#!/usr/bin/env python3
"""Halo DB 只读导出（Phase 2 T1 补充源）：posts + snapshots → .migration/。

公共 API 只有摘要、head-content 404；正文真源在 postgres extensions 表：
  /registry/blog.halo.run/posts/{name}          → spec.headSnapshot 指针
  /registry/content.halo.run/snapshots/{uuid}    → raw/content 全量
data 列是 bytea，经 base64 通道取出后 json 解析（psql 文本输出会搅转义）。
"""
import base64
import json
import subprocess
import sys
from pathlib import Path

OUT = Path(".migration")


def psql_b64(name: str) -> dict:
    sql = "SELECT encode(data,'base64') FROM extensions WHERE name = '%s';" % name
    raw = subprocess.run(
        ["docker", "exec", "postgres-db", "psql", "-U", "wait4x", "-d", "halo_db", "-t", "-A", "-c", sql],
        capture_output=True, text=True, check=True,
    ).stdout.strip()
    return json.loads(base64.b64decode(raw))


def names_like(pattern: str) -> list[str]:
    sql = "SELECT name FROM extensions WHERE name LIKE '%s';" % pattern
    raw = subprocess.run(
        ["docker", "exec", "postgres-db", "psql", "-U", "wait4x", "-d", "halo_db", "-t", "-A", "-c", sql],
        capture_output=True, text=True, check=True,
    ).stdout.split()
    return raw


def main() -> int:
    posts = [psql_b64(n) for n in names_like("%/posts/%")]
    snaps = {n.rsplit("/", 1)[1]: psql_b64(n) for n in names_like("%/snapshots/%")}

    # 每篇 post 挂上 head snapshot 的 raw/content 全文
    # 结构：snapshot.spec.{rawPatch,contentPatch} 直接是字符串（实测解剖）
    for p in posts:
        spec = p.get("spec", {})
        # head 与 release 各挂一个 snapshot；实测存在"head 空、release 满"与
        # 反例（mian-sha），取 content 更长者作为迁移源
        cands = [snaps.get(k, {}).get("spec", {}) for k in (spec.get("releaseSnapshot"), spec.get("headSnapshot")) if k]
        snap = max(cands, key=lambda sp: len(sp.get("contentPatch", "") or ""), default={})
        p["_snapshot"] = {
            "rawType": snap.get("rawType"),
            "raw": snap.get("rawPatch", ""),
            "content": snap.get("contentPatch", ""),
            "snapshotCount": len(cands),
        }

    (OUT / "posts_db.json").write_text(json.dumps(posts, ensure_ascii=False, indent=1))
    ok = sum(1 for p in posts if p["_snapshot"]["content"])
    print(f"db export: posts={len(posts)}（含正文 {ok}）；snapshots={len(snaps)}")
    return 0


if __name__ == "__main__":
    sys.exit(main())