from __future__ import annotations

import uuid
from datetime import datetime
from typing import Any

from sqlalchemy import JSON, DateTime, Float, ForeignKey, Integer, String, Text
from sqlalchemy.orm import Mapped, mapped_column

from app.db.base import Base, TimestampMixin, UUIDPrimaryKeyMixin


class SourceDocument(Base, UUIDPrimaryKeyMixin, TimestampMixin):
    """A single piece of collected content. Kept lean on purpose - raw HTML
    is not stored, only extracted text - per the "avoid saving unnecessary
    full HTML indefinitely" rule in the spec."""

    __tablename__ = "source_documents"

    source_id: Mapped[uuid.UUID] = mapped_column(
        ForeignKey("sources.id", ondelete="CASCADE"), index=True
    )
    company_id: Mapped[uuid.UUID] = mapped_column(
        ForeignKey("companies.id", ondelete="CASCADE"), index=True
    )
    url: Mapped[str] = mapped_column(String(1000))
    canonical_url: Mapped[str] = mapped_column(String(1000), index=True)
    title: Mapped[str | None] = mapped_column(String(500), nullable=True)
    author: Mapped[str | None] = mapped_column(String(200), nullable=True)
    publication_date: Mapped[datetime | None] = mapped_column(
        DateTime(timezone=True), nullable=True
    )
    retrieved_date: Mapped[datetime] = mapped_column(DateTime(timezone=True))
    content_text: Mapped[str] = mapped_column(Text)
    content_hash: Mapped[str] = mapped_column(String(64), index=True)
    metadata_json: Mapped[dict[str, Any]] = mapped_column(JSON, default=dict)
    language: Mapped[str | None] = mapped_column(String(16), nullable=True)
    http_status: Mapped[int | None] = mapped_column(Integer, nullable=True)
    extraction_method: Mapped[str] = mapped_column(String(50))
    trust_score: Mapped[float] = mapped_column(Float, default=0.7)