mitai-jinkendo/backend/profile_evaluator.py

"""
Training Type Profiles - Master Evaluator
Comprehensive activity evaluation across all 7 dimensions.

Issue: #15
Date: 2026-03-23
"""
from typing import Dict, Optional, List
from datetime import datetime
import logging

from rule_engine import RuleEvaluator, IntensityZoneEvaluator, TrainingEffectsEvaluator

logger = logging.getLogger(__name__)


class TrainingProfileEvaluator:
    """
    Master class for comprehensive activity evaluation.

    Evaluates an activity against a training type profile across 7 dimensions:
    1. Minimum Requirements (Quality Gates)
    2. Intensity Zones (HR zones)
    3. Training Effects (Abilities)
    4. Periodization (Frequency & Recovery)
    5. Performance Indicators (KPIs)
    6. Safety (Warnings)
    7. AI Context
    """

    def __init__(self, parameters_registry: Dict[str, Dict]):
        """
        Initialize evaluator with parameter registry.

        Args:
            parameters_registry: Dict mapping parameter_key -> config
        """
        self.parameters_registry = parameters_registry
        self.rule_evaluator = RuleEvaluator()
        self.zone_evaluator = IntensityZoneEvaluator()
        self.effects_evaluator = TrainingEffectsEvaluator()

    def evaluate_activity(
        self,
        activity: Dict,
        training_type_profile: Optional[Dict],
        context: Optional[Dict] = None
    ) -> Dict:
        """
        Complete evaluation of an activity against its training type profile.

        Args:
            activity: Activity data dictionary
            training_type_profile: Training type profile (JSONB)
            context: {
                "user_profile": {...},
                "recent_activities": [...],
                "historical_activities": [...]
            }

        Returns:
            {
                "evaluated_at": ISO timestamp,
                "profile_version": str,
                "rule_set_results": {
                    "minimum_requirements": {...},
                    "intensity_zones": {...},
                    "training_effects": {...},
                    "periodization": {...},
                    "performance_indicators": {...},
                    "safety": {...}
                },
                "overall_score": float (0-1),
                "quality_label": str,
                "recommendations": [str],
                "warnings": [str]
            }
        """
        # No profile? Return unvalidated result
        if not training_type_profile:
            return self._create_unvalidated_result()

        rule_sets = training_type_profile.get("rule_sets", {})
        context = context or {}

        results = {
            "evaluated_at": datetime.now().isoformat(),
            "profile_version": training_type_profile.get("version", "unknown"),
            "rule_set_results": {}
        }

        # ━━━ 1. MINIMUM REQUIREMENTS ━━━
        if "minimum_requirements" in rule_sets:
            results["rule_set_results"]["minimum_requirements"] = \
                self.rule_evaluator.evaluate_rule_set(
                    rule_sets["minimum_requirements"],
                    activity,
                    self.parameters_registry
                )

        # ━━━ 2. INTENSITY ZONES ━━━
        if "intensity_zones" in rule_sets:
            results["rule_set_results"]["intensity_zones"] = \
                self.zone_evaluator.evaluate(
                    rule_sets["intensity_zones"],
                    activity,
                    context.get("user_profile", {})
                )

        # ━━━ 3. TRAINING EFFECTS ━━━
        if "training_effects" in rule_sets:
            results["rule_set_results"]["training_effects"] = \
                self.effects_evaluator.evaluate(
                    rule_sets["training_effects"],
                    activity,
                    results["rule_set_results"].get("intensity_zones")
                )

        # ━━━ 4. PERIODIZATION ━━━
        if "periodization" in rule_sets:
            results["rule_set_results"]["periodization"] = \
                self._evaluate_periodization(
                    rule_sets["periodization"],
                    activity,
                    context.get("recent_activities", [])
                )

        # ━━━ 5. PERFORMANCE INDICATORS ━━━
        if "performance_indicators" in rule_sets:
            results["rule_set_results"]["performance_indicators"] = \
                self._evaluate_performance(
                    rule_sets["performance_indicators"],
                    activity,
                    context.get("historical_activities", [])
                )

        # ━━━ 6. SAFETY WARNINGS ━━━
        if "safety" in rule_sets:
            results["rule_set_results"]["safety"] = \
                self._evaluate_safety(
                    rule_sets["safety"],
                    activity
                )

        # ━━━ OVERALL SCORE & QUALITY LABEL ━━━
        overall_score = self._calculate_overall_score(results["rule_set_results"])
        results["overall_score"] = overall_score
        results["quality_label"] = self._get_quality_label(overall_score)

        # ━━━ RECOMMENDATIONS & WARNINGS ━━━
        results["recommendations"] = self._generate_recommendations(results)
        results["warnings"] = self._collect_warnings(results)

        return results

    def _create_unvalidated_result(self) -> Dict:
        """Creates result for activities without profile."""
        return {
            "evaluated_at": datetime.now().isoformat(),
            "profile_version": None,
            "rule_set_results": {},
            "overall_score": None,
            "quality_label": None,
            "recommendations": ["Kein Trainingsprofil konfiguriert"],
            "warnings": []
        }

    def _evaluate_periodization(
        self,
        config: Dict,
        activity: Dict,
        recent_activities: List[Dict]
    ) -> Dict:
        """
        Evaluates periodization compliance (frequency & recovery).

