oxirs_vec/

enhanced_performance_monitoring.rs

//! Enhanced Performance Monitoring and Analytics System
//!
//! This module provides comprehensive performance monitoring, analytics dashboard,
//! and quality assurance metrics for the vector search engine.

use anyhow::{anyhow, Result};
use parking_lot::RwLock;
use serde::{Deserialize, Serialize};
use std::collections::{HashMap, VecDeque};
use std::sync::Arc;
use std::time::{Duration, Instant, SystemTime, UNIX_EPOCH};

/// Configuration for performance monitoring
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct MonitoringConfig {
    /// Enable real-time monitoring
    pub enable_real_time: bool,
    /// Enable detailed query logging
    pub enable_query_logging: bool,
    /// Enable system resource monitoring
    pub enable_system_monitoring: bool,
    /// Enable quality metrics collection
    pub enable_quality_metrics: bool,
    /// Metrics retention period
    pub retention_period: Duration,
    /// Sampling rate for metrics (0.0 to 1.0)
    pub sampling_rate: f32,
    /// Alert thresholds
    pub alert_thresholds: AlertThresholds,
    /// Export configuration
    pub export_config: ExportConfig,
}

impl Default for MonitoringConfig {
    fn default() -> Self {
        Self {
            enable_real_time: true,
            enable_query_logging: true,
            enable_system_monitoring: true,
            enable_quality_metrics: true,
            retention_period: Duration::from_secs(24 * 60 * 60), // 24 hours
            sampling_rate: 1.0,
            alert_thresholds: AlertThresholds::default(),
            export_config: ExportConfig::default(),
        }
    }
}

/// Alert threshold configuration
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct AlertThresholds {
    /// Maximum acceptable query latency (ms)
    pub max_query_latency: f64,
    /// Maximum acceptable error rate (0.0 to 1.0)
    pub max_error_rate: f32,
    /// Minimum acceptable recall@10
    pub min_recall_at_10: f32,
    /// Maximum memory usage (MB)
    pub max_memory_usage: u64,
    /// Maximum CPU usage (0.0 to 1.0)
    pub max_cpu_usage: f32,
}

impl Default for AlertThresholds {
    fn default() -> Self {
        Self {
            max_query_latency: 1000.0, // 1 second
            max_error_rate: 0.05,      // 5%
            min_recall_at_10: 0.90,    // 90%
            max_memory_usage: 8192,    // 8GB
            max_cpu_usage: 0.80,       // 80%
        }
    }
}

/// Export configuration for metrics
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ExportConfig {
    /// Export format
    pub format: ExportFormat,
    /// Export interval
    pub export_interval: Duration,
    /// Export destination
    pub destination: ExportDestination,
    /// Include detailed metrics
    pub include_detailed: bool,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum ExportFormat {
    JSON,
    CSV,
    Prometheus,
    InfluxDB,
    ElasticSearch,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum ExportDestination {
    File(String),
    Http(String),
    Database(String),
    Console,
}

impl Default for ExportConfig {
    fn default() -> Self {
        Self {
            format: ExportFormat::JSON,
            export_interval: Duration::from_secs(60),
            destination: ExportDestination::Console,
            include_detailed: false,
        }
    }
}

/// Comprehensive performance monitor
pub struct EnhancedPerformanceMonitor {
    config: MonitoringConfig,
    query_metrics: Arc<RwLock<QueryMetricsCollector>>,
    system_metrics: Arc<RwLock<SystemMetricsCollector>>,
    quality_metrics: Arc<RwLock<QualityMetricsCollector>>,
    alert_manager: AlertManager,
    analytics_engine: AnalyticsEngine,
    dashboard_data: Arc<RwLock<DashboardData>>,
}

impl EnhancedPerformanceMonitor {
    /// Create new performance monitor
    pub fn new(config: MonitoringConfig) -> Self {
        Self {
            config: config.clone(),
            query_metrics: Arc::new(RwLock::new(QueryMetricsCollector::new())),
            system_metrics: Arc::new(RwLock::new(SystemMetricsCollector::new())),
            quality_metrics: Arc::new(RwLock::new(QualityMetricsCollector::new())),
            alert_manager: AlertManager::new(config.alert_thresholds.clone()),
            analytics_engine: AnalyticsEngine::new(),
            dashboard_data: Arc::new(RwLock::new(DashboardData::new())),
        }
    }

