crackerjack 0.33.0__py3-none-any.whl → 0.33.2__py3-none-any.whl

crackerjack/services/quality_intelligence.py

@@ -0,0 +1,785 @@
+"""Advanced ML-based quality intelligence with anomaly detection and predictive analytics."""
+
+import json
+import typing as t
+from dataclasses import dataclass, field
+from datetime import datetime
+from enum import Enum
+from pathlib import Path
+
+import numpy as np
+from scipy import stats
+
+from crackerjack.services.quality_baseline_enhanced import (
+    AlertSeverity,
+    EnhancedQualityBaselineService,
+    TrendDirection,
+)
+
+
+class AnomalyType(str, Enum):
+    """Types of anomalies that can be detected."""
+
+    SPIKE = "spike"  # Sudden increase in metrics
+    DROP = "drop"  # Sudden decrease in metrics
+    DRIFT = "drift"  # Gradual change over time
+    OSCILLATION = "oscillation"  # Unusual fluctuation patterns
+    OUTLIER = "outlier"  # Statistical outlier
+
+
+class PatternType(str, Enum):
+    """Types of patterns that can be identified."""
+
+    CYCLIC = "cyclic"  # Regular recurring patterns
+    SEASONAL = "seasonal"  # Time-based patterns
+    CORRELATION = "correlation"  # Metric correlation patterns
+    REGRESSION = "regression"  # Quality regression patterns
+    IMPROVEMENT = "improvement"  # Quality improvement patterns
+
+
+@dataclass
+class QualityAnomaly:
+    """Detected quality anomaly with ML confidence."""
+
+    anomaly_type: AnomalyType
+    metric_name: str
+    detected_at: datetime
+    confidence: float  # 0.0 to 1.0
+    severity: AlertSeverity
+    description: str
+    actual_value: float
+    expected_value: float
+    deviation_sigma: float  # Standard deviations from normal
+    context: dict[str, t.Any] = field(default_factory=dict[str, t.Any])
+
+    def to_dict(self) -> dict[str, t.Any]:
+        data = {
+            "anomaly_type": self.anomaly_type,
+            "metric_name": self.metric_name,
+            "detected_at": self.detected_at.isoformat(),
+            "confidence": self.confidence,
+            "severity": self.severity,
+            "description": self.description,
+            "actual_value": self.actual_value,
+            "expected_value": self.expected_value,
+            "deviation_sigma": self.deviation_sigma,
+            "context": self.context,
+        }
+        return data
+
+
+@dataclass
+class QualityPattern:
+    """Identified quality pattern with statistical analysis."""
+
+    pattern_type: PatternType
+    metric_names: list[str]
+    detected_at: datetime
+    confidence: float
+    description: str
+    period_days: int
+    correlation_strength: float  # For correlation patterns
+    trend_direction: TrendDirection
+    statistical_significance: float  # p-value
+    context: dict[str, t.Any] = field(default_factory=dict[str, t.Any])
+
+    def to_dict(self) -> dict[str, t.Any]:
+        return {
+            "pattern_type": self.pattern_type,
+            "metric_names": self.metric_names,
+            "detected_at": self.detected_at.isoformat(),
+            "confidence": self.confidence,
+            "description": self.description,
+            "period_days": self.period_days,
+            "correlation_strength": self.correlation_strength,
+            "trend_direction": self.trend_direction,
+            "statistical_significance": self.statistical_significance,
+            "context": self.context,
+        }
+
+
+@dataclass
+class QualityPrediction:
+    """Advanced quality prediction with confidence intervals."""
