crackerjack 0.18.2__py3-none-any.whl → 0.45.2__py3-none-any.whl

crackerjack/services/ai/predictive_analytics.py

@@ -0,0 +1,510 @@
+"""Advanced predictive analytics engine for quality metrics and trends."""
+
+import logging
+import statistics
+import typing as t
+from collections import defaultdict, deque
+from dataclasses import dataclass, field
+from datetime import datetime, timedelta
+from pathlib import Path
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class TrendAnalysis:
+    """Trend analysis result for a metric."""
+
+    metric_type: str
+    trend_direction: str  # increasing, decreasing, stable, volatile
+    trend_strength: float  # 0.0 to 1.0
+    predicted_values: list[float]
+    confidence_intervals: list[tuple[float, float]]
+    analysis_period: timedelta
+    last_updated: datetime
+
+
+@dataclass
+class Prediction:
+    """Prediction for a specific metric at a future time."""
+
+    metric_type: str
+    predicted_at: datetime
+    predicted_for: datetime
+    predicted_value: float
+    confidence_interval: tuple[float, float]
+    confidence_level: float
+    model_accuracy: float
+    metadata: dict[str, t.Any] = field(default_factory=dict[str, t.Any])
+
+
+@dataclass
+class CapacityForecast:
+    """Capacity planning forecast."""
+
+    resource_type: str
+    current_usage: float
+    predicted_usage: list[tuple[datetime, float]]
+    capacity_threshold: float
+    estimated_exhaustion: datetime | None
+    recommended_actions: list[str]
+    confidence: float
+
+
+class MovingAveragePredictor:
+    """Simple moving average predictor."""
+
+    def __init__(self, window_size: int = 10):
+        self.window_size = window_size
+
+    def predict(self, values: list[float], periods: int = 1) -> list[float]:
+        """Predict future values using moving average."""
+        if len(values) < self.window_size:
+            return [values[-1]] * periods if values else [0.0] * periods
+
+        recent_values = values[-self.window_size :]
+        ma = statistics.mean(recent_values)
+
+        return [ma] * periods
+
+
+class LinearTrendPredictor:
+    """Linear trend-based predictor."""
+
+    def predict(self, values: list[float], periods: int = 1) -> list[float]:
+        """Predict future values using linear regression."""
+        if len(values) < 2:
+            return [values[-1]] * periods if values else [0.0] * periods
+
+        # Simple linear regression
+        n = len(values)
+        x = list[t.Any](range(n))
+        y = values
+
+        # Calculate slope and intercept
+        x_mean = statistics.mean(x)
+        y_mean = statistics.mean(y)
+
+        numerator = sum((x[i] - x_mean) * (y[i] - y_mean) for i in range(n))
+        denominator = sum((x[i] - x_mean) ** 2 for i in range(n))
+
+        if denominator == 0:
+            return [y_mean] * periods
+
+        slope = numerator / denominator
+        intercept = y_mean - slope * x_mean
+
+        # Predict future values
+        predictions = []
+        for i in range(1, periods + 1):
+            future_x = n + i - 1
+            prediction = slope * future_x + intercept
+            predictions.append(prediction)
+
+        return predictions
+
+
+class SeasonalPredictor:
+    """Seasonal pattern-based predictor."""
+
+    def __init__(self, season_length: int = 24):
+        self.season_length = season_length
+
+    def predict(self, values: list[float], periods: int = 1) -> list[float]:
+        """Predict using seasonal patterns."""
+        if len(values) < self.season_length:
+            return [values[-1]] * periods if values else [0.0] * periods
+
+        predictions = []
+        for i in range(periods):
+            # Use seasonal pattern
+            season_index = (len(values) + i) % self.season_length
+            if season_index < len(values):
+                seasonal_value = values[-(self.season_length - season_index)]
+                predictions.append(seasonal_value)
+            else:
+                predictions.append(values[-1])
+
+        return predictions
+
+
+class PredictiveAnalyticsEngine:
+    """Advanced predictive analytics system for quality metrics."""
+
+    def __init__(self, history_limit: int = 1000):
+        """Initialize predictive analytics engine."""