    /// Record a query execution
    pub fn record_query(&self, query_info: QueryInfo) {
        if !self.config.enable_real_time {
            return;
        }

        // Check sampling rate
        if {
            #[allow(unused_imports)]
            use scirs2_core::random::{Random, Rng};
            let mut rng = Random::seed(42);
            rng.gen_range(0.0..1.0)
        } > self.config.sampling_rate
        {
            return;
        }

        // Record query metrics
        {
            let mut metrics = self.query_metrics.write();
            metrics.record_query(query_info.clone());
        }

        // Check for alerts
        self.alert_manager.check_query_alerts(&query_info);

        // Update dashboard data
        {
            let mut dashboard = self.dashboard_data.write();
            dashboard.update_query_stats(&query_info);
        }

        // Run analytics
        self.analytics_engine.analyze_query(&query_info);
    }

    /// Record system metrics
    pub fn record_system_metrics(&self, metrics: SystemMetrics) {
        if !self.config.enable_system_monitoring {
            return;
        }

        {
            let mut collector = self.system_metrics.write();
            collector.record_metrics(metrics.clone());
        }

        // Check for system alerts
        self.alert_manager.check_system_alerts(&metrics);

        // Update dashboard
        {
            let mut dashboard = self.dashboard_data.write();
            dashboard.update_system_stats(&metrics);
        }
    }

    /// Record quality metrics
    pub fn record_quality_metrics(&self, metrics: QualityMetrics) {
        if !self.config.enable_quality_metrics {
            return;
        }

        {
            let mut collector = self.quality_metrics.write();
            collector.record_metrics(metrics.clone());
        }

        // Check for quality alerts
        self.alert_manager.check_quality_alerts(&metrics);

        // Update dashboard
        {
            let mut dashboard = self.dashboard_data.write();
            dashboard.update_quality_stats(&metrics);
        }
    }

    /// Get current dashboard data
    pub fn get_dashboard_data(&self) -> DashboardData {
        self.dashboard_data.read().clone()
    }

    /// Generate comprehensive analytics report
    pub fn generate_analytics_report(&self) -> AnalyticsReport {
        let query_stats = self.query_metrics.read().get_statistics();
        let system_stats = self.system_metrics.read().get_statistics();
        let quality_stats = self.quality_metrics.read().get_statistics();
        let alerts = self.alert_manager.get_active_alerts();

        AnalyticsReport {
            timestamp: SystemTime::now(),
            query_statistics: query_stats,
            system_statistics: system_stats,
            quality_statistics: quality_stats,
            active_alerts: alerts,
            trends: self.analytics_engine.get_trends(),
            recommendations: self.analytics_engine.get_recommendations(),
        }
    }

    /// Export metrics
    pub fn export_metrics(&self) -> Result<String> {
        let report = self.generate_analytics_report();

        match self.config.export_config.format {
            ExportFormat::JSON => serde_json::to_string_pretty(&report)
                .map_err(|e| anyhow!("JSON serialization error: {}", e)),
            ExportFormat::CSV => self.generate_csv_export(&report),
            ExportFormat::Prometheus => self.generate_prometheus_export(&report),
            ExportFormat::InfluxDB => self.generate_influxdb_export(&report),
            ExportFormat::ElasticSearch => self.generate_elasticsearch_export(&report),
        }
    }

    /// Generate CSV export
    fn generate_csv_export(&self, report: &AnalyticsReport) -> Result<String> {
        let mut csv = String::new();
        csv.push_str("metric,value,timestamp\n");

        csv.push_str(&format!(
            "total_queries,{},{}\n",
            report.query_statistics.total_queries,
            report
                .timestamp
                .duration_since(UNIX_EPOCH)
                .expect("SystemTime should be after UNIX_EPOCH")
                .as_secs()
        ));

        csv.push_str(&format!(
            "avg_latency,{:.2},{}\n",
            report.query_statistics.average_latency.as_millis(),
            report
                .timestamp
                .duration_since(UNIX_EPOCH)
                .expect("SystemTime should be after UNIX_EPOCH")
                .as_secs()
        ));