+
+    metric_name: str
+    predicted_value: float
+    confidence_lower: float
+    confidence_upper: float
+    confidence_level: float  # e.g., 0.95 for 95% confidence
+    prediction_horizon_days: int
+    prediction_method: str
+    created_at: datetime
+    factors: list[str] = field(default_factory=list)
+    risk_assessment: str = "low"  # low, medium, high
+
+    def to_dict(self) -> dict[str, t.Any]:
+        return {
+            "metric_name": self.metric_name,
+            "predicted_value": self.predicted_value,
+            "confidence_lower": self.confidence_lower,
+            "confidence_upper": self.confidence_upper,
+            "confidence_level": self.confidence_level,
+            "prediction_horizon_days": self.prediction_horizon_days,
+            "prediction_method": self.prediction_method,
+            "created_at": self.created_at.isoformat(),
+            "factors": self.factors,
+            "risk_assessment": self.risk_assessment,
+        }
+
+
+@dataclass
+class QualityInsights:
+    """Comprehensive quality insights with ML analysis."""
+
+    anomalies: list[QualityAnomaly]
+    patterns: list[QualityPattern]
+    predictions: list[QualityPrediction]
+    recommendations: list[str]
+    overall_health_score: float  # 0.0 to 1.0
+    risk_level: str  # low, medium, high, critical
+    generated_at: datetime = field(default_factory=datetime.now)
+
+    def to_dict(self) -> dict[str, t.Any]:
+        return {
+            "anomalies": [a.to_dict() for a in self.anomalies],
+            "patterns": [p.to_dict() for p in self.patterns],
+            "predictions": [p.to_dict() for p in self.predictions],
+            "recommendations": self.recommendations,
+            "overall_health_score": self.overall_health_score,
+            "risk_level": self.risk_level,
+            "generated_at": self.generated_at.isoformat(),
+        }
+
+
+class QualityIntelligenceService:
+    """Advanced ML-based quality intelligence service."""
+
+    def __init__(
+        self,
+        quality_service: EnhancedQualityBaselineService,
+        anomaly_sensitivity: float = 2.0,  # Standard deviations for anomaly detection
+        min_data_points: int = 10,
+    ) -> None:
+        self.quality_service = quality_service
+        self.anomaly_sensitivity = anomaly_sensitivity
+        self.min_data_points = min_data_points
+
+    def detect_anomalies(
+        self, days: int = 30, metrics: list[str] | None = None
+    ) -> list[QualityAnomaly]:
+        """Detect anomalies in quality metrics using statistical analysis."""
+        metrics = self._get_default_metrics() if metrics is None else metrics
+
+        baselines = self.quality_service.get_recent_baselines(limit=days * 2)
+        if len(baselines) < self.min_data_points:
+            return []
+
+        anomalies = []
+        for metric_name in metrics:
+            metric_anomalies = self._detect_metric_anomalies(metric_name, baselines)
+            anomalies.extend(metric_anomalies)
+
+        return anomalies
+
+    def _get_default_metrics(self) -> list[str]:
+        """Get default metrics list[t.Any] for anomaly detection."""
+        return [
+            "quality_score",
+            "coverage_percent",
+            "hook_failures",
+            "security_issues",
+            "type_errors",
+            "linting_issues",
+        ]
+
+    def _detect_metric_anomalies(
+        self, metric_name: str, baselines: list[t.Any]
+    ) -> list[QualityAnomaly]:
+        """Detect anomalies for a specific metric."""
+        values, timestamps = self._extract_metric_values(metric_name, baselines)
+
+        if len(values) < self.min_data_points:
+            return []
+
+        stats_data = self._calculate_statistical_metrics(values)
+        if stats_data is None:  # No variation
+            return []
+
+        return self._identify_outlier_anomalies(
+            metric_name, values, timestamps, stats_data
+        )
+
+    def _extract_metric_values(
+        self, metric_name: str, baselines: list[t.Any]
+    ) -> tuple[list[float], list[t.Any]]:
+        """Extract metric values and timestamps from baselines."""
