i + 1);
        }
    }

    metrics.total_symbols = TOTAL_SYMBOLS;
    metrics.warmup_duration = phase1_start.elapsed();
    println!(
        "✓ Phase 1 Complete: {} symbols in {:?}",
        TOTAL_SYMBOLS, metrics.warmup_duration
    );

    // Phase 2: Warmup (feed 50 bars to each pipeline)
    println!(
        "\n🔥 Phase 2: Warming up pipelines ({} bars per symbol)...",
        WARMUP_BARS
    );
    let phase2_start = Instant::now();

    for (symbol, pipeline) in pipelines.iter_mut() {
        let base_price = 100.0 + (symbol.chars().last().unwrap() as u32 as f64);
        for bar_idx in 0..WARMUP_BARS {
            let bar = generate_synthetic_bar(base_price, bar_idx);
            pipeline.update(&bar);
        }
    }

    let warmup_elapsed = phase2_start.elapsed();
    println!(
        "✓ Phase 2 Complete: Warmup finished in {:?}",
        warmup_elapsed
    );

    // Capture post-warmup memory
    sys.refresh_all();
    let post_warmup = MemoryCheckpoint::capture(&sys, TOTAL_SYMBOLS, phase2_start);
    metrics.checkpoints.push(post_warmup.clone());
    println!(
        "  RSS after warmup: {:.2} MB ({:.2} KB/symbol)",
        post_warmup.rss_bytes as f64 / 1_048_576.0,
        post_warmup.memory_per_symbol() / 1024.0
    );

    // Phase 3: Stress test (10K update cycles on all symbols)
    const UPDATE_CYCLES: usize = 10_000;
    const CYCLE_CHECKPOINTS: [usize; 5] = [1_000, 2_500, 5_000, 7_500, 10_000];

    println!(
        "\n💪 Phase 3: Stress testing with {} update cycles...",
        UPDATE_CYCLES
    );
    let phase3_start = Instant::now();

    for cycle in 0..UPDATE_CYCLES {
        // Update all pipelines in this cycle
        for (symbol, pipeline) in pipelines.iter_mut() {
            let base_price = 100.0 + (symbol.chars().last().unwrap() as u32 as f64);
            let bar = generate_synthetic_bar(base_price, WARMUP_BARS + cycle);
            pipeline.update(&bar);
            metrics.total_updates += 1;
        }

        // Memory checkpoints
        if CYCLE_CHECKPOINTS.contains(&(cycle + 1)) {
            sys.refresh_all();
            let checkpoint = MemoryCheckpoint::capture(&sys, TOTAL_SYMBOLS, phase3_start);
            metrics.checkpoints.push(checkpoint.clone());
            println!(
                "  ✓ Cycle {}/{}: RSS {:.2} MB ({:.2} KB/symbol)",
                cycle + 1,
                UPDATE_CYCLES,
                checkpoint.rss_bytes as f64 / 1_048_576.0,
                checkpoint.memory_per_symbol() / 1024.0
            );
        }

        // Progress indicator every 1000 cycles
        if (cycle + 1) % 1_000 == 0 && !CYCLE_CHECKPOINTS.contains(&(cycle + 1)) {
            println!("  ... Cycle {}/{}", cycle + 1, UPDATE_CYCLES);
        }
    }

    metrics.stress_duration = phase3_start.elapsed();
    metrics.end_time = Instant::now();
    println!(
        "✓ Phase 3 Complete: {} update cycles in {:?}",
        UPDATE_CYCLES, metrics.stress_duration
    );

    // Final memory capture
    sys.refresh_all();
    let final_checkpoint = MemoryCheckpoint::capture(&sys, TOTAL_SYMBOLS, phase3_start);
    metrics.checkpoints.push(final_checkpoint.clone());

    // Print comprehensive summary
    metrics.print_summary();

    // Assertions
    let final_mb = final_checkpoint.rss_bytes as f64 / 1_048_576.0;
    assert!(
        final_mb <= 500.0,
        "Memory usage exceeded 500MB target: {:.2} MB",
        final_mb
    );

    assert!(
        !metrics.memory_leak_detected(),
        "Memory leak detected: RSS grew >5% after warmup stabilization"
    );

    // Check linear scaling (memory per symbol should be consistent)
    if metrics.checkpoints.len() >= 3 {
        let first = &metrics.checkpoints[1]; // After 1K symbols
        let last = &metrics.checkpoints[metrics.checkpoints.len() - 1];

        let first_per_symbol = first.memory_per_symbol();
        let last_per_symbol = last.memory_per_symbol();
        let variance_pct = ((last_per_symbol - first_per_symbol).abs() / first_per_symbol) * 100.0;

        println!(
            "Linear Scaling Check: {:.2} KB/symbol (1K) vs {:.2} KB/symbol (100K) = {:.2}% variance",
            first_per_symbol / 1024.0,
            last_per_symbol / 1024.0,
            variance_pct
        );

        assert!(
            variance_pct < 20.0,
            "Memory per symbol variance too high: {:.2}% (expected <20%)",
            variance_pct
        );
    }

    println!("\n✅ Wave D Memory Stress Test: PASSED");
    println!("   - Memory usage: {:.2} MB / 500 MB target", final_mb);
    println!("   - No memory leaks detected");
    println!("   - Linear scaling confirmed");
}

#[test]
fn wave_d_memory_scaling_small() {
    // Quick test with 1K symbols for CI/CD (non-ignored)
    println!("\n🧪 Wave D Memory Scaling Test - 1K Symbols (Quick)");

    let mut sys = System::new_all();
    sys.refresh_all();

    let baseline = MemoryCheckpoint::capture(&sys, 0, Instant::now());
    println!(
        "Baseline RSS: {:.2} MB",
        baseline.rss_bytes as f64 / 1_048_576.0
    );

    const SYMBOLS: usize = 1_000;
    let config = FeatureConfig::default();
    let mut pipelines: Vec<FeatureExtractionPipeline> = Vec::with_capacity(SYMBOLS);

    // Allocate 1K pipelines
    for _ in 0..SYMBOLS {
        pipelines.push(FeatureExtractionPipeline::with_config(config.clone()));
    }

    // Warmup each pipeline
    for (i, pipeline) in pipelines.iter_mut().enumerate() {
        let base_price = 100.0 + i as f64;
        for bar_idx in 0..50 {
            let bar = generate_synthetic_bar(base_price, bar_idx);
            pipeline.update(&bar);
        }
    }

    sys.refresh_all();
    let after = MemoryCheckpoint::capture(&sys, SYMBOLS, Instant::now());

    // Calculate delta memory (excluding baseline process overhead)
    let delta_mb = (after.rss_bytes as i64 - baseline.rss_bytes as i64) as f64 / 1_048_576.0;
    let per_symbol_kb = (delta_mb * 1024.0) / SYMBOLS as f64;

    println!(
        "After 1K symbols: Total {:.2} MB, Delta {:.2} MB ({:.2} KB/symbol)",
        after.rss_bytes as f64 / 1_048_576.0,
        delta_mb,
        per_symbol_kb
    );

    // For 1K symbols, expect <50MB delta (50KB per symbol including overhead)
    assert!(
        delta_mb <= 50.0,
        "Memory delta too high for 1K symbols: {:.2} MB (expected <50MB)",
        delta_mb
    );

    // Verify each symbol uses reasonable memory (expected ~4.6KB, allow up to 50KB with overhead)
    assert!(
        per_symbol_kb <= 50.0,
        "Memory per symbol too high: {:.2} KB (expected <50KB)",
        per_symbol_kb
    );

    println!("✅ Small-scale memory test: PASSED");
}