        Ok(csv)
    }

    /// Generate Prometheus export
    fn generate_prometheus_export(&self, report: &AnalyticsReport) -> Result<String> {
        let mut prometheus = String::new();

        prometheus.push_str("# HELP vector_search_queries_total Total number of queries\n");
        prometheus.push_str("# TYPE vector_search_queries_total counter\n");
        prometheus.push_str(&format!(
            "vector_search_queries_total {}\n",
            report.query_statistics.total_queries
        ));

        prometheus.push_str("# HELP vector_search_latency_seconds Query latency in seconds\n");
        prometheus.push_str("# TYPE vector_search_latency_seconds histogram\n");
        prometheus.push_str(&format!(
            "vector_search_latency_seconds {:.6}\n",
            report.query_statistics.average_latency.as_secs_f64()
        ));

        Ok(prometheus)
    }

    /// Generate InfluxDB line protocol export.
    ///
    /// Line protocol format: `measurement[,tag_key=tag_val]* field_key=field_val[,…] [timestamp]`
    fn generate_influxdb_export(&self, report: &AnalyticsReport) -> Result<String> {
        let ts_ns = report
            .timestamp
            .duration_since(UNIX_EPOCH)
            .map_err(|e| anyhow!("SystemTime before UNIX_EPOCH: {}", e))?
            .as_nanos();

        let mut output = String::new();

        // Query statistics measurement
        output.push_str(&format!(
            "vector_search_queries total_queries={}u,successful_queries={}u,failed_queries={}u,cache_hit_rate={:.6},avg_latency_ms={:.6} {}\n",
            report.query_statistics.total_queries,
            report.query_statistics.successful_queries,
            report.query_statistics.failed_queries,
            report.query_statistics.cache_hit_rate,
            report.query_statistics.average_latency.as_secs_f64() * 1000.0,
            ts_ns,
        ));

        // System statistics measurement
        output.push_str(&format!(
            "vector_search_system cpu_usage={:.6},memory_usage={}u,peak_memory_usage={}u,peak_cpu_usage={:.6} {}\n",
            report.system_statistics.current_cpu_usage,
            report.system_statistics.current_memory_usage,
            report.system_statistics.peak_memory_usage,
            report.system_statistics.peak_cpu_usage,
            ts_ns,
        ));

        // Quality statistics measurement
        output.push_str(&format!(
            "vector_search_quality precision_at_10={:.6},recall_at_10={:.6},f1_score={:.6},trend_precision={:.6},trend_recall={:.6} {}\n",
            report.quality_statistics.average_precision_at_10,
            report.quality_statistics.average_recall_at_10,
            report.quality_statistics.average_f1_score,
            report.quality_statistics.trend_precision,
            report.quality_statistics.trend_recall,
            ts_ns,
        ));

        Ok(output)
    }

    /// Generate Elasticsearch bulk-index JSON export.
    ///
    /// Each document follows the `{ "index": {} }\n{ ...fields... }\n` ndjson format.
    fn generate_elasticsearch_export(&self, report: &AnalyticsReport) -> Result<String> {
        let ts_secs = report
            .timestamp
            .duration_since(UNIX_EPOCH)
            .map_err(|e| anyhow!("SystemTime before UNIX_EPOCH: {}", e))?
            .as_secs();

        let doc = serde_json::json!({
            "@timestamp": ts_secs,
            "query_statistics": {
                "total_queries": report.query_statistics.total_queries,
                "successful_queries": report.query_statistics.successful_queries,
                "failed_queries": report.query_statistics.failed_queries,
                "average_latency_ms": report.query_statistics.average_latency.as_secs_f64() * 1000.0,
                "min_latency_ms": report.query_statistics.min_latency.as_secs_f64() * 1000.0,
                "max_latency_ms": report.query_statistics.max_latency.as_secs_f64() * 1000.0,
                "cache_hit_rate": report.query_statistics.cache_hit_rate,
                "throughput_qps": report.query_statistics.throughput_qps,
            },
            "system_statistics": {
                "cpu_usage": report.system_statistics.current_cpu_usage,
                "peak_cpu_usage": report.system_statistics.peak_cpu_usage,
                "memory_usage_bytes": report.system_statistics.current_memory_usage,
                "peak_memory_usage_bytes": report.system_statistics.peak_memory_usage,
            },
            "quality_statistics": {
                "precision_at_10": report.quality_statistics.average_precision_at_10,
                "recall_at_10": report.quality_statistics.average_recall_at_10,
                "f1_score": report.quality_statistics.average_f1_score,
                "trend_precision": report.quality_statistics.trend_precision,
                "trend_recall": report.quality_statistics.trend_recall,
            },
            "active_alerts": report.active_alerts.len(),
            "recommendations": report.recommendations.len(),
        });

        let action_line = "{\"index\":{}}\n";
        let doc_line = serde_json::to_string(&doc)
            .map_err(|e| anyhow!("Elasticsearch JSON serialization error: {}", e))?;

