diary-news/backend/app/models/article.py

"""文章主表:原文 + 译文 + ML 字段预留。"""
from __future__ import annotations

from datetime import datetime

from sqlalchemy import (
    BigInteger,
    DateTime,
    Float,
    ForeignKey,
    Index,
    Integer,
    String,
    Text,
    func,
)
from sqlalchemy.dialects.postgresql import JSONB
from sqlalchemy.orm import Mapped, mapped_column, relationship

from app.database import Base


class Article(Base):
    __tablename__ = "articles"

    id: Mapped[int] = mapped_column(BigInteger, primary_key=True)

    # === 来源 ===
    source_id: Mapped[int] = mapped_column(
        ForeignKey("sources.id", ondelete="CASCADE"), nullable=False, index=True
    )
    source: Mapped["Source"] = relationship(back_populates="articles", lazy="joined")  # noqa: F821

    # === 原文标识 ===
    url: Mapped[str] = mapped_column(Text, nullable=False)
    url_hash: Mapped[str] = mapped_column(String(40), unique=True, nullable=False, index=True)
    guid: Mapped[str | None] = mapped_column(String(255), index=True)  # 源站给的 ID

    # === 原文内容 ===
    title: Mapped[str] = mapped_column(Text, nullable=False)
    body_html: Mapped[str | None] = mapped_column(Text)  # 抽取后保留结构
    body_text: Mapped[str] = mapped_column(Text, nullable=False, default="")
    lang_src: Mapped[str | None] = mapped_column(String(8))
    author: Mapped[str | None] = mapped_column(String(255))
    image_url: Mapped[str | None] = mapped_column(Text)

    # === 译文 ===
    title_zh: Mapped[str | None] = mapped_column(Text)
    body_zh_html: Mapped[str | None] = mapped_column(Text)
    body_zh_text: Mapped[str | None] = mapped_column(Text)
    body_zh_formatted: Mapped[str | None] = mapped_column(Text)  # LLM 排版后
    summary_zh: Mapped[str | None] = mapped_column(Text)

    # === 翻译状态 ===
    translation_status: Mapped[str] = mapped_column(
        String(16), default="pending", nullable=False, index=True
    )
    # pending / ok / partial / failed / n/a
    translation_engine: Mapped[str | None] = mapped_column(String(16))
    # tencent / nllb / cache / skip
    translation_chars: Mapped[int] = mapped_column(Integer, default=0, nullable=False)
    translated_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True))

    # === LLM 增强状态(每个独立状态)===
    format_status: Mapped[str] = mapped_column(
        String(16), default="n/a", nullable=False
    )  # pending/ok/failed/n/a
    classify_status: Mapped[str] = mapped_column(
        String(16), default="n/a", nullable=False
    )
    image_ai_status: Mapped[str] = mapped_column(
        String(16), default="n/a", nullable=False
    )
    commentary_status: Mapped[str] = mapped_column(
        String(16), default="n/a", nullable=False
    )

    # === LLM 增强内容 ===
    image_ai_url: Mapped[str | None] = mapped_column(Text)  # AI 生成的插图

    # === ML 字段(预留,MVP 全 null)===
    category: Mapped[str | None] = mapped_column(String(32), index=True)
    commentary: Mapped[str | None] = mapped_column(Text)
    entities: Mapped[dict | None] = mapped_column(JSONB)
    sentiment: Mapped[float | None] = mapped_column(Float)
    topic_id: Mapped[str | None] = mapped_column(String(64), index=True)
    bias: Mapped[str | None] = mapped_column(String(16))  # left/center/right

    # === 去重 ===
    duplicate_of: Mapped[int | None] = mapped_column(
        ForeignKey("articles.id", ondelete="SET NULL"), index=True
    )

    # === 时间 ===
    published_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), index=True)
    fetched_at: Mapped[datetime] = mapped_column(
        DateTime(timezone=True), server_default=func.now(), nullable=False, index=True
    )
    created_at: Mapped[datetime] = mapped_column(
        DateTime(timezone=True), server_default=func.now(), nullable=False
    )

    __table_args__ = (
        Index("ix_articles_source_published", "source_id", "published_at"),
        Index("ix_articles_status_published", "translation_status", "published_at"),
    )

    def __repr__(self) -> str:
        return f"<Article id={self.id} src={self.source_id} status={self.translation_status}>"