server-core/scripts/recover_from_binlog.py

#!/usr/bin/env python3
"""从 binlog 解码文件中恢复被误删的数据（DELETE → INSERT）"""
import re
import asyncio
import asyncmy

BINLOG_FILE = "/tmp/binlog_decode.sql"

# 只恢复这些表（非政策/活动/任务/岗位/社区内容）
RECOVER_TABLES = {
    "opc_services", "service_orders", "service_views", "service_referrals",
    "service_providers", "service_favorites", "service_inquiries", "service_reviews",
    "market_items", "market_categories", "market_purchase_orders", "market_purchases",
    "finance_records", "settlement_logs", "invoices", "contracts", "payment_bindings",
    "compute_recharge_orders", "compute_recharges", "compute_usage_records",
    "compute_balance_allocations", "investment_intents", "investor_preferences",
    "courses", "training_courses", "training_chapters", "training_lessons",
    "training_enrollments", "training_course_enrollments", "training_lesson_progress",
    "training_certificates", "skill_tests", "certifications", "certification_types",
    "opc_certifications",
    "credit_ledger", "credit_scores", "credit_dimensions", "credit_levels",
    "credit_rules", "badges", "user_badges",
    "messages", "notifications", "announcements", "survey_logs", "plan_logs",
    "invite_records", "ratings",
    "park_admissions", "company_apply_requests", "company_claim_requests",
}


def parse_binlog(path):
    """解析解码后的 binlog，返回 {table: [row_dict, ...]}"""
    results = {}
    current_table = None
    current_row = {}
    in_where = False

    with open(path, "r", encoding="utf-8", errors="replace") as f:
        for line in f:
            line = line.strip()
            # 匹配 DELETE FROM
            m = re.match(r"### DELETE FROM `opc`\.`(\w+)`", line)
            if m:
                # 保存上一行
                if current_table and current_row:
                    results.setdefault(current_table, []).append(current_row)
                current_table = m.group(1)
                current_row = {}
                in_where = False
                continue
            if line == "### WHERE":
                in_where = True
                continue
            if in_where and line.startswith("###   @"):
                # 解析 @N=value
                m2 = re.match(r"###   @(\d+)=(.*)", line)
                if m2:
                    idx = int(m2.group(1))
                    val = m2.group(2)
                    # 处理 NULL
                    if val == "NULL":
                        val = None
                    else:
                        # 去掉引号（简单处理）
                        if val.startswith("'") and val.endswith("'"):
                            val = val[1:-1]
                            # 处理转义
                            val = val.replace("\\'", "'").replace("\\\\", "\\")
                    current_row[idx] = val
                continue
            # 遇到其他行，结束当前事件
            if current_table and current_row and not line.startswith("###"):
                results.setdefault(current_table, []).append(current_row)
                current_table = None
                current_row = {}
                in_where = False

        # 最后一个
        if current_table and current_row:
            results.setdefault(current_table, []).append(current_row)

    return results


async def get_columns(conn, table):
    """获取表的列名顺序"""
    async with conn.cursor() as cur:
        await cur.execute(
            """SELECT COLUMN_NAME FROM INFORMATION_SCHEMA.COLUMNS
               WHERE TABLE_SCHEMA='opc' AND TABLE_NAME=%s ORDER BY ORDINAL_POSITION""",
            (table,))
        return [r[0] for r in await cur.fetchall()]


async def main():
    print("解析 binlog...")
    data = parse_binlog(BINLOG_FILE)
    print(f"解析到 {len(data)} 个表的删除事件")

    conn = await asyncmy.connect(
        host="47.108.226.213", port=8091,
        user="root", password="sjnxhyjashaywiuwhaja", database="opc")

    total_recovered = 0
    for table, rows in data.items():
        if table not in RECOVER_TABLES:
            continue
        if not rows:
            continue

        cols = await get_columns(conn, table)
        if not cols:
            print(f"  ⚠ {table}: 无法获取列名")
            continue

        # 过滤掉已经存在的记录（通过主键判断）
        # 先查现有主键
        pk_col = cols[0]  # 假设第一列是主键
        existing_ids = set()
        try:
            async with conn.cursor() as cur:
                await cur.execute(f"SELECT `{pk_col}` FROM `{table}`")
                existing_ids = {str(r[0]) for r in await cur.fetchall()}
        except Exception:
            pass

        inserted = 0
        skipped = 0
        async with conn.cursor() as cur:
            for row in rows:
                # 按列索引映射
                values = []
                for i, col in enumerate(cols, 1):
                    val = row.get(i)
                    values.append(val)

                # 跳过已存在的
                pk_val = str(values[0]) if values[0] is not None else None
                if pk_val and pk_val in existing_ids:
                    skipped += 1
                    continue

                placeholders = ",".join(["%s"] * len(cols))
                col_names = ",".join([f"`{c}`" for c in cols])
                try:
                    await cur.execute(
                        f"INSERT INTO `{table}` ({col_names}) VALUES ({placeholders})",
                        values)
                    inserted += 1
                    if pk_val:
                        existing_ids.add(pk_val)
                except Exception as e:
                    print(f"  ✗ {table}: {e}")
                    skipped += 1

        await conn.commit()
        total_recovered += inserted
        print(f"  ✓ {table}: 恢复 {inserted} 条（跳过 {skipped} 条已存在）")

    await conn.ensure_closed()
    print(f"\n总计恢复: {total_recovered} 条")


if __name__ == "__main__":
    asyncio.run(main())