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

class PredictionResponse(BaseModel):
    id: str
    original_image_url: str
    prediction_image_url: str
    overlay_image_url: str
    confidence_score: float
    predicted_area_percentage: float
    has_stroke: bool
    processing_time: float
    created_at: datetime
    doctor_expectation: Optional[bool] = None  # Doctor's diagnosis: True=expects stroke, False=expects normal, None=not provided
    stroke_type: Optional[str] = None  # Doctor's case type classification: "normal", "ischemia", "bleeding", or None

class PaginationInfo(BaseModel):
    page: int
    limit: int
    total: int
    totalPages: int
    hasNext: bool
    hasPrev: bool

class PredictionHistoryResponse(BaseModel):
    items: List[PredictionResponse]
    pagination: PaginationInfo

class SystemInfo(BaseModel):
    model_loaded: bool
    device: str
    config_validation: bool
    mongodb_connected: bool
    cloudinary_configured: bool
    timestamp: datetime

class HealthCheck(BaseModel):
    status: str
    model_loaded: bool
    device: str
    optional_services: dict
    timestamp: datetime

class TrendData(BaseModel):
    """Trend data comparing current period to previous period"""
    total_predictions: Optional[float] = 0.0
    stroke_detected: Optional[float] = 0.0
    normal_cases: Optional[float] = 0.0
    detection_rate: Optional[float] = 0.0
    average_confidence: Optional[float] = 0.0

class DetailedAnalytics(BaseModel):
    """Detailed analytics metrics"""
    detection_accuracy: float = 0.0
    detection_accuracy_trend: float = 0.0
    average_processing_time: float = 0.0
    processing_time_trend: float = 0.0
    false_positive_rate: float = 0.0
    false_positive_trend: float = 0.0
    false_negative_rate: float = 0.0
    false_negative_trend: float = 0.0

class PerformanceMetrics(BaseModel):
    """ML model performance metrics"""
    sensitivity: float = 0.0  # True Positive Rate
    specificity: float = 0.0  # True Negative Rate
    positive_predictive_value: float = 0.0  # Precision
    negative_predictive_value: float = 0.0
    f1_score: float = 0.0
    auc_roc: float = 0.0

class SystemHealth(BaseModel):
    """System health and performance metrics"""
    api_response_time_ms: float = 0.0
    model_inference_time_s: float = 0.0
    error_rate: float = 0.0
    memory_usage_percent: float = 0.0
    cpu_usage_percent: float = 0.0
    storage_usage_percent: float = 0.0
    uptime_percentage: float = 100.0

class StrokeTypeBreakdown(BaseModel):
    """Breakdown of cases by doctor-classified stroke type"""
    normal_count: int = 0
    ischemia_count: int = 0
    bleeding_count: int = 0
    unclassified_count: int = 0  # Cases where doctor did not provide stroke_type
    normal_percentage: float = 0.0
    ischemia_percentage: float = 0.0
    bleeding_percentage: float = 0.0

class StrokeTypeAccuracy(BaseModel):
    """AI accuracy breakdown per stroke type category"""
    normal_accuracy: float = 0.0  # AI correctly predicted no stroke for doctor-classified normal
    ischemia_accuracy: float = 0.0  # AI correctly predicted stroke for doctor-classified ischemia
    bleeding_accuracy: float = 0.0  # AI correctly predicted stroke for doctor-classified bleeding

class Statistics(BaseModel):
    # Core statistics
    total_predictions: int
    stroke_detected: int
    normal_cases: int
    stroke_detection_rate: float
    average_confidence: float

    # Stroke type breakdown (optional)
    stroke_type_breakdown: Optional[StrokeTypeBreakdown] = None

    # AI accuracy per stroke type (optional)
    stroke_type_accuracy: Optional[StrokeTypeAccuracy] = None

    # Trend data (optional)
    trends: Optional[TrendData] = None

    # Detailed analytics (optional)
    detailed_analytics: Optional[DetailedAnalytics] = None

    # Performance metrics (optional)
    performance_metrics: Optional[PerformanceMetrics] = None

    # System health (optional)
    system_health: Optional[SystemHealth] = None