+        values = []
+        timestamps = []
+
+        for baseline in baselines:
+            metric_value = self._get_baseline_metric_value(baseline, metric_name)
+            if metric_value is not None:
+                values.append(metric_value)
+                timestamps.append(baseline.timestamp)
+
+        return values, timestamps
+
+    def _get_baseline_metric_value(
+        self, baseline: t.Any, metric_name: str
+    ) -> float | None:
+        """Get metric value from baseline object."""
+        metric_mapping = {
+            "quality_score": baseline.quality_score,
+            "coverage_percent": baseline.coverage_percent,
+            "hook_failures": baseline.hook_failures,
+            "security_issues": baseline.security_issues,
+            "type_errors": baseline.type_errors,
+            "linting_issues": baseline.linting_issues,
+        }
+        return metric_mapping.get(metric_name)
+
+    def _calculate_statistical_metrics(
+        self, values: list[float]
+    ) -> dict[str, float] | None:
+        """Calculate statistical metrics for anomaly detection."""
+        values_array = np.array(values)
+        mean_val = np.mean(values_array)
+        std_val = np.std(values_array)
+
+        if std_val == 0:
+            return None  # No variation to detect anomalies
+
+        z_scores = np.abs((values_array - mean_val) / std_val)
+
+        return {
+            "mean": mean_val,
+            "std": std_val,
+            "z_scores": z_scores,
+            "values_array": values_array,
+        }
+
+    def _identify_outlier_anomalies(
+        self,
+        metric_name: str,
+        values: list[float],
+        timestamps: list[t.Any],
+        stats_data: dict[str, float],
+    ) -> list[QualityAnomaly]:
+        """Identify outlier anomalies based on z-scores."""
+        anomalies = []
+        z_scores = stats_data["z_scores"]
+        mean_val = stats_data["mean"]
+        std_val = stats_data["std"]
+
+        for i, (value, timestamp, z_score) in enumerate(
+            zip(values, timestamps, z_scores)
+        ):
+            if z_score > self.anomaly_sensitivity:
+                anomaly = self._create_anomaly_object(
+                    metric_name,
+                    value,
+                    timestamp,
+                    z_score,
+                    mean_val,
+                    std_val,
+                    i,
+                    len(values),
+                )
+                anomalies.append(anomaly)
+
+        return anomalies
+
+    def _create_anomaly_object(
+        self,
+        metric_name: str,
+        value: float,
+        timestamp: t.Any,
+        z_score: float,
+        mean_val: float,
+        std_val: float,
+        position: int,
+        data_points: int,
+    ) -> QualityAnomaly:
+        """Create QualityAnomaly object from detected outlier."""
+        anomaly_type, severity = self._determine_anomaly_type_and_severity(
+            value, mean_val, z_score
+        )
+        confidence = min(1.0, z_score / 4.0)  # Scale to 0-1
+
+        return QualityAnomaly(
+            anomaly_type=anomaly_type,
+            metric_name=metric_name,
+            detected_at=timestamp,
+            confidence=confidence,
+            severity=severity,
+            description=f"{metric_name} {anomaly_type} detected: {value:.2f} (expected ~{mean_val:.2f})",
+            actual_value=value,
+            expected_value=mean_val,
+            deviation_sigma=z_score,
+            context={
+                "metric_mean": mean_val,
+                "metric_std": std_val,
+                "data_points": data_points,
+                "position_in_series": position,
+            },
+        )
+
+    def _determine_anomaly_type_and_severity(
+        self, value: float, mean_val: float, z_score: float
+    ) -> tuple[AnomalyType, AlertSeverity]:
+        """Determine anomaly type and severity based on value and z-score."""
+        if value > mean_val:
+            anomaly_type = AnomalyType.SPIKE
+        else:
+            anomaly_type = AnomalyType.DROP
+
+        severity = AlertSeverity.CRITICAL if z_score > 3.0 else AlertSeverity.WARNING
+
+        return anomaly_type, severity
+
+    def identify_patterns(self, days: int = 60) -> list[QualityPattern]:
+        """Identify patterns in quality metrics using correlation and trend analysis."""