        Ok(format!("{}{}\n", action_line, doc_line))
    }

    /// Start background monitoring
    pub fn start_background_monitoring(&self) {
        // In a real implementation, this would start background threads
        // for continuous system monitoring, metric collection, etc.
    }

    /// Stop monitoring
    pub fn stop_monitoring(&self) {
        // Clean shutdown of monitoring systems
    }
}

/// Query information for monitoring
#[derive(Debug, Clone)]
pub struct QueryInfo {
    pub query_id: String,
    pub query_type: QueryType,
    pub query_text: Option<String>,
    pub vector_dimensions: Option<usize>,
    pub k_value: Option<usize>,
    pub threshold: Option<f32>,
    pub start_time: Instant,
    pub end_time: Instant,
    pub success: bool,
    pub error_message: Option<String>,
    pub results_count: usize,
    pub index_used: Option<String>,
    pub cache_hit: bool,
}

#[derive(Debug, Clone)]
pub enum QueryType {
    KNNSearch,
    ThresholdSearch,
    SimilarityCalculation,
    TextEmbedding,
    IndexUpdate,
    Custom(String),
}

/// Query metrics collector
#[derive(Debug)]
pub struct QueryMetricsCollector {
    queries: VecDeque<QueryInfo>,
    statistics: QueryStatistics,
    max_retention: usize,
}

impl Default for QueryMetricsCollector {
    fn default() -> Self {
        Self::new()
    }
}

impl QueryMetricsCollector {
    pub fn new() -> Self {
        Self {
            queries: VecDeque::new(),
            statistics: QueryStatistics::default(),
            max_retention: 10000, // Keep last 10k queries
        }
    }

    pub fn record_query(&mut self, query: QueryInfo) {
        let latency = query.end_time.duration_since(query.start_time);

        // Update statistics
        self.statistics.total_queries += 1;
        if query.success {
            self.statistics.successful_queries += 1;
        } else {
            self.statistics.failed_queries += 1;
        }

        // Update latency statistics
        if self.statistics.total_queries == 1 {
            self.statistics.average_latency = latency;
            self.statistics.min_latency = latency;
            self.statistics.max_latency = latency;
        } else {
            // Update running average
            let total_time = self
                .statistics
                .average_latency
                .mul_f64(self.statistics.total_queries as f64 - 1.0)
                + latency;
            self.statistics.average_latency =
                total_time.div_f64(self.statistics.total_queries as f64);

            if latency < self.statistics.min_latency {
                self.statistics.min_latency = latency;
            }
            if latency > self.statistics.max_latency {
                self.statistics.max_latency = latency;
            }
        }

        // Update latency distribution
        let latency_ms = latency.as_millis() as f64;
        if latency_ms < 10.0 {
            self.statistics.latency_distribution.p10 += 1;
        } else if latency_ms < 50.0 {
            self.statistics.latency_distribution.p50 += 1;
        } else if latency_ms < 100.0 {
            self.statistics.latency_distribution.p90 += 1;
        } else if latency_ms < 500.0 {
            self.statistics.latency_distribution.p95 += 1;
        } else {
            self.statistics.latency_distribution.p99 += 1;
        }

        // Cache hit rate
        if query.cache_hit {
            self.statistics.cache_hit_rate =
                (self.statistics.cache_hit_rate * (self.statistics.total_queries - 1) as f32 + 1.0)
                    / self.statistics.total_queries as f32;
        } else {
            self.statistics.cache_hit_rate = (self.statistics.cache_hit_rate
                * (self.statistics.total_queries - 1) as f32)
                / self.statistics.total_queries as f32;
        }

        // Store query for retention
        self.queries.push_back(query);
        if self.queries.len() > self.max_retention {
            self.queries.pop_front();
        }
    }

    pub fn get_statistics(&self) -> QueryStatistics {
        self.statistics.clone()
    }

    pub fn get_recent_queries(&self, count: usize) -> Vec<&QueryInfo> {
        self.queries.iter().rev().take(count).collect()
    }
}

