"""
Reflection & Analytics Repositories.

CRUD-Operationen fuer Lesson-Reflections und Analytics-Abfragen (Phase 5).
"""
from datetime import datetime
from typing import Optional, List, Dict, Any

from sqlalchemy.orm import Session as DBSession

from .db_models import LessonSessionDB, LessonPhaseEnum, LessonReflectionDB
from .analytics import (
    LessonReflection, SessionSummary, TeacherAnalytics, AnalyticsCalculator,
)


class ReflectionRepository:
    """Repository fuer LessonReflection CRUD-Operationen."""

    def __init__(self, db: DBSession):
        self.db = db

    # ==================== CREATE ====================

    def create(self, reflection: LessonReflection) -> LessonReflectionDB:
        """Erstellt eine neue Reflection."""
        db_reflection = LessonReflectionDB(
            id=reflection.reflection_id,
            session_id=reflection.session_id,
            teacher_id=reflection.teacher_id,
            notes=reflection.notes,
            overall_rating=reflection.overall_rating,
            what_worked=reflection.what_worked,
            improvements=reflection.improvements,
            notes_for_next_lesson=reflection.notes_for_next_lesson,
        )
        self.db.add(db_reflection)
        self.db.commit()
        self.db.refresh(db_reflection)
        return db_reflection

    # ==================== READ ====================

    def get_by_id(self, reflection_id: str) -> Optional[LessonReflectionDB]:
        """Holt eine Reflection nach ID."""
        return self.db.query(LessonReflectionDB).filter(
            LessonReflectionDB.id == reflection_id
        ).first()

    def get_by_session(self, session_id: str) -> Optional[LessonReflectionDB]:
        """Holt die Reflection einer Session."""
        return self.db.query(LessonReflectionDB).filter(
            LessonReflectionDB.session_id == session_id
        ).first()

    def get_by_teacher(
        self,
        teacher_id: str,
        limit: int = 20,
        offset: int = 0
    ) -> List[LessonReflectionDB]:
        """Holt alle Reflections eines Lehrers."""
        return self.db.query(LessonReflectionDB).filter(
            LessonReflectionDB.teacher_id == teacher_id
        ).order_by(
            LessonReflectionDB.created_at.desc()
        ).offset(offset).limit(limit).all()

    # ==================== UPDATE ====================

    def update(self, reflection: LessonReflection) -> Optional[LessonReflectionDB]:
        """Aktualisiert eine Reflection."""
        db_reflection = self.get_by_id(reflection.reflection_id)
        if not db_reflection:
            return None

        db_reflection.notes = reflection.notes
        db_reflection.overall_rating = reflection.overall_rating
        db_reflection.what_worked = reflection.what_worked
        db_reflection.improvements = reflection.improvements
        db_reflection.notes_for_next_lesson = reflection.notes_for_next_lesson

        self.db.commit()
        self.db.refresh(db_reflection)
        return db_reflection

    # ==================== DELETE ====================

    def delete(self, reflection_id: str) -> bool:
        """Loescht eine Reflection."""
        db_reflection = self.get_by_id(reflection_id)
        if not db_reflection:
            return False

        self.db.delete(db_reflection)
        self.db.commit()
        return True

    # ==================== CONVERSION ====================

    def to_dataclass(self, db_reflection: LessonReflectionDB) -> LessonReflection:
        """Konvertiert DB-Model zu Dataclass."""
        return LessonReflection(
            reflection_id=db_reflection.id,
            session_id=db_reflection.session_id,
            teacher_id=db_reflection.teacher_id,
            notes=db_reflection.notes or "",
            overall_rating=db_reflection.overall_rating,
            what_worked=db_reflection.what_worked or [],
            improvements=db_reflection.improvements or [],
            notes_for_next_lesson=db_reflection.notes_for_next_lesson or "",
            created_at=db_reflection.created_at,
            updated_at=db_reflection.updated_at,
        )


class AnalyticsRepository:
    """Repository fuer Analytics-Abfragen."""

    def __init__(self, db: DBSession):
        self.db = db

    def get_session_summary(self, session_id: str) -> Optional[SessionSummary]:
        """
        Berechnet die Summary einer abgeschlossenen Session.

        Args:
            session_id: ID der Session

        Returns:
            SessionSummary oder None wenn Session nicht gefunden
        """
        db_session = self.db.query(LessonSessionDB).filter(
            LessonSessionDB.id == session_id
        ).first()

        if not db_session:
            return None

        # Session-Daten zusammenstellen
        session_data = {
            "session_id": db_session.id,
            "teacher_id": db_session.teacher_id,
            "class_id": db_session.class_id,
            "subject": db_session.subject,
            "topic": db_session.topic,
            "lesson_started_at": db_session.lesson_started_at,
            "lesson_ended_at": db_session.lesson_ended_at,
            "phase_durations": db_session.phase_durations or {},
        }

        # Phase History aus DB oder JSON
        phase_history = db_session.phase_history or []

        # Summary berechnen
        return AnalyticsCalculator.calculate_session_summary(
            session_data, phase_history
        )

    def get_teacher_analytics(
        self,
        teacher_id: str,
        period_start: Optional[datetime] = None,
        period_end: Optional[datetime] = None
    ) -> TeacherAnalytics:
        """
        Berechnet aggregierte Statistiken fuer einen Lehrer.