+        baselines = self.quality_service.get_recent_baselines(limit=days * 2)
+        if len(baselines) < self.min_data_points:
+            return []
+
+        metrics_data = self._extract_metrics_data(baselines)
+        return self._find_correlation_patterns(metrics_data, days)
+
+    def _extract_metrics_data(self, baselines: list[t.Any]) -> dict[str, list[float]]:
+        """Extract metric data from baselines for correlation analysis."""
+        metrics_data = {
+            "quality_score": [],
+            "coverage_percent": [],
+            "hook_failures": [],
+            "security_issues": [],
+            "type_errors": [],
+            "linting_issues": [],
+        }
+
+        for baseline in baselines:
+            metrics_data["quality_score"].append(baseline.quality_score)
+            metrics_data["coverage_percent"].append(baseline.coverage_percent)
+            metrics_data["hook_failures"].append(baseline.hook_failures)
+            metrics_data["security_issues"].append(baseline.security_issues)
+            metrics_data["type_errors"].append(baseline.type_errors)
+            metrics_data["linting_issues"].append(baseline.linting_issues)
+
+        return metrics_data
+
+    def _find_correlation_patterns(
+        self, metrics_data: dict[str, list[float]], days: int
+    ) -> list[QualityPattern]:
+        """Find correlation patterns between metrics."""
+        patterns = []
+        metric_names = list[t.Any](metrics_data.keys())
+
+        for i, metric1 in enumerate(metric_names):
+            for metric2 in metric_names[i + 1 :]:
+                pattern = self._analyze_metric_correlation(
+                    metric1, metric2, metrics_data, days
+                )
+                if pattern:
+                    patterns.append(pattern)
+
+        return patterns
+
+    def _analyze_metric_correlation(
+        self,
+        metric1: str,
+        metric2: str,
+        metrics_data: dict[str, list[float]],
+        days: int,
+    ) -> QualityPattern | None:
+        """Analyze correlation between two metrics."""
+        values1 = np.array(metrics_data[metric1])
+        values2 = np.array(metrics_data[metric2])
+
+        if len(values1) < self.min_data_points:
+            return None
+
+        correlation, p_value = stats.pearsonr(values1, values2)
+
+        # Strong correlation threshold
+        if abs(correlation) > 0.7 and p_value < 0.05:
+            return self._create_correlation_pattern(
+                metric1, metric2, correlation, p_value, values1, days
+            )
+
+        return None
+
+    def _create_correlation_pattern(
+        self,
+        metric1: str,
+        metric2: str,
+        correlation: float,
+        p_value: float,
+        values1: np.ndarray,
+        days: int,
+    ) -> QualityPattern:
+        """Create a quality pattern from correlation analysis."""
+        trend_dir, description = self._get_correlation_trend_and_description(
+            metric1, metric2, correlation
+        )
+
+        return QualityPattern(
+            pattern_type=PatternType.CORRELATION,
+            metric_names=[metric1, metric2],
+            detected_at=datetime.now(),
+            confidence=abs(correlation),
+            description=description,
+            period_days=days,
+            correlation_strength=abs(correlation),
+            trend_direction=trend_dir,
+            statistical_significance=p_value,
+            context={
+                "correlation_coefficient": correlation,
+                "sample_size": len(values1),
+                "strength": self._get_correlation_strength_label(correlation),
+            },
+        )
+
+    def _get_correlation_trend_and_description(
+        self, metric1: str, metric2: str, correlation: float
+    ) -> tuple[TrendDirection, str]:
+        """Get trend direction and description for correlation."""
+        if correlation > 0:
+            return (
+                TrendDirection.IMPROVING,
+                f"Strong positive correlation between {metric1} and {metric2}",
+            )
+        return (
+            TrendDirection.DECLINING,
+            f"Strong negative correlation between {metric1} and {metric2}",
+        )
+
+    def _get_correlation_strength_label(self, correlation: float) -> str:
+        """Get strength label for correlation coefficient."""
