"""SQLAlchemy models for knowledge graph storage."""

from sqlalchemy import Column, String, Text, Float, DateTime, ARRAY, ForeignKey, UniqueConstraint
from sqlalchemy.dialects.postgresql import UUID, VECTOR
from sqlalchemy.orm import declarative_base
from sqlalchemy.sql import func
import uuid
from datetime import datetime

Base = declarative_base()


class Entity(Base):
    """Knowledge graph entity."""
    __tablename__ = "entities"

    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
    domain = Column(String(100), nullable=False, index=True)
    name = Column(String(500), nullable=False)
    description = Column(Text)
    entity_type = Column(String(100), nullable=False)  # transceiver, standard, vendor, etc
    embedding = Column(VECTOR(384))  # bge-m3 384-dim
    confidence = Column(Float, default=1.0)
    metadata = Column(String)  # JSON metadata
    created_at = Column(DateTime, default=datetime.utcnow)
    updated_at = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)

    __table_args__ = (
        UniqueConstraint('domain', 'entity_type', 'name', name='unique_entity'),
    )


class Relation(Base):
    """Knowledge graph relation between entities."""
    __tablename__ = "relations"

    source_id = Column(UUID(as_uuid=True), ForeignKey("entities.id"), primary_key=True)
    relation_type = Column(String(100), primary_key=True)  # supported_by, manufactured_by, etc
    target_id = Column(UUID(as_uuid=True), ForeignKey("entities.id"), primary_key=True)
    strength = Column(Float, default=1.0)  # confidence in relation
    metadata = Column(String)  # JSON metadata
    created_at = Column(DateTime, default=datetime.utcnow)


class Document(Base):
    """Ingested document for knowledge graph."""
    __tablename__ = "documents"

    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
    domain = Column(String(100), nullable=False, index=True)
    source = Column(String(100), nullable=False)  # blog, datasheet, standard, etc
    title = Column(String(500), nullable=False)
    content = Column(Text, nullable=False)
    entity_ids = Column(ARRAY(UUID(as_uuid=True)))  # linked entity IDs
    embedding = Column(VECTOR(384))  # Document-level embedding
    token_count = Column(Float)
    created_at = Column(DateTime, default=datetime.utcnow)


class QueryLog(Base):
    """Query execution audit trail for evaluation."""
    __tablename__ = "query_logs"

    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
    domain = Column(String(100), nullable=False, index=True)
    query_text = Column(Text, nullable=False)
    retrieved_doc_ids = Column(ARRAY(UUID(as_uuid=True)))
    ground_truth_doc_ids = Column(ARRAY(UUID(as_uuid=True)))
    relevance_scores = Column(ARRAY(Float))
    latency_ms = Column(Float)
    entity_count = Column(Float)
    created_at = Column(DateTime, default=datetime.utcnow)


class EvaluationResult(Base):
    """Evaluation metrics snapshot."""
    __tablename__ = "evaluation_results"

    id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
    domain = Column(String(100), nullable=False, index=True)
    eval_set_name = Column(String(100), nullable=False)
    metric_name = Column(String(100), nullable=False)
    metric_value = Column(Float, nullable=False)
    baseline_value = Column(Float)  # FTS baseline for comparison
    improvement_pct = Column(Float)
    sample_count = Column(Float)
    created_at = Column(DateTime, default=datetime.utcnow)