life-echo/api/scripts/migrate_chapters_to_sections.py

"""
将 chapters 的 content + images 迁移到 chapter_sections，并删除 chapters.content / chapters.images。

前置：已执行 api/migrations/add_chapter_sections.sql（创建 chapter_sections 表、chapters.cover_image 列）。

用法（在 api 目录下）:
    python -m scripts.migrate_chapters_to_sections
"""
import json
import os
import sys
import uuid

sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))

from sqlalchemy import text
from app.core.db import sync_engine as engine
from app.features.memoir.memoir_images.parser import split_narrative_to_sections
from app.core.logging import get_logger, setup_logging

setup_logging()
logger = get_logger(__name__)


def run():
    with engine.connect() as conn:
        # 检查是否还存在 content 列（若已删则跳过）
        r = conn.execute(text("""
            SELECT column_name FROM information_schema.columns
            WHERE table_schema = 'public' AND table_name = 'chapters' AND column_name = 'content'
        """))
        if r.fetchone() is None:
            logger.info("chapters.content 已不存在，跳过迁移")
            return

        # 读取所有有 content 的章节（原始列）
        rows = conn.execute(text("""
            SELECT id, content, images FROM chapters WHERE content IS NOT NULL AND trim(content) != ''
        """)).fetchall()

        for row in rows:
            ch_id, content, images_raw = row[0], row[1], row[2]
            images = json.loads(images_raw) if isinstance(images_raw, str) else (images_raw or [])
            if not isinstance(images, list):
                images = []

            sections = split_narrative_to_sections(content or "")
            if not sections:
                # 无占位符：整段为一条 section，无图
                section_id = str(uuid.uuid4()).replace("-", "")[:32]
                conn.execute(text("""
                    INSERT INTO chapter_sections (id, chapter_id, order_index, content, image, updated_at)
                    VALUES (:id, :ch_id, 0, :content, NULL, NOW())
                """), {"id": section_id, "ch_id": ch_id, "content": (content or "").strip()})
                conn.commit()
                logger.info("章节 %s: 1 条 section（无图）", ch_id)
                continue

            first_cover = None
            img_index = 0
            for order_idx, seg in enumerate(sections):
                section_id = str(uuid.uuid4()).replace("-", "")[:32]
                seg_content = seg.get("content") or ""
                ph = seg.get("placeholder_info")
                image_json = None
                if ph is not None and img_index < len(images):
                    image_json = json.dumps(images[img_index]) if isinstance(images[img_index], dict) else None
                    if first_cover is None and image_json:
                        first_cover = image_json
                    img_index += 1

                conn.execute(text("""
                    INSERT INTO chapter_sections (id, chapter_id, order_index, content, image, updated_at)
                    VALUES (:id, :ch_id, :ord, :content, :img::jsonb, NOW())
                """), {
                    "id": section_id,
                    "ch_id": ch_id,
                    "ord": order_idx,
                    "content": seg_content,
                    "img": image_json,
                })
            if first_cover:
                conn.execute(
                    text("UPDATE chapters SET cover_image = :img::jsonb WHERE id = :id"),
                    {"img": first_cover, "id": ch_id},
                )
            conn.commit()
            logger.info("章节 %s: %d 条 sections", ch_id, len(sections))

        # 删除 chapters.content 和 chapters.images
        conn.execute(text("ALTER TABLE chapters DROP COLUMN IF EXISTS content"))
        conn.execute(text("ALTER TABLE chapters DROP COLUMN IF EXISTS images"))
        conn.commit()
        logger.info("已删除 chapters.content 与 chapters.images")


if __name__ == "__main__":
    run()