+        self.history_limit = history_limit
+
+        # Data storage
+        self.metric_history: dict[str, deque[tuple[datetime, float]]] = defaultdict(
+            lambda: deque[tuple[datetime, float]](maxlen=history_limit)
+        )
+
+        # Predictors
+        self.predictors = {
+            "moving_average": MovingAveragePredictor(window_size=10),
+            "linear_trend": LinearTrendPredictor(),
+            "seasonal": SeasonalPredictor(season_length=24),
+        }
+
+        # Cached analyses
+        self.trend_analyses: dict[str, TrendAnalysis] = {}
+        self.predictions_cache: dict[str, list[Prediction]] = defaultdict(list)
+
+        # Configuration
+        self.metric_configs = {
+            "test_pass_rate": {
+                "critical_threshold": 0.8,
+                "optimal_range": (0.95, 1.0),
+                "predictor": "moving_average",
+            },
+            "coverage_percentage": {
+                "critical_threshold": 0.7,
+                "optimal_range": (0.9, 1.0),
+                "predictor": "linear_trend",
+            },
+            "execution_time": {
+                "critical_threshold": 300.0,
+                "predictor": "seasonal",
+            },
+            "memory_usage": {
+                "critical_threshold": 1024.0,
+                "predictor": "linear_trend",
+            },
+            "complexity_score": {
+                "critical_threshold": 15.0,
+                "predictor": "moving_average",
+            },
+        }
+
+    def add_metric(
+        self,
+        metric_type: str,
+        value: float,
+        timestamp: datetime | None = None,
+    ) -> None:
+        """Add new metric data point."""
+        if timestamp is None:
+            timestamp = datetime.now()
+
+        self.metric_history[metric_type].append((timestamp, value))
+
+        # Update trend analysis if we have enough data
+        if len(self.metric_history[metric_type]) >= 10:
+            self._update_trend_analysis(metric_type)
+
+    def _update_trend_analysis(self, metric_type: str) -> None:
+        """Update trend analysis for a metric."""
+        history: list[tuple[datetime, float]] = list(self.metric_history[metric_type])
+        values = [point[1] for point in history]
+        timestamps = [point[0] for point in history]
+
+        # Calculate trend direction and strength
+        trend_direction, trend_strength = self._calculate_trend(values)
+
+        # Generate predictions
+        config = self.metric_configs.get(metric_type, {})
+        predictor_name = config.get("predictor", "moving_average")
+        predictor = self.predictors[predictor_name]
+
+        predicted_values = predictor.predict(values, periods=24)  # 24 periods ahead
+        confidence_intervals = self._calculate_confidence_intervals(
+            values, predicted_values
+        )
+
+        self.trend_analyses[metric_type] = TrendAnalysis(
+            metric_type=metric_type,
+            trend_direction=trend_direction,
+            trend_strength=trend_strength,
+            predicted_values=predicted_values,
+            confidence_intervals=confidence_intervals,
+            analysis_period=timestamps[-1] - timestamps[0],
+            last_updated=datetime.now(),
+        )
+
+    def _calculate_trend(self, values: list[float]) -> tuple[str, float]:
+        """Calculate trend direction and strength."""
+        if len(values) < 3:
+            return "stable", 0.0
+
+        # Use linear regression to determine trend
+        n = len(values)
+        x: list[int] = list(range(n))
+        y = values
+
+        x_mean = statistics.mean(x)
+        y_mean = statistics.mean(y)
+
+        numerator = sum((x[i] - x_mean) * (y[i] - y_mean) for i in range(n))
+        denominator = sum((x[i] - x_mean) ** 2 for i in range(n))
+
+        if denominator == 0:
+            return "stable", 0.0
+
+        slope = numerator / denominator
+
+        # Calculate R-squared for trend strength
+        y_pred = [slope * xi + (y_mean - slope * x_mean) for xi in x]
+        ss_res = sum((y[i] - y_pred[i]) ** 2 for i in range(n))
+        ss_tot = sum((y[i] - y_mean) ** 2 for i in range(n))
+
+        r_squared = 1 - (ss_res / ss_tot) if ss_tot != 0 else 0
+        trend_strength = max(0.0, min(1.0, r_squared))
+
+        # Determine direction
+        if abs(slope) < 0.01:
+            direction = "stable"
+        elif slope > 0:
+            direction = "increasing"
+        else:
+            direction = "decreasing"
+
+        # Check for volatility
+        if trend_strength < 0.3:
+            recent_std = statistics.stdev(values[-10:]) if len(values) >= 10 else 0
+            overall_std = statistics.stdev(values)
+            if recent_std > overall_std * 1.5:
+                direction = "volatile"
+
+        return direction, trend_strength
+
+    def _calculate_confidence_intervals(
+        self, historical: list[float], predictions: list[float]
+    ) -> list[tuple[float, float]]:
+        """Calculate confidence intervals for predictions."""
