"""
Clarification Module for Echolalia Assistant
반향어 분석을 위한 명확화 질문 생성 모듈
"""
import re
import json
import logging
from typing import Dict, Any, List, Tuple, Optional
from dataclasses import dataclass
from enum import Enum

logger = logging.getLogger(__name__)


class AmbiguityType(Enum):
    """반향어 분석을 위한 모호성 유형"""
    CONTEXT_SITUATION = "context_situation"  # 상황 설명
    PREVIOUS_UTTERANCE = "previous_utterance"  # 이전 발화
    CHILD_AGE = "child_age"  # 아이 나이
    VOCABULARY_LEVEL = "vocabulary_level"  # 어휘 수준
    EMOTIONAL_STATE = "emotional_state"  # 감정 상태
    GENERAL = "general"  # 일반적 모호성


@dataclass
class AmbiguityResult:
    """모호성 탐지 결과"""
    is_ambiguous: bool
    ambiguity_score: float  # 0-1, 높을수록 모호함
    missing_facets: List[AmbiguityType]
    reason: str


@dataclass
class ClarifyingQuestion:
    """명확화 질문"""
    question_type: AmbiguityType
    question_text: str
    options: Optional[List[str]] = None
    is_open_ended: bool = False


class AmbiguityDetector:
    """반향어 분석을 위한 모호성 탐지기"""
    
    def __init__(self, config: Dict[str, Any] = None):
        self.config = config or {}
        self.min_length = self.config.get('min_query_length', 5)
        self.threshold = self.config.get('ambiguity_threshold', 0.5)
        
        # 패턴 기반 탐지
        self.patterns = {
            'context_situation': [
                r'상황', r'맥락', r'상황 설명', r'언제', r'어디서', r'어떤 상황'
            ],
            'previous_utterance': [
                r'이전', r'전에', r'앞서', r'질문', r'말', r'발화'
            ],
            'child_age': [
                r'\d+세', r'나이', r'연령', r'몇 살'
            ],
            'vocabulary_level': [
                r'어휘', r'언어 수준', r'수준', r'능력'
            ],
            'emotional_state': [
                r'감정', r'기분', r'상태', r'느낌'
            ]
        }
    
    def detect(self, query: str, existing_info: Optional[Dict[str, Any]] = None) -> AmbiguityResult:
        """모호성 탐지"""
        query_clean = query.strip()
        existing_info = existing_info or {}
        
        # 기본 길이 체크
        if len(query_clean) < self.min_length:
            return AmbiguityResult(
                is_ambiguous=True,
                ambiguity_score=0.9,
                missing_facets=[AmbiguityType.GENERAL],
                reason="질문이 너무 짧습니다"
            )
        
        missing_facets = []
        query_lower = query_clean.lower()
        
        # 각 패턴 확인
        for facet, patterns in self.patterns.items():
            found = False
            for pattern in patterns:
                if re.search(pattern, query_lower):
                    found = True
                    break
            
            # 기존 정보에도 없는 경우
            facet_key = facet
            if facet_key not in existing_info or not existing_info[facet_key]:
                if not found:
                    try:
                        missing_facets.append(AmbiguityType(facet))
                    except ValueError:
                        pass
        
        # 상황 설명과 이전 발화는 특히 중요
        if not existing_info.get('context_situation'):
            if AmbiguityType.CONTEXT_SITUATION not in missing_facets:
                missing_facets.append(AmbiguityType.CONTEXT_SITUATION)
        
        ambiguity_score = len(missing_facets) / len(self.patterns) if missing_facets else 0.0
        is_ambiguous = ambiguity_score >= self.threshold
        
        reason = f"부족한 정보: {', '.join([f.value for f in missing_facets])}" if missing_facets else "충분한 정보"
        
        return AmbiguityResult(
            is_ambiguous=is_ambiguous,
            ambiguity_score=ambiguity_score,
            missing_facets=missing_facets,
            reason=reason
        )


class CQGenerator:
    """명확화 질문 생성기"""
    
    def __init__(self, config: Dict[str, Any] = None):
        self.config = config or {}
        