/// Query statistics
#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct QueryStatistics {
    pub total_queries: u64,
    pub successful_queries: u64,
    pub failed_queries: u64,
    pub average_latency: Duration,
    pub min_latency: Duration,
    pub max_latency: Duration,
    pub latency_distribution: LatencyDistribution,
    pub cache_hit_rate: f32,
    pub throughput_qps: f32,
}

#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct LatencyDistribution {
    pub p10: u64, // < 10ms
    pub p50: u64, // < 50ms
    pub p90: u64, // < 100ms
    pub p95: u64, // < 500ms
    pub p99: u64, // >= 500ms
}

/// System metrics
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct SystemMetrics {
    pub timestamp: SystemTime,
    pub cpu_usage: f32,
    pub memory_usage: u64, // bytes
    pub memory_total: u64, // bytes
    pub disk_usage: u64,   // bytes
    pub disk_total: u64,   // bytes
    pub network_in: u64,   // bytes/sec
    pub network_out: u64,  // bytes/sec
    pub open_file_descriptors: u32,
    pub thread_count: u32,
}

/// System metrics collector
#[derive(Debug)]
pub struct SystemMetricsCollector {
    metrics_history: VecDeque<SystemMetrics>,
    statistics: SystemStatistics,
    max_retention: usize,
}

impl Default for SystemMetricsCollector {
    fn default() -> Self {
        Self::new()
    }
}

impl SystemMetricsCollector {
    pub fn new() -> Self {
        Self {
            metrics_history: VecDeque::new(),
            statistics: SystemStatistics::default(),
            max_retention: 1440, // 24 hours of minute-by-minute data
        }
    }

    pub fn record_metrics(&mut self, metrics: SystemMetrics) {
        // Update statistics
        self.statistics.current_cpu_usage = metrics.cpu_usage;
        self.statistics.current_memory_usage = metrics.memory_usage;

        if metrics.cpu_usage > self.statistics.peak_cpu_usage {
            self.statistics.peak_cpu_usage = metrics.cpu_usage;
        }

        if metrics.memory_usage > self.statistics.peak_memory_usage {
            self.statistics.peak_memory_usage = metrics.memory_usage;
        }

        // Store metrics
        self.metrics_history.push_back(metrics);
        if self.metrics_history.len() > self.max_retention {
            self.metrics_history.pop_front();
        }
    }

    pub fn get_statistics(&self) -> SystemStatistics {
        self.statistics.clone()
    }

    pub fn get_recent_metrics(&self, count: usize) -> Vec<&SystemMetrics> {
        self.metrics_history.iter().rev().take(count).collect()
    }
}

/// System statistics
#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct SystemStatistics {
    pub current_cpu_usage: f32,
    pub peak_cpu_usage: f32,
    pub current_memory_usage: u64,
    pub peak_memory_usage: u64,
    pub average_cpu_usage: f32,
    pub average_memory_usage: u64,
}

/// Quality metrics
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct QualityMetrics {
    pub timestamp: SystemTime,
    pub precision_at_1: f32,
    pub precision_at_5: f32,
    pub precision_at_10: f32,
    pub recall_at_1: f32,
    pub recall_at_5: f32,
    pub recall_at_10: f32,
    pub f1_score: f32,
    pub mrr: f32,  // Mean Reciprocal Rank
    pub ndcg: f32, // Normalized Discounted Cumulative Gain
    pub query_coverage: f32,
    pub result_diversity: f32,
}

/// Quality metrics collector
#[derive(Debug)]
pub struct QualityMetricsCollector {
    metrics_history: VecDeque<QualityMetrics>,
    statistics: QualityStatistics,
    max_retention: usize,
}

impl Default for QualityMetricsCollector {
    fn default() -> Self {
        Self::new()
    }
}

impl QualityMetricsCollector {
    pub fn new() -> Self {
        Self {
            metrics_history: VecDeque::new(),
            statistics: QualityStatistics::default(),
            max_retention: 1000,
        }
    }

    pub fn record_metrics(&mut self, metrics: QualityMetrics) {
        // Update running statistics
        let count = self.metrics_history.len() as f32;
        if count > 0.0 {
            self.statistics.average_precision_at_10 =
                (self.statistics.average_precision_at_10 * count + metrics.precision_at_10)
                    / (count + 1.0);
            self.statistics.average_recall_at_10 = (self.statistics.average_recall_at_10 * count
                + metrics.recall_at_10)
                / (count + 1.0);
            self.statistics.average_f1_score =
                (self.statistics.average_f1_score * count + metrics.f1_score) / (count + 1.0);
        } else {
            self.statistics.average_precision_at_10 = metrics.precision_at_10;
            self.statistics.average_recall_at_10 = metrics.recall_at_10;
            self.statistics.average_f1_score = metrics.f1_score;
        }

