#!/usr/bin/env python3
"""瞬间迁移（Phase 2 T4）：Halo moments → content/moments/*.md

Halo moment.content.html 为富文本（内联样式很脏），经 pandoc 提纯为
纯文本 md；medium（配图）当前库存为空，字段留待将来有图时扩展。
"""
import json
import subprocess
from datetime import datetime, timezone
from pathlib import Path

OUT = Path("content/moments")


def main() -> int:
    moments = json.load(open(".migration/moments.json", encoding="utf-8"))
    n = 0
    for m in moments:
        spec = m["spec"]
        if spec.get("visible") == "PRIVATE":
            continue
        html = (spec.get("content") or {}).get("html", "")
        text = subprocess.run(
            ["/home/wait4x/.local/bin/pandoc", "-f", "html", "-t", "plain", "--wrap=none"],
            input=html.encode(), capture_output=True, check=True,
        ).stdout.decode("utf-8").strip()
        if not text:
            continue
        ts = spec.get("releaseTime") or m["metadata"].get("creationTimestamp", "")
        day = datetime.fromisoformat(ts.replace("Z", "+00:00")).astimezone(timezone.utc).strftime("%Y-%m-%d")
        out = OUT / f"{day}.md"
        i = 2
        while out.exists():
            out = OUT / f"{day}-{i}.md"
            i += 1
        tags = spec.get("tags") or []
        # schema 约束 text/image 在 frontmatter（body 不参与校验），一句话即 text 字段
        out.write_text(
            "---\n"
            f"pubDate: {day}\n"
            f"tags: [{', '.join(tags)}]\n"
            f'text: "{text.replace(chr(34), chr(39))}"\n'
            "---\n",
            encoding="utf-8",
        )
        n += 1
    print(f"migrated {n} moments")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())