+        abs_corr = abs(correlation)
+        if abs_corr > 0.9:
+            return "very strong"
+        elif abs_corr > 0.7:
+            return "strong"
+        return "moderate"
+
+    def generate_advanced_predictions(
+        self, horizon_days: int = 14, confidence_level: float = 0.95
+    ) -> list[QualityPrediction]:
+        """Generate advanced predictions with confidence intervals."""
+        baselines = self.quality_service.get_recent_baselines(limit=90)
+        if len(baselines) < self.min_data_points:
+            return []
+
+        predictions = []
+        metrics = ["quality_score", "coverage_percent"]
+
+        for metric_name in metrics:
+            values, timestamps = self._extract_time_series(baselines, metric_name)
+
+            if len(values) < self.min_data_points:
+                continue
+
+            prediction = self._create_metric_prediction(
+                metric_name, values, horizon_days, confidence_level
+            )
+            predictions.append(prediction)
+
+        return predictions
+
+    def _extract_time_series(
+        self, baselines: list[t.Any], metric_name: str
+    ) -> tuple[list[t.Any], list[t.Any]]:
+        """Extract time series data for specified metric."""
+        values = []
+        timestamps = []
+
+        for baseline in baselines:
+            if metric_name == "quality_score":
+                values.append(baseline.quality_score)
+            elif metric_name == "coverage_percent":
+                values.append(baseline.coverage_percent)
+            timestamps.append(baseline.timestamp)
+
+        return values, timestamps
+
+    def _create_metric_prediction(
+        self,
+        metric_name: str,
+        values: list[t.Any],
+        horizon_days: int,
+        confidence_level: float,
+    ) -> QualityPrediction:
+        """Create prediction for a single metric."""
+        regression_results = self._perform_linear_regression(values, horizon_days)
+        confidence_bounds = self._calculate_confidence_interval(
+            values, regression_results, confidence_level
+        )
+        risk_level = self._assess_prediction_risk(
+            metric_name, regression_results["predicted_value"]
+        )
+
+        return QualityPrediction(
+            metric_name=metric_name,
+            predicted_value=float(regression_results["predicted_value"]),
+            confidence_lower=float(confidence_bounds["lower"]),
+            confidence_upper=float(confidence_bounds["upper"]),
+            confidence_level=confidence_level,
+            prediction_horizon_days=horizon_days,
+            prediction_method="linear_regression_with_confidence_intervals",
+            created_at=datetime.now(),
+            factors=["historical_trend", "statistical_analysis"],
+            risk_assessment=risk_level,
+        )
+
+    def _perform_linear_regression(
+        self, values: list[t.Any], horizon_days: int
+    ) -> dict[str, t.Any]:
+        """Perform linear regression and predict future value."""
+        values_array = np.array(values)
+        time_indices = np.arange(len(values))
+
+        slope, intercept, r_value, p_value, std_err = stats.linregress(
+            time_indices, values_array
+        )
+
+        future_index = len(values) + horizon_days
+        predicted_value = slope * future_index + intercept
+
+        return {
+            "slope": slope,
+            "intercept": intercept,
+            "predicted_value": predicted_value,
+            "time_indices": time_indices,
+            "values_array": values_array,
+            "horizon_days": horizon_days,
+        }
+
+    def _calculate_confidence_interval(
+        self,
+        values: list[t.Any],
+        regression_results: dict[str, t.Any],
+        confidence_level: float,
+    ) -> dict[str, t.Any]:
+        """Calculate confidence interval for prediction."""