        Simplified for MVP - full implementation later.
        """
        if not config.get("enabled", False):
            return {"enabled": False}

        # Basic frequency check
        training_type_id = activity.get("training_type_id")
        same_type_this_week = sum(
            1 for a in recent_activities
            if a.get("training_type_id") == training_type_id
        )

        frequency_config = config.get("frequency", {})
        optimal = frequency_config.get("per_week_optimal", 3)

        return {
            "enabled": True,
            "weekly_count": same_type_this_week,
            "optimal_count": optimal,
            "frequency_status": "optimal" if same_type_this_week <= optimal else "over_optimal",
            "recovery_adequate": True,  # Simplified for MVP
            "warning": None
        }

    def _evaluate_performance(
        self,
        config: Dict,
        activity: Dict,
        historical_activities: List[Dict]
    ) -> Dict:
        """
        Evaluates performance development.

        Simplified for MVP - full implementation later.
        """
        if not config.get("enabled", False):
            return {"enabled": False}

        return {
            "enabled": True,
            "trend": "stable",  # Simplified
            "metrics_comparison": {},
            "benchmark_level": "intermediate"
        }

    def _evaluate_safety(self, config: Dict, activity: Dict) -> Dict:
        """
        Evaluates safety warnings.
        """
        if not config.get("enabled", False):
            return {"enabled": False, "warnings": []}

        warnings_config = config.get("warnings", [])
        triggered_warnings = []

        for warning_rule in warnings_config:
            param_key = warning_rule.get("parameter")
            operator = warning_rule.get("operator")
            threshold = warning_rule.get("value")
            severity = warning_rule.get("severity", "medium")
            message = warning_rule.get("message", "")

            actual_value = activity.get(param_key)

            if actual_value is not None:
                operator_func = RuleEvaluator.OPERATORS.get(operator)
                if operator_func and operator_func(actual_value, threshold):
                    triggered_warnings.append({
                        "severity": severity,
                        "message": message,
                        "parameter": param_key,
                        "actual_value": actual_value,
                        "threshold": threshold
                    })

        return {
            "enabled": True,
            "warnings": triggered_warnings
        }

    def _calculate_overall_score(self, rule_set_results: Dict) -> float:
        """
        Calculates weighted overall score.

        Weights:
        - Minimum Requirements: 40%
        - Intensity Zones: 20%
        - Periodization: 20%
        - Performance: 10%
        - Training Effects: 10%
        """
        weights = {
            "minimum_requirements": 0.4,
            "intensity_zones": 0.2,
            "periodization": 0.2,
            "performance_indicators": 0.1,
            "training_effects": 0.1
        }

        total_score = 0.0
        total_weight = 0.0

        for rule_set_name, weight in weights.items():
            result = rule_set_results.get(rule_set_name)
            if result and result.get("enabled"):
                score = result.get("score", 0.5)

                # Special handling for different result types
                if rule_set_name == "intensity_zones":
                    score = result.get("duration_quality", 0.5)
                elif rule_set_name == "periodization":
                    score = 1.0 if result.get("recovery_adequate", False) else 0.5

                total_score += score * weight
                total_weight += weight

        return round(total_score / total_weight, 2) if total_weight > 0 else 0.5

    def _get_quality_label(self, score: Optional[float]) -> Optional[str]:
        """Converts score to quality label."""
        if score is None:
            return None

        if score >= 0.9:
            return "excellent"
        elif score >= 0.7:
            return "good"
        elif score >= 0.5:
            return "acceptable"
        else:
            return "poor"

    def _generate_recommendations(self, results: Dict) -> List[str]:
        """Generates actionable recommendations."""
        recommendations = []

        # Check minimum requirements
        min_req = results["rule_set_results"].get("minimum_requirements", {})
        if min_req.get("enabled") and not min_req.get("passed"):
            for failed in min_req.get("failed_rules", []):
                param = failed.get("parameter")
                actual = failed.get("actual_value")
                expected = failed.get("expected_value")
                reason = failed.get("reason", "")
                symbol = failed.get("operator_symbol", "")

                recommendations.append(
                    f"{param}: {actual} {symbol} {expected} - {reason}"
                )

        # Check intensity zones
        zone_result = results["rule_set_results"].get("intensity_zones", {})
        if zone_result.get("enabled") and zone_result.get("recommendation"):
            recommendations.append(zone_result["recommendation"])

        # Default recommendation if excellent
        if results.get("quality_label") == "excellent" and not recommendations:
            recommendations.append("Hervorragendes Training! Weiter so.")

        return recommendations

    def _collect_warnings(self, results: Dict) -> List[str]:
        """Collects all warnings from safety checks."""
        safety_result = results["rule_set_results"].get("safety", {})
        if not safety_result.get("enabled"):
            return []

        warnings = []
        for warning in safety_result.get("warnings", []):
            severity_icon = "🔴" if warning["severity"] == "high" else "⚠️"
            warnings.append(f"{severity_icon} {warning['message']}")

        return warnings