        // Store metrics
        self.metrics_history.push_back(metrics);
        if self.metrics_history.len() > self.max_retention {
            self.metrics_history.pop_front();
        }
    }

    pub fn get_statistics(&self) -> QualityStatistics {
        self.statistics.clone()
    }
}

/// Quality statistics
#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct QualityStatistics {
    pub average_precision_at_10: f32,
    pub average_recall_at_10: f32,
    pub average_f1_score: f32,
    pub trend_precision: f32,
    pub trend_recall: f32,
}

/// Alert manager
pub struct AlertManager {
    thresholds: AlertThresholds,
    active_alerts: Arc<RwLock<Vec<Alert>>>,
}

impl AlertManager {
    pub fn new(thresholds: AlertThresholds) -> Self {
        Self {
            thresholds,
            active_alerts: Arc::new(RwLock::new(Vec::new())),
        }
    }

    pub fn check_query_alerts(&self, query: &QueryInfo) {
        let latency_ms = query.end_time.duration_since(query.start_time).as_millis() as f64;

        if latency_ms > self.thresholds.max_query_latency {
            self.add_alert(Alert {
                alert_type: AlertType::HighLatency,
                severity: AlertSeverity::Warning,
                message: format!(
                    "Query latency {}ms exceeds threshold {}ms",
                    latency_ms, self.thresholds.max_query_latency
                ),
                timestamp: SystemTime::now(),
                source: "query_monitor".to_string(),
            });
        }
    }

    pub fn check_system_alerts(&self, metrics: &SystemMetrics) {
        if metrics.cpu_usage > self.thresholds.max_cpu_usage {
            self.add_alert(Alert {
                alert_type: AlertType::HighCpuUsage,
                severity: AlertSeverity::Warning,
                message: format!(
                    "CPU usage {:.1}% exceeds threshold {:.1}%",
                    metrics.cpu_usage * 100.0,
                    self.thresholds.max_cpu_usage * 100.0
                ),
                timestamp: SystemTime::now(),
                source: "system_monitor".to_string(),
            });
        }

        let memory_mb = metrics.memory_usage / (1024 * 1024);
        if memory_mb > self.thresholds.max_memory_usage {
            self.add_alert(Alert {
                alert_type: AlertType::HighMemoryUsage,
                severity: AlertSeverity::Critical,
                message: format!(
                    "Memory usage {}MB exceeds threshold {}MB",
                    memory_mb, self.thresholds.max_memory_usage
                ),
                timestamp: SystemTime::now(),
                source: "system_monitor".to_string(),
            });
        }
    }

    pub fn check_quality_alerts(&self, metrics: &QualityMetrics) {
        if metrics.recall_at_10 < self.thresholds.min_recall_at_10 {
            self.add_alert(Alert {
                alert_type: AlertType::LowRecall,
                severity: AlertSeverity::Warning,
                message: format!(
                    "Recall@10 {:.3} below threshold {:.3}",
                    metrics.recall_at_10, self.thresholds.min_recall_at_10
                ),
                timestamp: SystemTime::now(),
                source: "quality_monitor".to_string(),
            });
        }
    }

    fn add_alert(&self, alert: Alert) {
        let mut alerts = self.active_alerts.write();
        alerts.push(alert);

        // Keep only recent alerts (last hour)
        let cutoff = SystemTime::now() - Duration::from_secs(3600);
        alerts.retain(|a| a.timestamp > cutoff);
    }

    pub fn get_active_alerts(&self) -> Vec<Alert> {
        self.active_alerts.read().clone()
    }
}

/// Alert information
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Alert {
    pub alert_type: AlertType,
    pub severity: AlertSeverity,
    pub message: String,
    pub timestamp: SystemTime,
    pub source: String,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum AlertType {
    HighLatency,
    HighCpuUsage,
    HighMemoryUsage,
    LowRecall,
    HighErrorRate,
    IndexCorruption,
    ServiceDown,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum AlertSeverity {
    Info,
    Warning,
    Critical,
    Emergency,
}