        # 질문 템플릿
        self.templates = {
            AmbiguityType.CONTEXT_SITUATION: "어떤 상황에서 이 말을 했나요? (예: 식사 시간, 놀이 시간, 이별 상황 등)",
            AmbiguityType.PREVIOUS_UTTERANCE: "아이에게 했던 질문이나 말이 있나요?",
            AmbiguityType.CHILD_AGE: "아이의 나이를 알려주세요.",
            AmbiguityType.VOCABULARY_LEVEL: "아이의 어휘 수준은 어느 정도인가요? (초급/중급/고급)",
            AmbiguityType.EMOTIONAL_STATE: "아이의 감정 상태는 어떤가요? (불안, 평온, 흥분 등)",
            AmbiguityType.GENERAL: "질문을 더 구체적으로 말씀해 주시겠어요?"
        }
        
        # 선택지
        self.options = {
            AmbiguityType.CONTEXT_SITUATION: [
                "식사 시간", "놀이 시간", "외출 준비", "수업 시간", 
                "휴식 시간", "이별/분리 상황", "기타"
            ],
            AmbiguityType.VOCABULARY_LEVEL: ["초급", "중급", "고급"],
            AmbiguityType.EMOTIONAL_STATE: ["불안", "평온", "흥분", "화남", "슬픔", "기타"]
        }
    
    def generate(
        self,
        ambiguity_result: AmbiguityResult,
        original_query: str = "",
        max_questions: int = 2
    ) -> List[ClarifyingQuestion]:
        """명확화 질문 생성"""
        if not ambiguity_result.is_ambiguous:
            return []
        
        questions = []
        for facet in ambiguity_result.missing_facets[:max_questions]:
            question = self._generate_question(facet, original_query)
            if question:
                questions.append(question)
        
        return questions[:max_questions]
    
    def _generate_question(self, facet: AmbiguityType, original_query: str = "") -> Optional[ClarifyingQuestion]:
        """특정 facet에 대한 질문 생성"""
        template = self.templates.get(facet)
        if not template:
            return None
        
        options = self.options.get(facet)
        is_open_ended = (options is None)
        
        return ClarifyingQuestion(
            question_type=facet,
            question_text=template,
            options=options,
            is_open_ended=is_open_ended
        )


class QueryRewriter:
    """쿼리 재작성기"""
    
    def rewrite(
        self,
        original_query: str,
        clarifications: Dict[AmbiguityType, str]
    ) -> str:
        """명확화 응답을 포함하여 쿼리 재작성"""
        if not clarifications:
            return original_query
        
        context_parts = []
        for facet, response in clarifications.items():
            if response and response.strip():
                context_parts.append(response.strip())
        
        if context_parts:
            rewritten = f"{original_query} (상황: {', '.join(context_parts)})"
        else:
            rewritten = original_query
        
        return rewritten


class ClarificationModule:
    """통합 명확화 모듈"""
    
    def __init__(self, config: Dict[str, Any] = None):
        self.config = config or {}
        self.detector = AmbiguityDetector(config)
        self.generator = CQGenerator(config)
        self.rewriter = QueryRewriter()
        
        self.max_rounds = self.config.get('max_clarification_rounds', 2)
        self.current_round = 0
        self.clarifications = {}
        self.original_query_cache = ""
    
    def reset(self) -> None:
        """상태 초기화"""
        self.current_round = 0
        self.clarifications = {}
        self.original_query_cache = ""
    
    def should_clarify(
        self,
        query: str,
        existing_info: Optional[Dict[str, Any]] = None
    ) -> Tuple[bool, AmbiguityResult]:
        """명확화가 필요한지 확인"""
        if self.current_round >= self.max_rounds:
            return False, None
        
        existing_info = existing_info or {}
        ambiguity_result = self.detector.detect(query, existing_info)
        
        return ambiguity_result.is_ambiguous, ambiguity_result
    
    def get_clarifying_questions(
        self,
        ambiguity_result: AmbiguityResult,
        original_query: str = ""
    ) -> List[ClarifyingQuestion]:
        """명확화 질문 가져오기"""
        self.current_round += 1
        if original_query:
            self.original_query_cache = original_query
        return self.generator.generate(ambiguity_result, self.original_query_cache)
    
    def process_response(
        self,
        original_query: str,
        question: ClarifyingQuestion,
        response: str
    ) -> str:
        """사용자 응답 처리"""
        self.clarifications[question.question_type] = response
        rewritten_query = self.rewriter.rewrite(original_query, self.clarifications)
        return rewritten_query