        Args:
            teacher_id: ID des Lehrers
            period_start: Beginn des Zeitraums (default: 30 Tage zurueck)
            period_end: Ende des Zeitraums (default: jetzt)

        Returns:
            TeacherAnalytics mit aggregierten Statistiken
        """
        from datetime import timedelta

        if not period_end:
            period_end = datetime.utcnow()
        if not period_start:
            period_start = period_end - timedelta(days=30)

        # Sessions im Zeitraum abfragen
        sessions_query = self.db.query(LessonSessionDB).filter(
            LessonSessionDB.teacher_id == teacher_id,
            LessonSessionDB.lesson_started_at >= period_start,
            LessonSessionDB.lesson_started_at <= period_end
        ).all()

        # Sessions zu Dictionaries konvertieren
        sessions_data = []
        for db_session in sessions_query:
            sessions_data.append({
                "session_id": db_session.id,
                "teacher_id": db_session.teacher_id,
                "class_id": db_session.class_id,
                "subject": db_session.subject,
                "topic": db_session.topic,
                "lesson_started_at": db_session.lesson_started_at,
                "lesson_ended_at": db_session.lesson_ended_at,
                "phase_durations": db_session.phase_durations or {},
                "phase_history": db_session.phase_history or [],
            })

        return AnalyticsCalculator.calculate_teacher_analytics(
            sessions_data, period_start, period_end
        )

    def get_phase_duration_trends(
        self,
        teacher_id: str,
        phase: str,
        limit: int = 20
    ) -> List[Dict[str, Any]]:
        """
        Gibt die Dauer-Trends fuer eine bestimmte Phase zurueck.

        Args:
            teacher_id: ID des Lehrers
            phase: Phasen-ID (einstieg, erarbeitung, etc.)
            limit: Max Anzahl der Datenpunkte

        Returns:
            Liste von Datenpunkten [{date, planned, actual, difference}]
        """
        sessions = self.db.query(LessonSessionDB).filter(
            LessonSessionDB.teacher_id == teacher_id,
            LessonSessionDB.current_phase == LessonPhaseEnum.ENDED
        ).order_by(
            LessonSessionDB.lesson_ended_at.desc()
        ).limit(limit).all()

        trends = []
        for db_session in sessions:
            history = db_session.phase_history or []
            for entry in history:
                if entry.get("phase") == phase:
                    planned = (db_session.phase_durations or {}).get(phase, 0) * 60
                    actual = entry.get("duration_seconds", 0) or 0
                    trends.append({
                        "date": db_session.lesson_started_at.isoformat() if db_session.lesson_started_at else None,
                        "session_id": db_session.id,
                        "subject": db_session.subject,
                        "planned_seconds": planned,
                        "actual_seconds": actual,
                        "difference_seconds": actual - planned,
                    })
                    break

        return list(reversed(trends))  # Chronologisch sortieren

    def get_overtime_analysis(
        self,
        teacher_id: str,
        limit: int = 30
    ) -> Dict[str, Any]:
        """
        Analysiert Overtime-Muster.

        Args:
            teacher_id: ID des Lehrers
            limit: Anzahl der zu analysierenden Sessions

        Returns:
            Dict mit Overtime-Statistiken pro Phase
        """
        sessions = self.db.query(LessonSessionDB).filter(
            LessonSessionDB.teacher_id == teacher_id,
            LessonSessionDB.current_phase == LessonPhaseEnum.ENDED
        ).order_by(
            LessonSessionDB.lesson_ended_at.desc()
        ).limit(limit).all()

        phase_overtime: Dict[str, List[int]] = {
            "einstieg": [],
            "erarbeitung": [],
            "sicherung": [],
            "transfer": [],
            "reflexion": [],
        }

        for db_session in sessions:
            history = db_session.phase_history or []
            phase_durations = db_session.phase_durations or {}

            for entry in history:
                phase = entry.get("phase", "")
                if phase in phase_overtime:
                    planned = phase_durations.get(phase, 0) * 60
                    actual = entry.get("duration_seconds", 0) or 0
                    overtime = max(0, actual - planned)
                    phase_overtime[phase].append(overtime)

        # Statistiken berechnen
        result = {}
        for phase, overtimes in phase_overtime.items():
            if overtimes:
                result[phase] = {
                    "count": len([o for o in overtimes if o > 0]),
                    "total": len(overtimes),
                    "avg_overtime_seconds": sum(overtimes) / len(overtimes),
                    "max_overtime_seconds": max(overtimes),
                    "overtime_percentage": len([o for o in overtimes if o > 0]) / len(overtimes) * 100,
                }
            else:
                result[phase] = {
                    "count": 0,
                    "total": 0,
                    "avg_overtime_seconds": 0,
                    "max_overtime_seconds": 0,
                    "overtime_percentage": 0,
                }

        return result