/// Analytics engine for trend analysis and recommendations
pub struct AnalyticsEngine {
    trends: Arc<RwLock<HashMap<String, TrendData>>>,
    recommendations: Arc<RwLock<Vec<Recommendation>>>,
}

impl Default for AnalyticsEngine {
    fn default() -> Self {
        Self::new()
    }
}

impl AnalyticsEngine {
    pub fn new() -> Self {
        Self {
            trends: Arc::new(RwLock::new(HashMap::new())),
            recommendations: Arc::new(RwLock::new(Vec::new())),
        }
    }

    pub fn analyze_query(&self, _query: &QueryInfo) {
        // Placeholder for trend analysis
        // In a full implementation, this would update trend data
    }

    pub fn get_trends(&self) -> HashMap<String, TrendData> {
        self.trends.read().clone()
    }

    pub fn get_recommendations(&self) -> Vec<Recommendation> {
        self.recommendations.read().clone()
    }
}

/// Trend data for analytics
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct TrendData {
    pub metric_name: String,
    pub values: Vec<f64>,
    pub timestamps: Vec<SystemTime>,
    pub trend_direction: TrendDirection,
    pub confidence: f32,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum TrendDirection {
    Increasing,
    Decreasing,
    Stable,
    Volatile,
}

/// Recommendation for system optimization
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Recommendation {
    pub category: RecommendationCategory,
    pub priority: RecommendationPriority,
    pub title: String,
    pub description: String,
    pub estimated_impact: String,
    pub implementation_effort: String,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum RecommendationCategory {
    Performance,
    Quality,
    ResourceOptimization,
    Configuration,
    Maintenance,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum RecommendationPriority {
    Low,
    Medium,
    High,
    Critical,
}

/// Dashboard data structure
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct DashboardData {
    pub last_updated: SystemTime,
    pub query_stats: QueryStatistics,
    pub system_stats: SystemStatistics,
    pub quality_stats: QualityStatistics,
    pub alerts_count: usize,
    pub trends: HashMap<String, Vec<f64>>,
}

impl DashboardData {
    pub fn new() -> Self {
        Self {
            last_updated: SystemTime::now(),
            query_stats: QueryStatistics::default(),
            system_stats: SystemStatistics::default(),
            quality_stats: QualityStatistics::default(),
            alerts_count: 0,
            trends: HashMap::new(),
        }
    }

    pub fn update_query_stats(&mut self, _query: &QueryInfo) {
        // Update query statistics
        self.last_updated = SystemTime::now();
        // Additional updates would be implemented here
    }

    pub fn update_system_stats(&mut self, _metrics: &SystemMetrics) {
        // Update system statistics
        self.last_updated = SystemTime::now();
        // Additional updates would be implemented here
    }

    pub fn update_quality_stats(&mut self, _metrics: &QualityMetrics) {
        // Update quality statistics
        self.last_updated = SystemTime::now();
        // Additional updates would be implemented here
    }
}

impl Default for DashboardData {
    fn default() -> Self {
        Self::new()
    }
}

/// Complete analytics report
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct AnalyticsReport {
    pub timestamp: SystemTime,
    pub query_statistics: QueryStatistics,
    pub system_statistics: SystemStatistics,
    pub quality_statistics: QualityStatistics,
    pub active_alerts: Vec<Alert>,
    pub trends: HashMap<String, TrendData>,
    pub recommendations: Vec<Recommendation>,
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_query_metrics_collection() {
        let mut collector = QueryMetricsCollector::new();

        let query = QueryInfo {
            query_id: "test_query".to_string(),
            query_type: QueryType::KNNSearch,
            query_text: Some("test query".to_string()),
            vector_dimensions: Some(384),
            k_value: Some(10),
            threshold: None,
            start_time: Instant::now() - Duration::from_millis(50),
            end_time: Instant::now(),
            success: true,
            error_message: None,
            results_count: 5,
            index_used: Some("hnsw".to_string()),
            cache_hit: false,
        };

        collector.record_query(query);

        let stats = collector.get_statistics();
        assert_eq!(stats.total_queries, 1);
        assert_eq!(stats.successful_queries, 1);
        assert_eq!(stats.failed_queries, 0);
    }

