lesson-json-parse-failure-handling.json

{
  "task_id": "lesson-json-parse-failure-handling",
  "title": "JSON 解析失败处理 — 截断 / 格式错误",
  "domain": "devops",
  "tags": [
    "json",
    "parse",
    "truncated",
    "llm",
    "output"
  ],
  "problem": "从 LLM 输出或 API 返回中解析 JSON 时报 `json.decoder.JSONDecodeError`。常见于模型输出被截断、前后有多余字符。",
  "solution": "```python\nimport json\nimport re\n\ndef safe_json_parse(text: str) -> dict | None:\n    \"\"\"尝试多种策略解析 JSON\"\"\"\n    \n    # 策略 1：直接解析\n    try:\n        return json.loads(text)\n    except json.JSONDecodeError:\n        pass\n    \n    # 策略 2：提取代码块中的 JSON\n    m = re.search(r'```(?:json)?\\s*\\n?(.*?)\\n?```', text, re.DOTALL)\n    if m:\n        try:\n            return json.loads(m.group(1))\n        except json.JSONDecodeError:\n            pass\n    \n    # 策略 3：提取第一个 { } 间的完整内容\n    m = re.search(r'(\\{.*\\})', text, re.DOTALL)\n    if m:\n        try:\n            return json.loads(m.group(1))\n        except json.JSONDecodeError:\n            pass\n    \n    # 策略 4：使用 strict=False（允许控制字符）\n    try:\n        return json.loads(text, strict=False)\n    except json.JSONDecodeError:\n        pass\n    \n    return None\n\n# 使用\nresult = safe_json_parse(model_output)\nif result is None:\n    print(\"JSON 解析失败，尝试重新生成\")\n```",
  "source": "lessons/contrib/json-parse-failure-handling.md",
  "test_cmd": ""
}