guanghulab/scripts/sync-to-sg.py

#!/usr/bin/env python3
"""
光湖直连同步器 · 本地→新加坡
绕过代理，直连新加坡 Gatekeeper，分 chunk 传输文件到代码仓库。
用法: python3 sync-to-sg.py --file brain/xxx.hdlp --target brain/xxx.hdlp --msg "feat: xxx"
"""

import argparse
import base64
import json
import os
import subprocess
import sys
import time

GK_URL = "http://43.156.237.110:3911/exec"
GK_KEY = "zy_gtw_c38be341d65043eee0ba51a214a267832a9ae46dd73e423c"
REPO_PATH = "/opt/zhuyuan/guanghulab"
CHUNK_SIZE = 6000  # bytes，新加坡 gatekeeper 安全上限
CMD_TIMEOUT = 30   # 每条命令超时秒数


def gk_exec(cmd: str) -> dict:
    """通过 Gatekeeper 执行命令，绕过代理直连"""
    env = os.environ.copy()
    env.pop("http_proxy", None)
    env.pop("https_proxy", None)
    env.pop("HTTP_PROXY", None)
    env.pop("HTTPS_PROXY", None)

    result = subprocess.run(
        ["curl", "-s", "--connect-timeout", str(CMD_TIMEOUT),
         "-X", "POST", GK_URL,
         "-H", "Content-Type: application/json",
         "-H", f"Authorization: Bearer {GK_KEY}",
         "-d", json.dumps({"cmd": cmd})],
        capture_output=True, text=True, timeout=CMD_TIMEOUT + 5, env=env
    )
    try:
        return json.loads(result.stdout)
    except json.JSONDecodeError:
        return {"ok": False, "error": f"gatekeeper 返回非 JSON: {result.stdout[:200]}", "stderr": result.stderr}


def sync_file(local_path: str, target_path: str, commit_msg: str) -> bool:
    """同步单个文件到新加坡仓库"""
    # 读文件
    with open(local_path, "rb") as f:
        content = f.read()
    expected_bytes = len(content)
    b64 = base64.b64encode(content).decode("ascii")
    print(f"[sync] {local_path} → {target_path}  ({expected_bytes} bytes, {len(b64)} b64 chars)")

    # 清空临时目录
    r = gk_exec(f"rm -f /tmp/sync_chunk_* /tmp/sync_b64.txt")
    if not r.get("ok"):
        print(f"[error] 清空临时目录失败: {r}")
        return False

    # 分 chunk 上传
    chunks = [b64[i:i+CHUNK_SIZE] for i in range(0, len(b64), CHUNK_SIZE)]
    print(f"[sync] 分 {len(chunks)} chunks 上传...")

    for i, chunk in enumerate(chunks):
        # base64 编码 chunk 内容避免 shell 转义问题
        chunk_b64 = base64.b64encode(chunk.encode()).decode()
        cmd = f'echo "{chunk_b64}" | base64 -d >> /tmp/sync_b64.txt'
        r = gk_exec(cmd)
        if not r.get("ok"):
            print(f"[error] chunk {i+1}/{len(chunks)} 上传失败: {r}")
            return False
        print(f"  chunk {i+1}/{len(chunks)} ✓")

    # 解码并写入
    full_target = f"{REPO_PATH}/{target_path}"
    target_dir = os.path.dirname(full_target)
    write_cmd = (
        f"mkdir -p {target_dir} && "
        f"base64 -d /tmp/sync_b64.txt > {full_target} && "
        f"echo BYTES=$(wc -c < {full_target})"
    )
    r = gk_exec(write_cmd)
    if not r.get("ok"):
        print(f"[error] 写入失败: {r}")
        return False

    # 校验字节数
    stdout = r.get("stdout", "")
    if f"BYTES={expected_bytes}" not in stdout:
        actual = stdout.replace("BYTES=", "").strip()
        print(f"[error] 字节校验失败: 期望={expected_bytes}, 实际={actual}")
        return False
    print(f"[sync] 写入+校验通过 ({expected_bytes} bytes)")

    # Git 提交（pull→commit→push）
    escaped_msg = commit_msg.replace('"', '\\"').replace("'", "\\'")
    git_cmd = (
        f"cd {REPO_PATH} && "
        f"git config user.email zhuyuan@guanghulab.com && "
        f"git config user.name '铸渊 ICE-GL-ZY001' && "
        f"git pull --rebase origin main 2>&1; "
        f"git add {target_path} && "
        f"GIT_OUT=$(git commit -m \"{escaped_msg}\" 2>&1); echo \"COMMIT: $GIT_OUT\"; "
        f"if echo \"$GIT_OUT\" | grep -q 'nothing to commit'; then echo 'NO_CHANGE'; else git push origin main 2>&1; fi"
    )
    r = gk_exec(git_cmd)
    stdout = r.get("stdout", "")
    stderr = r.get("stderr", "")

    if "NO_CHANGE" in stdout:
        print(f"[sync] 文件无变更，跳过 commit+push")
    elif r.get("ok"):
        print(f"[sync] git pull + commit + push 成功")
        if stdout:
            print(f"  输出: {stdout[:300]}")
    else:
        # git 可能返回非零但实际成功了（committer 警告等）
        if "COMMIT:" in stdout and "nothing to commit" not in stdout:
            print(f"[sync] git 操作完成（有警告但 commit 已执行）")
        else:
            print(f"[error] git 操作失败: stdout={stdout[:300]}, stderr={stderr[:200]}")
            return False

    # 清理
    gk_exec("rm -f /tmp/sync_chunk_* /tmp/sync_b64.txt")
    return True


def main():
    parser = argparse.ArgumentParser(description="光湖直连同步器 · 本地→新加坡")
    parser.add_argument("--file", required=True, help="本地文件路径")
    parser.add_argument("--target", required=True, help="目标路径(相对于仓库根)")
    parser.add_argument("--msg", required=True, help="commit message")
    args = parser.parse_args()

    if not os.path.exists(args.file):
        print(f"[error] 文件不存在: {args.file}")
        sys.exit(1)

    print(f"═" * 50)
    print(f"光湖直连同步器 · 本地→新加坡 BS-SG-001")
    print(f"═" * 50)

    start = time.time()
    ok = sync_file(args.file, args.target, args.msg)
    elapsed = time.time() - start

    if ok:
        print(f"✅ 同步完成 ({elapsed:.1f}s)")
    else:
        print(f"❌ 同步失败 ({elapsed:.1f}s)")
        sys.exit(1)


if __name__ == "__main__":
    main()