    #[test]
    fn test_alert_generation() {
        let thresholds = AlertThresholds {
            max_query_latency: 100.0,
            max_error_rate: 0.05,
            min_recall_at_10: 0.90,
            max_memory_usage: 1024,
            max_cpu_usage: 0.80,
        };

        let alert_manager = AlertManager::new(thresholds);

        let query = QueryInfo {
            query_id: "slow_query".to_string(),
            query_type: QueryType::KNNSearch,
            query_text: None,
            vector_dimensions: None,
            k_value: None,
            threshold: None,
            start_time: Instant::now() - Duration::from_millis(200), // Exceeds threshold
            end_time: Instant::now(),
            success: true,
            error_message: None,
            results_count: 5,
            index_used: None,
            cache_hit: false,
        };

        alert_manager.check_query_alerts(&query);

        let alerts = alert_manager.get_active_alerts();
        assert_eq!(alerts.len(), 1);
        assert!(matches!(alerts[0].alert_type, AlertType::HighLatency));
    }

    #[test]
    fn test_performance_monitor() {
        let config = MonitoringConfig::default();
        let monitor = EnhancedPerformanceMonitor::new(config);

        let query = QueryInfo {
            query_id: "test".to_string(),
            query_type: QueryType::KNNSearch,
            query_text: None,
            vector_dimensions: Some(384),
            k_value: Some(10),
            threshold: None,
            start_time: Instant::now() - Duration::from_millis(25),
            end_time: Instant::now(),
            success: true,
            error_message: None,
            results_count: 8,
            index_used: Some("hnsw".to_string()),
            cache_hit: true,
        };

        monitor.record_query(query);

        let dashboard = monitor.get_dashboard_data();
        assert!(dashboard.last_updated <= SystemTime::now());
    }

    #[test]
    fn test_influxdb_export_format() {
        let mut config = MonitoringConfig::default();
        config.export_config.format = ExportFormat::InfluxDB;
        let monitor = EnhancedPerformanceMonitor::new(config);

        let result = monitor.export_metrics();
        assert!(result.is_ok(), "InfluxDB export should succeed");
        let output = result.unwrap();
        // InfluxDB line protocol: measurement fields timestamp
        assert!(
            output.contains("vector_search_queries"),
            "Should contain query measurement"
        );
        assert!(
            output.contains("total_queries="),
            "Should contain total_queries field"
        );
        assert!(
            output.contains("vector_search_system"),
            "Should contain system measurement"
        );
        assert!(
            output.contains("vector_search_quality"),
            "Should contain quality measurement"
        );
        // Each line must have exactly the format: measurement fields timestamp
        for line in output.lines() {
            let parts: Vec<&str> = line.splitn(3, ' ').collect();
            assert_eq!(
                parts.len(),
                3,
                "Each InfluxDB line must have measurement, fields, and timestamp: {:?}",
                line
            );
        }
    }

    #[test]
    fn test_elasticsearch_export_format() {
        let mut config = MonitoringConfig::default();
        config.export_config.format = ExportFormat::ElasticSearch;
        let monitor = EnhancedPerformanceMonitor::new(config);

        let result = monitor.export_metrics();
        assert!(result.is_ok(), "ElasticSearch export should succeed");
        let output = result.unwrap();

        // Elasticsearch bulk format: action line + doc line
        let lines: Vec<&str> = output.lines().collect();
        assert!(
            lines.len() >= 2,
            "Should have at least action and doc lines"
        );
        assert_eq!(lines[0], "{\"index\":{}}", "First line must be bulk action");

        // The doc line must be valid JSON
        let doc: serde_json::Value =
            serde_json::from_str(lines[1]).expect("Elasticsearch document line must be valid JSON");
        assert!(
            doc.get("@timestamp").is_some(),
            "Document must contain @timestamp"
        );
        assert!(
            doc.get("query_statistics").is_some(),
            "Document must contain query_statistics"
        );
        assert!(
            doc.get("system_statistics").is_some(),
            "Document must contain system_statistics"
        );
    }

    #[test]
    fn test_export_json_format() {
        let config = MonitoringConfig::default(); // default format is JSON
        let monitor = EnhancedPerformanceMonitor::new(config);

        let result = monitor.export_metrics();
        assert!(result.is_ok(), "JSON export should succeed");
        let output = result.unwrap();
        let parsed: serde_json::Value =
            serde_json::from_str(&output).expect("JSON export must be valid JSON");
        assert!(parsed.get("query_statistics").is_some());
    }
}