Parse it into this format

```
{
  "metadata": {
    "equal_contribution_detected": true, 
    "extracted_emails": [
      "email1@example.com",
      "email2@example.com"
    ]
  },
  "authors": [
    {
      "rank": 1,
      "first_name": "String", (drop middle name)
      "last_name": "String",
      "organization": "String (Optional)",
      "school_department": "String (Optional)"
    },
    {
      "rank": 2,
...
    }
  ]
}
```


```
user@fedora ~/Downloads> tree
.
└── example_data
    ├── 10
    │   └── 一种自动构建STRIDE威胁规则模型和更新规则库的方法_付昌兰.pdf
    ├── 11
    │   └── 预测控制在步进梁加热炉先进控制中的应用研究_汪开红.pdf
    ├── 12
    │   └── 城市空中交通系统最优规模评估与调度_郭戈.pdf
    ├── 13
    │   └── 基于CORBA协议开放式算法仿真平台的体系结构设计_陈治纲.pdf
    ├── 14
    │   └── 基于机器学习的粪污还田过程低氨排放优化策略构建_刘德钊.pdf
    ├── 15
    │   └── 基于贝叶斯稀疏自编码器的蓄积性疲劳特征重构对机器学习模型预测制造业员工职业紧张的提升作用_宋涛.pdf
    ├── 16
    │   └── 间歇过程优化与先进控制综述_陈治纲.pdf
    ├── 17
    │   └── 面向开放互联网的科学数据挖掘与理解_卢彬.pdf
    ├── en_1
    │   ├── 1406.2661v1.pdf
    │   └── authors.json
    ├── en_2
    │   ├── 1412.6980v9.pdf
    │   └── authors.json
    ├── en_3
    │   ├── 1512.03385v1.pdf
    │   └── authors.json
    ├── en_4
    │   ├── 1706.03762v7.pdf
    │   └── authors.json
    ├── en_5
    │   ├── 2005.14165v4.pdf
    │   └── authors.json
    ├── en_6
    │   ├── 2201.11903v6.pdf
    │   └── authors.json
    ├── en_7
    │   ├── 2302.13971v1.pdf
    │   └── authors.json
    ├── en_8
    │   ├── authors.json
    │   └── NeurIPS-2024-improving-deep-reinforcement-learning-by-reducing-the-chain-effect-of-value-and-policy-churn-Paper-Conference.pdf
    ├── en_9
    │   ├── authors.json
    │   └── NeurIPS-2024-remap-neural-model-reprogramming-with-network-inversion-and-retrieval-augmented-mapping-for-adaptive-motion-forecasting-Paper-Conference.pdf
    └── README.md

19 directories, 27 files
```