+        if len(historical) < 2:
+            return [(pred, pred) for pred in predictions]
+
+        # Use historical standard deviation for confidence intervals
+        std_dev = statistics.stdev(historical)
+        confidence_multiplier = 1.96  # 95% confidence
+
+        intervals = []
+        for pred in predictions:
+            lower = pred - confidence_multiplier * std_dev
+            upper = pred + confidence_multiplier * std_dev
+            intervals.append((lower, upper))
+
+        return intervals
+
+    def predict_metric(
+        self,
+        metric_type: str,
+        periods_ahead: int = 1,
+        predictor_name: str | None = None,
+    ) -> list[Prediction]:
+        """Generate predictions for a metric."""
+        if metric_type not in self.metric_history:
+            return []
+
+        history: list[tuple[datetime, float]] = list(self.metric_history[metric_type])
+        values = [point[1] for point in history]
+        last_timestamp = history[-1][0] if history else datetime.now()
+
+        # Select predictor
+        if predictor_name is None:
+            config = self.metric_configs.get(metric_type, {})
+            predictor_name = t.cast(str, config.get("predictor", "moving_average"))
+
+        predictor = self.predictors[predictor_name]
+        predicted_values = predictor.predict(values, periods_ahead)
+
+        # Calculate confidence intervals
+        confidence_intervals = self._calculate_confidence_intervals(
+            values, predicted_values
+        )
+
+        # Calculate model accuracy
+        accuracy = self._calculate_model_accuracy(metric_type, predictor_name)
+
+        # Generate predictions
+        predictions = []
+        for i, (pred_value, conf_interval) in enumerate(
+            zip(predicted_values, confidence_intervals)
+        ):
+            prediction_time = last_timestamp + timedelta(hours=i + 1)
+
+            prediction = Prediction(
+                metric_type=metric_type,
+                predicted_at=datetime.now(),
+                predicted_for=prediction_time,
+                predicted_value=pred_value,
+                confidence_interval=conf_interval,
+                confidence_level=0.95,
+                model_accuracy=accuracy,
+                metadata={"predictor": predictor_name},
+            )
+            predictions.append(prediction)
+
+        # Cache predictions
+        self.predictions_cache[metric_type] = predictions
+
+        return predictions
+
+    def _calculate_model_accuracy(self, metric_type: str, predictor_name: str) -> float:
+        """Calculate historical accuracy of the prediction model."""
+        if len(self.metric_history[metric_type]) < 20:
+            return 0.5  # Default accuracy for insufficient data
+
+        history: list[tuple[datetime, float]] = list(self.metric_history[metric_type])
+        values = [point[1] for point in history]
+
+        # Use last 10 points for validation
+        train_data = values[:-10]
+        validation_data = values[-10:]
+
+        if len(train_data) < 5:
+            return 0.5
+
+        # Generate predictions for validation period
+        predictor = self.predictors[predictor_name]
+        predictions = predictor.predict(train_data, periods=len(validation_data))
+
+        # Calculate accuracy (inverse of mean absolute error)
+        mae = statistics.mean(
+            abs(pred - actual) for pred, actual in zip(predictions, validation_data)
+        )
+
+        # Convert to accuracy score (0-1)
+        if mae == 0:
+            return 1.0
+
+        avg_value = statistics.mean(validation_data)
+        relative_error = mae / abs(avg_value) if avg_value != 0 else mae
+
+        return max(0.1, min(1.0, 1.0 - relative_error))
+
+    def analyze_capacity_requirements(
+        self, resource_type: str, current_usage: float, threshold: float = 0.8
+    ) -> CapacityForecast:
+        """Analyze capacity requirements and forecast exhaustion."""
