import json
import re
from collections import defaultdict


def find_duplicate_devices_unique_mac(log_file_path):
    # 按 deviceSn 分組，每組內(nèi)用集合去重 deviceMac
    devices_by_sn = defaultdict(list)
    mac_seen_by_sn = defaultdict(set)  # 用于跟蹤每個(gè) SN 下已見過的 deviceMac

    # 讀取日志文件
    with open(log_file_path, 'r', encoding='utf-8') as file:
        for line_num, line in enumerate(file, 1):
            # 使用正則表達(dá)式提取 JSON 部分
            match = re.search(r'\[.*\]', line)
            if match:
                try:
                    # 解析 JSON 數(shù)組
                    devices = json.loads(match.group())

                    # 將每個(gè)設(shè)備添加到對應(yīng) deviceSn 的分組中，并去重 deviceMac
                    for device in devices:
                        device_sn = device.get('deviceSn')
                        device_mac = device.get('deviceMac')

                        if device_sn and device_mac:
                            # 如果這個(gè) SN 下還沒見過這個(gè) MAC，則添加
                            if device_mac not in mac_seen_by_sn[device_sn]:
                                devices_by_sn[device_sn].append(device)
                                mac_seen_by_sn[device_sn].add(device_mac)
                except json.JSONDecodeError as e:
                    print(f"第 {line_num} 行解析 JSON 時(shí)出錯(cuò): {e}")
                    continue

    # 找出重復(fù)的 deviceSn（去重 MAC 后仍然有多個(gè)記錄的）
    duplicate_devices = {}
    for device_sn, devices in devices_by_sn.items():
        if len(devices) > 1:
            duplicate_devices[device_sn] = devices

    return duplicate_devices


def print_duplicate_devices(duplicate_devices):
    if not duplicate_devices:
        print("沒有找到重復(fù)的 deviceSn")
        return

    print("找到以下重復(fù)的 deviceSn (已對 deviceMac 去重):\n")
    for device_sn, devices in duplicate_devices.items():
        print(f"deviceSn: {device_sn} (去重后出現(xiàn) {len(devices)} 次)")
        print("-" * 50)

        for i, device in enumerate(devices, 1):
            print(f"第 {i} 條記錄:")
            # 美化輸出 JSON
            print(json.dumps(device, indent=2, ensure_ascii=False))
            print()

        print("=" * 80)


# 版本2：更簡潔的實(shí)現(xiàn)，直接輸出去重結(jié)果
def find_and_print_duplicates_unique(log_file):
    # 存儲(chǔ)去重后的設(shè)備
    unique_devices_by_sn = defaultdict(list)
    seen_mac_by_sn = defaultdict(set)

    with open(log_file, 'r') as f:
        for line in f:
            # 提取 JSON 數(shù)組部分
            json_match = re.search(r'\[.*\]', line)
            if json_match:
                try:
                    devices = json.loads(json_match.group())
                    for device in devices:
                        sn = device.get('deviceSn')
                        mac = device.get('deviceMac')

                        if sn and mac:
                            # 如果這個(gè) MAC 還沒在這個(gè) SN 組中出現(xiàn)過
                            if mac not in seen_mac_by_sn[sn]:
                                unique_devices_by_sn[sn].append(device)
                                seen_mac_by_sn[sn].add(mac)
                except:
                    continue

    # 找出并打印重復(fù)項(xiàng)
    print("重復(fù)的設(shè)備SN及其數(shù)據(jù) (已對deviceMac去重):")
    print("=" * 80)

    found_duplicates = False
    for sn, devices in unique_devices_by_sn.items():
        if len(devices) > 1:
            found_duplicates = True
            print(f"\n設(shè)備SN: {sn} (去重后出現(xiàn) {len(devices)} 次)")
            print("-" * 50)

            for i, device in enumerate(devices, 1):
                print(f"記錄 {i}:")
                # 格式化時(shí)間戳
                if 'productionDate' in device:
                    import datetime
                    timestamp = device['productionDate'] / 1000
                    dt = datetime.datetime.fromtimestamp(timestamp)
                    device['productionDate_formatted'] = dt.strftime('%Y-%m-%d %H:%M:%S')

                print(json.dumps(device, indent=2, ensure_ascii=False))
                print()

    if not found_duplicates:
        print("沒有找到重復(fù)的 deviceSn (或所有重復(fù)都是相同的 deviceMac)")


# 主程序
if __name__ == "__main__":
    log_file_path = "syncDevice_2026-01-07.log"

    try:
        print("=" * 80)
        print("方法1：詳細(xì)版")
        print("=" * 80)
        # 查找重復(fù)設(shè)備（去重 MAC）
        duplicate_devices = find_duplicate_devices_unique_mac(log_file_path)

        # 打印結(jié)果
        print_duplicate_devices(duplicate_devices)

        # 統(tǒng)計(jì)信息
        print("\n統(tǒng)計(jì)信息:")
        print(f"總共有 {len(duplicate_devices)} 個(gè)重復(fù)的 deviceSn")
        for device_sn, devices in duplicate_devices.items():
            print(f"  - {device_sn}: {len(devices)} 條不重復(fù)的記錄")

        print("\n" + "=" * 80)
        print("方法2：簡潔版")
        print("=" * 80)
        find_and_print_duplicates_unique(log_file_path)

    except FileNotFoundError:
        print(f"錯(cuò)誤: 找不到文件 {log_file_path}")
    except Exception as e:
        print(f"處理文件時(shí)出錯(cuò): {e}")