+        slope = regression_results["slope"]
+        intercept = regression_results["intercept"]
+        time_indices = regression_results["time_indices"]
+        values_array = regression_results["values_array"]
+        predicted_value = regression_results["predicted_value"]
+
+        residuals = values_array - (slope * time_indices + intercept)
+        residual_std = np.std(residuals)
+
+        future_index = len(values) + regression_results["horizon_days"]
+        t_value = stats.t.ppf((1 + confidence_level) / 2, len(values) - 2)
+
+        margin_error = self._calculate_margin_error(
+            t_value, residual_std, len(values), future_index, time_indices
+        )
+
+        return {
+            "lower": predicted_value - margin_error,
+            "upper": predicted_value + margin_error,
+        }
+
+    def _calculate_margin_error(
+        self,
+        t_value: float,
+        residual_std: float,
+        n_values: int,
+        future_index: int,
+        time_indices: np.ndarray,
+    ) -> float:
+        """Calculate margin of error for confidence interval."""
+        mean_time: float = float(np.mean(time_indices))
+        sum_sq_diff: float = float(np.sum((time_indices - mean_time) ** 2))
+        numerator: float = (future_index - mean_time) ** 2
+
+        sqrt_term: float = float(np.sqrt(1 + 1 / n_values + numerator / sum_sq_diff))
+        return t_value * residual_std * sqrt_term
+
+    def _assess_prediction_risk(self, metric_name: str, predicted_value: float) -> str:
+        """Assess risk level based on predicted value."""
+        if metric_name == "quality_score":
+            return self._assess_quality_score_risk(predicted_value)
+        # coverage_percent
+        return self._assess_coverage_risk(predicted_value)
+
+    def _assess_quality_score_risk(self, predicted_value: float) -> str:
+        """Assess risk for quality score predictions."""
+        if predicted_value < 70:
+            return "critical"
+        elif predicted_value < 80:
+            return "high"
+        elif predicted_value < 90:
+            return "medium"
+        return "low"
+
+    def _assess_coverage_risk(self, predicted_value: float) -> str:
+        """Assess risk for coverage predictions."""
+        if predicted_value < 70:
+            return "high"
+        elif predicted_value < 85:
+            return "medium"
+        return "low"
+
+    def generate_ml_recommendations(
+        self,
+        anomalies: list[QualityAnomaly],
+        patterns: list[QualityPattern],
+        predictions: list[QualityPrediction],
+    ) -> list[str]:
+        """Generate intelligent recommendations based on ML analysis."""
+        recommendations = []
+
+        # Anomaly-based recommendations
+        critical_anomalies = [
+            a for a in anomalies if a.severity == AlertSeverity.CRITICAL
+        ]
+        if critical_anomalies:
+            recommendations.append(
+                f"🚨 CRITICAL: {len(critical_anomalies)} critical anomalies detected - immediate investigation required"
+            )
+
+        quality_drops = [
+            a
+            for a in anomalies
+            if a.anomaly_type == AnomalyType.DROP and a.metric_name == "quality_score"
+        ]
+        if quality_drops:
+            recommendations.append(
+                "📉 Quality score drops detected - review recent commits and implement quality gates"
+            )
+
+        # Pattern-based recommendations
+        declining_correlations = [
+            p for p in patterns if p.trend_direction == TrendDirection.DECLINING
+        ]
+        if declining_correlations:
+            recommendations.append(
+                f"⚠️ Negative quality correlations identified - investigate dependencies between {declining_correlations[0].metric_names}"
+            )
+
+        strong_patterns = [p for p in patterns if p.confidence > 0.8]
+        if strong_patterns:
+            recommendations.append(
+                "📊 Strong quality patterns detected - leverage insights for predictive quality management"
+            )
+
+        # Prediction-based recommendations
+        high_risk_predictions = [
+            p for p in predictions if p.risk_assessment in ("high", "critical")
+        ]
+        if high_risk_predictions:
+            metrics = [p.metric_name for p in high_risk_predictions]
+            recommendations.append(
+                f"🔮 High-risk quality forecast for {', '.join(metrics)} - proactive intervention recommended"
+            )
+
+        low_confidence_predictions = [
+            p for p in predictions if p.confidence_upper - p.confidence_lower > 20
+        ]
+        if low_confidence_predictions:
+            recommendations.append(
+                "📈 Wide prediction intervals detected - increase data collection frequency for better forecasting"
+            )
+
+        # General ML insights
+        if len(anomalies) > 5:
+            recommendations.append(
+                f"🤖 High anomaly frequency ({len(anomalies)}) suggests systemic quality issues - consider ML-based automated quality monitoring"
+            )
+
+        if not recommendations:
+            recommendations.append(
+                "✅ Quality metrics show stable patterns with no significant anomalies detected - maintain current practices"
+            )
+
+        return recommendations
+
+    def generate_comprehensive_insights(
+        self, analysis_days: int = 30, prediction_days: int = 14
+    ) -> QualityInsights:
+        """Generate comprehensive quality insights with ML analysis."""