+        if resource_type not in self.metric_history:
+            return CapacityForecast(
+                resource_type=resource_type,
+                current_usage=current_usage,
+                predicted_usage=[],
+                capacity_threshold=threshold,
+                estimated_exhaustion=None,
+                recommended_actions=["Insufficient data for analysis"],
+                confidence=0.0,
+            )
+
+        # Get predictions for the next 30 days
+        predictions = self.predict_metric(resource_type, periods_ahead=24 * 30)
+
+        predicted_usage = [
+            (pred.predicted_for, pred.predicted_value) for pred in predictions
+        ]
+
+        # Find when threshold will be exceeded
+        estimated_exhaustion = None
+        for timestamp, usage in predicted_usage:
+            if usage >= threshold:
+                estimated_exhaustion = timestamp
+                break
+
+        # Generate recommendations
+        recommendations = self._generate_capacity_recommendations(
+            resource_type, current_usage, threshold, estimated_exhaustion
+        )
+
+        # Calculate confidence based on prediction accuracy
+        avg_accuracy = statistics.mean(pred.model_accuracy for pred in predictions)
+
+        return CapacityForecast(
+            resource_type=resource_type,
+            current_usage=current_usage,
+            predicted_usage=predicted_usage,
+            capacity_threshold=threshold,
+            estimated_exhaustion=estimated_exhaustion,
+            recommended_actions=recommendations,
+            confidence=avg_accuracy,
+        )
+
+    def _generate_capacity_recommendations(
+        self,
+        resource_type: str,
+        current_usage: float,
+        threshold: float,
+        estimated_exhaustion: datetime | None,
+    ) -> list[str]:
+        """Generate capacity planning recommendations."""
+        recommendations: list[str] = []
+
+        utilization = current_usage / threshold if threshold > 0 else 0
+
+        if estimated_exhaustion:
+            days_until = (estimated_exhaustion - datetime.now()).days
+            if days_until < 7:
+                recommendations.extend(
+                    (
+                        f"URGENT: {resource_type} capacity will be exceeded in {days_until} days",
+                        "Consider immediate scaling or optimization",
+                    )
+                )
+            elif days_until < 30:
+                recommendations.append(
+                    f"Plan capacity increase for {resource_type} within {days_until} days"
+                )
+            else:
+                recommendations.append(
+                    f"Monitor {resource_type} usage, capacity limit expected in {days_until} days"
+                )
+
+        if utilization > 0.7:
+            recommendations.extend(
+                (
+                    f"High {resource_type} utilization ({utilization:.1%})",
+                    "Consider proactive scaling",
+                )
+            )
+
+        if not recommendations:
+            recommendations.append(f"{resource_type} capacity is within normal limits")
+
+        return recommendations
+
+    def get_trend_summary(self) -> dict[str, dict[str, t.Any]]:
+        """Get summary of all trend analyses."""
+        summary = {}
+
+        for metric_type, analysis in self.trend_analyses.items():
+            summary[metric_type] = {
+                "trend_direction": analysis.trend_direction,
+                "trend_strength": analysis.trend_strength,
+                "next_predicted_value": analysis.predicted_values[0]
+                if analysis.predicted_values
+                else None,
+                "confidence_range": analysis.confidence_intervals[0]
+                if analysis.confidence_intervals
+                else None,
+                "last_updated": analysis.last_updated.isoformat(),
+            }
+
+        return summary
+
+    def export_analytics_data(self, output_path: str | Path) -> None:
+        """Export analytics data for external analysis."""
+        import json
+
+        data = {
+            "trend_analyses": {
+                metric_type: {
+                    "metric_type": analysis.metric_type,
+                    "trend_direction": analysis.trend_direction,
+                    "trend_strength": analysis.trend_strength,
+                    "predicted_values": analysis.predicted_values[:10],  # Limit size
+                    "analysis_period": analysis.analysis_period.total_seconds(),
+                    "last_updated": analysis.last_updated.isoformat(),
+                }
+                for metric_type, analysis in self.trend_analyses.items()
+            },
+            "predictions_summary": {
+                metric_type: len(predictions)
+                for metric_type, predictions in self.predictions_cache.items()
+            },
+            "exported_at": datetime.now().isoformat(),
+        }
+
+        with open(output_path, "w", encoding="utf-8") as f:
+            json.dump(data, f, indent=2)