inspiration-collector/scripts/sync_cards.py

#!/usr/bin/env python3
"""
sync_cards.py — Memos 卡片备份脚本（基于 uid 去重）

每30分钟由 cron 触发，将当前用户的所有 Memos 卡片备份到 cards/ 目录，
并推送到 Gitea inspiration-collector 仓库。
去重方式：按 memo.uid 比对，不重复写入。
"""

import os
import sys
import re
import subprocess
from datetime import datetime, timezone, timedelta

import requests

# ── 配置 ──
BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
CARDS_DIR = os.path.join(BASE_DIR, "cards")
GIT_DIR = BASE_DIR

MEMOS_URL = os.environ.get("MEMOS_URL", "http://localhost:5230")
MEMOS_TOKEN = os.environ.get("MEMOS_ACCESS_TOKEN", "")

TZ_BEIJING = timezone(timedelta(hours=8))

KNOWN_TAGS = {"#功能完善", "#90后四级副职培训班", "#灵感收集器"}


class MemosClient:
    def __init__(self, base_url, token):
        self.base_url = base_url.rstrip("/")
        self.session = requests.Session()
        self.session.headers.update({
            "Authorization": f"Bearer {token}",
            "Content-Type": "application/json",
        })

    def list_all_memos(self):
        all_memos = []
        page_token = ""
        while True:
            params = {"pageSize": 100}
            if page_token:
                params["pageToken"] = page_token
            resp = self.session.post(
                f"{self.base_url}/memos.api.v1.MemoService/ListMemos",
                json=params, timeout=15
            )
            resp.raise_for_status()
            data = resp.json()
            all_memos.extend(data.get("memos", []))
            next_token = data.get("nextPageToken", "")
            if not next_token:
                break
            page_token = next_token
        return all_memos


def sanitize_filename(text):
    text = re.sub(r'[<>:"/\\|?*]', "", text)
    text = re.sub(r'\s+', "_", text.strip())
    return text[:80] or "memo"


def extract_tags(content):
    """提取内容中的 #标签 """
    return [w for w in content.split() if w.startswith("#") and w in KNOWN_TAGS]


def load_existing_uids(cards_dir):
    """扫描现有卡片文件，收集所有 uid + 文件名作为去重依据"""
    uids = set()
    fnames = set()
    if not os.path.isdir(cards_dir):
        return uids, fnames
    for fname in os.listdir(cards_dir):
        if not fname.endswith(".md"):
            continue
        fnames.add(fname)
        fpath = os.path.join(cards_dir, fname)
        try:
            with open(fpath, "r", encoding="utf-8") as f:
                head = f.read(500)
            # 匹配 uid: "xxx" 或 id: xxx 或 uid: xxx（无引号）
            m = re.search(r'^(?:uid|id):\s*"?([^"\n]+)"?', head, re.MULTILINE)
            if m:
                val = m.group(1).strip()
                if val:
                    uids.add(val)
        except Exception:
            continue
    return uids, fnames


def memo_to_markdown(memo):
    uid = memo.get("uid", "")
    content = memo.get("content", "")
    create_time = memo.get("createTime", "")

    try:
        dt = datetime.fromisoformat(create_time.replace("Z", "+00:00"))
        dt_bj = dt.astimezone(TZ_BEIJING)
        time_str = dt_bj.strftime("%Y-%m-%dT%H:%M:%S+08:00")
        file_ts = dt_bj.strftime("%Y%m%d%H%M%S")
    except Exception:
        time_str = create_time
        file_ts = uid[:14] if uid else "unknown"

    first_line = content.strip().split("\n")[0][:60]
    filename = f"{file_ts}_{sanitize_filename(first_line)}.md"
    filepath = os.path.join(CARDS_DIR, filename)

    md = f"""---
title: "{first_line}"
date: {time_str}
uid: "{uid}"
source: "memos"
---

{content}
"""
    return filepath, md


def git_push(directory, message):
    try:
        subprocess.run(["git", "add", "-A"], cwd=directory, check=True,
                       capture_output=True, timeout=30)
        result = subprocess.run(
            ["git", "diff", "--cached", "--quiet"],
            cwd=directory, capture_output=True, timeout=15
        )
        if result.returncode == 0:
            return
        subprocess.run(
            ["git", "commit", "-m", message],
            cwd=directory, check=True, capture_output=True, timeout=30
        )
        subprocess.run(
            ["git", "push", "origin", "main"],
            cwd=directory, check=True, capture_output=True, timeout=60
        )
        print(f"[OK] 已推送: {message}")
    except subprocess.CalledProcessError as e:
        print(f"[WARN] git 操作失败: {e.stderr.decode()[:200]}")
    except Exception as e:
        print(f"[WARN] git 异常: {e}")


def main():
    if not MEMOS_TOKEN:
        print("[ERROR] MEMOS_ACCESS_TOKEN 未设置")
        sys.exit(1)

    os.makedirs(CARDS_DIR, exist_ok=True)

    # 加载已有 uid + 文件名
    existing_uids, existing_fnames = load_existing_uids(CARDS_DIR)
    print(f"[INFO] 已有 {len(existing_fnames)} 个文件，{len(existing_uids)} 个 uid")

    client = MemosClient(MEMOS_URL, MEMOS_TOKEN)
    print(f"[INFO] 正在从 Memos 拉取...")
    all_memos = client.list_all_memos()
    print(f"[INFO] 共获取 {len(all_memos)} 条卡片")

    new_count = 0
    for memo in all_memos:
        uid = memo.get("uid", "")
        if uid and uid in existing_uids:
            continue  # uid 已存在，跳过
        filepath, md = memo_to_markdown(memo)
        fname = os.path.basename(filepath)
        if fname in existing_fnames:
            continue  # 文件名已存在，跳过（兼容旧版无 uid 的卡片）
        with open(filepath, "w", encoding="utf-8") as f:
            f.write(md)
        new_count += 1
        print(f"  [+] {fname}")
        if uid:
            existing_uids.add(uid)
        existing_fnames.add(fname)

    if new_count > 0:
        print(f"[INFO] 新增 {new_count} 条卡片，推送至 Gitea...")
        git_push(GIT_DIR, f"cards: 备份 {new_count} 条 Memos 卡片")
    else:
        print(f"[INFO] 无新增卡片")


if __name__ == "__main__":
    main()