+        # Collect all analysis results
+        anomalies = self.detect_anomalies(days=analysis_days)
+        patterns = self.identify_patterns(days=analysis_days * 2)
+        predictions = self.generate_advanced_predictions(horizon_days=prediction_days)
+        recommendations = self.generate_ml_recommendations(
+            anomalies, patterns, predictions
+        )
+
+        # Calculate derived metrics
+        health_score, risk_level = self._calculate_health_metrics(
+            anomalies, predictions
+        )
+
+        return QualityInsights(
+            anomalies=anomalies,
+            patterns=patterns,
+            predictions=predictions,
+            recommendations=recommendations,
+            overall_health_score=health_score,
+            risk_level=risk_level,
+        )
+
+    def _calculate_health_metrics(
+        self, anomalies: list[QualityAnomaly], predictions: list[QualityPrediction]
+    ) -> tuple[float, str]:
+        """Calculate overall health score and risk level."""
+        anomaly_counts = self._count_anomalies_by_severity(anomalies)
+        risk_prediction_count = self._count_high_risk_predictions(predictions)
+
+        health_score = self._compute_health_score(anomaly_counts, risk_prediction_count)
+        risk_level = self._determine_risk_level(health_score)
+
+        return health_score, risk_level
+
+    def _count_anomalies_by_severity(
+        self, anomalies: list[QualityAnomaly]
+    ) -> dict[str, int]:
+        """Count anomalies by severity level."""
+        return {
+            "critical": len(
+                [a for a in anomalies if a.severity == AlertSeverity.CRITICAL]
+            ),
+            "warning": len(
+                [a for a in anomalies if a.severity == AlertSeverity.WARNING]
+            ),
+        }
+
+    def _count_high_risk_predictions(self, predictions: list[QualityPrediction]) -> int:
+        """Count predictions with high or critical risk assessment."""
+        return len(
+            [p for p in predictions if p.risk_assessment in ("high", "critical")]
+        )
+
+    def _compute_health_score(
+        self, anomaly_counts: dict[str, int], risk_predictions: int
+    ) -> float:
+        """Compute health score based on anomalies and risk predictions."""
+        health_score = 1.0
+        health_score -= (
+            anomaly_counts["critical"] * 0.2
+        )  # Critical anomalies heavily impact health
+        health_score -= (
+            anomaly_counts["warning"] * 0.1
+        )  # Warning anomalies moderately impact health
+        health_score -= risk_predictions * 0.15  # High-risk predictions impact health
+        return max(0.0, min(1.0, health_score))
+
+    def _determine_risk_level(self, health_score: float) -> str:
+        """Determine overall risk level based on health score."""
+        if health_score < 0.5:
+            return "critical"
+        elif health_score < 0.7:
+            return "high"
+        elif health_score < 0.85:
+            return "medium"
+        return "low"
+
+    def export_insights(self, insights: QualityInsights, output_path: Path) -> None:
+        """Export quality insights to JSON file."""
+        with output_path.open("w") as f:
+            json.dump(insights.to_dict(), f, indent=2, default=str)