Day 10 | 文件处理与正则

#!/usr/bin/env python3
"""
Day 10 - JSON 文件处理示例
演示 JSON 的读写和解析
"""

import json

# ========== 1. 写入 JSON 文件 ==========
print("=== 写入 JSON 文件 ===")

# 准备数据（字典格式）
config_data = {
    "app": {
        "name": "insightful-ops",
        "version": "1.0.0",
        "debug": False
    },
    "database": {
        "host": "localhost",
        "port": 3306,
        "name": "infra_db"
    },
    "servers": ["web-01", "web-02", "db-01"]
}

# 写入 JSON 文件
# indent=4: 格式化输出，每级缩进 4 个空格（便于阅读）
# ensure_ascii=False: 允许非 ASCII 字符（如中文）
with open("config.json", "w", encoding="utf-8") as f:
    json.dump(config_data, f, indent=4, ensure_ascii=False)

print("✅ 配置文件已写入 config.json")

# ========== 2. 读取 JSON 文件 ==========
print("\n=== 读取 JSON 文件 ===")

with open("config.json", "r", encoding="utf-8") as f:
    loaded_data = json.load(f)

print(f"应用名称: {loaded_data['app']['name']}")
print(f"数据库端口: {loaded_data['database']['port']}")
print(f"服务器列表: {loaded_data['servers']}")

# ========== 3. JSON 字符串处理 ==========
print("\n=== JSON 字符串处理 ===")

# 模拟 API 返回的 JSON 字符串
api_response = '{"status": "success", "data": {"users": 100, "servers": 5}}'

# 解析 JSON 字符串
response_dict = json.loads(api_response)
print(f"状态: {response_dict['status']}")
print(f"用户数: {response_dict['data']['users']}")

# 将字典转换为 JSON 字符串
json_string = json.dumps(response_dict, indent=2)
print(f"\nJSON 字符串:\n{json_string}")

#!/usr/bin/env python3
"""
Day 10 - YAML 文件处理示例
演示 YAML 的读写和转换
"""

import yaml
import json

# ========== 1. 创建 YAML 内容 ==========
print("=== 创建 YAML 内容 ===")

# YAML 支持注释和多行字符串
yaml_content = """
# 应用配置
app:
    name: insightful-ops
    version: 1.0.0
    debug: false

# 数据库配置
database:
    host: localhost
    port: 3306
    name: infra_db

# 服务器列表
servers:
    - web-01
    - web-02
    - db-01
"""

# ========== 2. 解析 YAML ==========
print("=== 解析 YAML ===")

# safe_load: 安全加载（只加载标准 YAML 标签，防止代码执行）
data = yaml.safe_load(yaml_content)
print(f"应用名称: {data['app']['name']}")
print(f"数据库端口: {data['database']['port']}")

# ========== 3. 写入 YAML 文件 ==========
print("\n=== 写入 YAML 文件 ===")

config_dict = {
    "app": {
        "name": "insightful-ops",
        "version": "1.0.0"
    },
    "database": {
        "host": "localhost",
        "port": 3306
    }
}

with open("config.yaml", "w", encoding="utf-8") as f:
    yaml.dump(config_dict, f, default_flow_style=False, allow_unicode=True)
    # default_flow_style=False: 使用块样式（更易读）
    # allow_unicode=True: 允许 Unicode 字符

print("✅ 配置文件已写入 config.yaml")

# ========== 4. YAML 转 JSON ==========
print("\n=== YAML 转 JSON ===")

# 读取 YAML
with open("config.yaml", "r", encoding="utf-8") as f:
    yaml_data = yaml.safe_load(f)

# 转换为 JSON
with open("config_from_yaml.json", "w", encoding="utf-8") as f:
    json.dump(yaml_data, f, indent=4, ensure_ascii=False)

print("✅ YAML 已转换为 JSON")

#!/usr/bin/env python3
"""
Day 10 - 正则表达式实战
演示从日志中提取关键信息
"""

import re

# ========== 1. 基础匹配 ==========
print("=== 基础匹配 ===")

# 匹配 IP 地址
text = "Server IP: 192.168.1.100"
pattern = r'\d+\.\d+\.\d+\.\d+'  # \d+ 匹配数字，\. 匹配点号

match = re.search(pattern, text)
if match:
    print(f"找到 IP: {match.group()}")  # 192.168.1.100

# ========== 2. 捕获组 ==========
print("\n=== 捕获组 ===")

# 从日志中提取 IP、方法和状态码
log_line = '192.168.1.10 - - [31/Jan/2026:14:00:00 +0800] "GET /index.html HTTP/1.1" 200 4096'

# 正则模式解释：
# (\d+\.\d+\.\d+\.\d+)  - 捕获组1：IP 地址
# .*?                   - 非贪婪匹配任意字符
# "(\w+)                - 捕获组2：请求方法（GET/POST等）
# .*?"                  - 匹配到引号结束
# \s(\d{3})\s           - 捕获组3：状态码（3位数字）
pattern = r'(\d+\.\d+\.\d+\.\d+).*?"(\w+) .*?"\s(\d{3})\s'

match = re.search(pattern, log_line)
if match:
    ip = match.group(1)      # 第一个捕获组
    method = match.group(2)   # 第二个捕获组
    status = match.group(3)  # 第三个捕获组
    print(f"IP: {ip}, 方法: {method}, 状态码: {status}")

# ========== 3. 查找所有匹配 ==========
print("\n=== 查找所有匹配 ===")

log_text = """
192.168.1.10 - - [31/Jan/2026:14:00:00] "GET /index.html" 200
192.168.1.11 - - [31/Jan/2026:14:01:00] "POST /api/data" 201
192.168.1.12 - - [31/Jan/2026:14:02:00] "GET /about.html" 404
"""

# 提取所有 IP 地址
ips = re.findall(r'\d+\.\d+\.\d+\.\d+', log_text)
print(f"所有 IP: {ips}")

# 提取所有状态码
status_codes = re.findall(r'\s(\d{3})\s', log_text)
print(f"所有状态码: {status_codes}")

# ========== 4. 文本替换 ==========
print("\n=== 文本替换 ===")

text = "Contact: email@example.com or phone: 138-0013-8000"

# 替换邮箱为 [EMAIL]
masked = re.sub(r'\w+@\w+\.\w+', '[EMAIL]', text)
print(f"脱敏后: {masked}")

# 替换手机号为 [PHONE]
masked = re.sub(r'\d{3}-\d{4}-\d{4}', '[PHONE]', masked)
print(f"进一步脱敏: {masked}")

# ========== 5. 复杂日志解析 ==========
print("\n=== 复杂日志解析 ===")

error_log = """
[2026-01-31 10:00:01] INFO: Service started
[2026-01-31 10:05:23] ERROR: Database connection timeout (IP: 192.168.1.5)
[2026-01-31 10:06:12] ERROR: Out of memory in worker process
[2026-01-31 10:07:00] INFO: Health check passed
"""

# 提取所有错误日志
error_pattern = r'\[([^\]]+)\]\s+ERROR:\s+(.+)'
errors = re.findall(error_pattern, error_log)

print("发现的错误:")
for timestamp, message in errors:
    print(f"  [{timestamp}] {message}")

# 从错误信息中提取 IP 地址
for timestamp, message in errors:
    ip_match = re.search(r'\d+\.\d+\.\d+\.\d+', message)
    if ip_match:
        print(f"  错误 IP: {ip_match.group()}")