from pydantic import BaseModel
from typing import List, Optional
from datetime import datetime

# ═══════════════════════════════════════════════════════════════════════════════
# AUTH SCHEMAS
# ═══════════════════════════════════════════════════════════════════════════════

class UserRegistration(BaseModel):
    """User registration request"""
    email: str
    password: str
    full_name: Optional[str] = ""
    dob: Optional[str] = ""
    sex: Optional[str] = ""
    medical_history: Optional[str] = ""
    geolocation: Optional[str] = ""

class UserCredentials(BaseModel):
    """User login credentials"""
    email: str
    password: str

class UserProfile(BaseModel):
    """User profile information"""
    user_id: str
    email: str
    full_name: Optional[str] = ""
    dob: Optional[str] = ""
    sex: Optional[str] = ""
    medical_history: Optional[str] = ""
    geolocation: Optional[str] = ""
    created_at: str
    conversation_count: int = 0

class UserProfileUpdate(BaseModel):
    """User profile update request"""
    email: Optional[str] = ""
    full_name: Optional[str] = ""
    dob: Optional[str] = ""
    sex: Optional[str] = ""
    medical_history: Optional[str] = ""
    geolocation: Optional[str] = ""

class PasswordChange(BaseModel):
    """Password change request"""
    old_password: str
    new_password: str

class TokenResponse(BaseModel):
    """JWT token response"""
    access_token: str
    token_type: str = "bearer"
    user: UserProfile

class AuthResponse(BaseModel):
    """Generic auth response"""
    success: bool
    message: str
    data: Optional[dict] = None

class Conversation(BaseModel):
    """Conversation metadata"""
    conversation_id: str
    user_id: str
    title: str
    created_at: str
    last_updated: str
    message_count: int = 0
    has_diagnosis: bool = False
    last_message_preview: str = ""

class ConversationMessage(BaseModel):
    """Individual message in conversation"""
    timestamp: str
    role: str  # "user" or "assistant"
    content: str
    language: Optional[str] = "en"
    symptoms: list[str] = []
    quick_replies: list[str] = []

class ConversationDetail(BaseModel):
    """Full conversation with messages and diagnosis"""
    conversation_id: str
    user_id: str
    title: str
    created_at: str
    messages: list[ConversationMessage] = []
    diagnosis: Optional[dict] = None
    recommendations: Optional[str] = None

# === INPUT ===

class PatientInfo(BaseModel):
    """Reçu depuis le frontend, contient les infos du patient."""
    age: int = 0
    sex: str = "Male"
    medical_history: str = ""
    geolocation: str = ""


class ChatMessage(BaseModel):
    """Message individuel dans l'historique de chat, avec un rôle (user ou assistant) et le contenu textuel."""
    role: str      # "user" or "assistant"
    content: str
    language: Optional[str] = "en"
class ChatRequest(BaseModel):
    """Reçu depuis le frontend, contient l'historique complet du chat et le nouveau message de l'utilisateur."""
    history: list[ChatMessage]   # full conversation so far
    message: str                 # new user message
    patient_info: PatientInfo  # infos du patient, pour contexte additionnel


# === OUTPUT ===


class DiagnosisResultPerDisease(BaseModel):
    """Prediction atomique pour une condition médicale donnée."""
    name: str        # exp. "Malaria"
    probability: float    # exp. 0.87
    level: str        # exp. level_low, level_medium, level_high
    symptoms: List[str]

class DiagnosisResponse(BaseModel):
    """Reponse envoyée au frontend, contient une liste de diagnostics possibles et une recommandation."""
    statistics: List[DiagnosisResultPerDisease]
    recommendation: str


class ChatResponse(BaseModel):
    lang: str
    reply: str
    quick_replies: list[str] = []
    # Quand le LLM a suffisament d'infos pour faire un diagnostic, on peut aussi inclure les résultats:
    symptoms: list[str] = []
    diagnosis: dict = {}   # reponse finale de notre modele
    diseases_info: list[dict] = []
    recommendations: str = ""