如何用Python将Broadcom FDMI设备日志转为按Port分组的JSON？

问题背景 #

我正在处理Broadcom设备生成的FDMI日志（由fdmiShow命令导出），现有Python解析代码输出的JSON将Node Name、Port Name、Serial Number、Model等属性分散存储，无法按每个Port块进行分组。需要修改代码，将每个Port对应的所有关联属性整合到同一个对象中，生成结构清晰的JSON输出。

日志示例（fdmiShow输出片段） #

FDMI Device Database:
-------------------------------------------------------------------------------
Node Name: switch-01
Port Name: 0
Serial Number: ABC123
Model: BRCM-57800
-------------------------------------------------------------------------------
Node Name: switch-01
Port Name: 1
Serial Number: ABC123
Model: BRCM-57800
-------------------------------------------------------------------------------
Node Name: storage-02
Port Name: 0
Serial Number: XYZ789
Model: BRCM-9200

当前解析代码（问题版本） #

import re
import json

def parse_fdmi_log(log_path):
    with open(log_path, 'r') as f:
        log_content = f.read()
    
    # 原代码错误地将属性拆分到不同集合中
    node_names = re.findall(r'Node Name: (.*)', log_content)
    port_names = re.findall(r'Port Name: (.*)', log_content)
    serial_numbers = re.findall(r'Serial Number: (.*)', log_content)
    models = re.findall(r'Model: (.*)', log_content)
    
    # 分散存储属性，无法关联到单个Port
    result = {
        "node_names": [n.strip() for n in node_names],
        "port_names": [p.strip() for p in port_names],
        "serial_numbers": [s.strip() for s in serial_numbers],
        "models": [m.strip() for m in models]
    }
    
    return json.dumps(result, indent=2)

现有输出（分散结构） #

{
  "node_names": [
    "switch-01",
    "switch-01",
    "storage-02"
  ],
  "port_names": [
    "0",
    "1",
    "0"
  ],
  "serial_numbers": [
    "ABC123",
    "ABC123",
    "XYZ789"
  ],
  "models": [
    "BRCM-57800",
    "BRCM-57800",
    "BRCM-9200"
  ]
}

期望输出（按Port分组） #

[
  {
    "node_name": "switch-01",
    "port_name": "0",
    "serial_number": "ABC123",
    "model": "BRCM-57800"
  },
  {
    "node_name": "switch-01",
    "port_name": "1",
    "serial_number": "ABC123",
    "model": "BRCM-57800"
  },
  {
    "node_name": "storage-02",
    "port_name": "0",
    "serial_number": "XYZ789",
    "model": "BRCM-9200"
  }
]

修改后的解析代码 #

import re
import json

def parse_fdmi_log(log_path):
    with open(log_path, 'r') as f:
        log_content = f.read()
    
    # 用长分隔线拆分每个独立的Port属性块
    port_blocks = re.split(r'-{5,}', log_content)
    parsed_ports = []
    
    for block in port_blocks:
        block = block.strip()
        # 跳过空块和日志标题块
        if not block or 'FDMI Device Database:' in block:
            continue
        
        # 从当前块提取所有关联属性，整合到单个对象
        port_data = {}
        if match := re.search(r'Node Name:\s*(.*)', block):
            port_data['node_name'] = match.group(1).strip()
        if match := re.search(r'Port Name:\s*(.*)', block):
            port_data['port_name'] = match.group(1).strip()
        if match := re.search(r'Serial Number:\s*(.*)', block):
            port_data['serial_number'] = match.group(1).strip()
        if match := re.search(r'Model:\s*(.*)', block):
            port_data['model'] = match.group(1).strip()
        
        # 只添加包含有效数据的Port对象
        if port_data:
            parsed_ports.append(port_data)
    
    return json.dumps(parsed_ports, indent=2)

# 测试调用
if __name__ == '__main__':
    print(parse_fdmi_log('fdmi_log.txt'))

关键修改点 #

1. 按块拆分日志：利用日志中重复的长分隔线（-----）将日志拆分为单个Port的属性块，确保每个块对应一个Port的完整信息
2. 块内属性整合：对每个块单独提取所有关联属性，直接存入同一个字典，实现Port属性的集中存储
3. 过滤无效数据：跳过空块和日志头部的标题块，避免生成无效的Port对象
4. 鲁棒性处理：使用条件判断兼容日志异常情况（如某个属性缺失），不会导致解析失败

内容的提问来源于stack exchange，提问作者itsavy