| 3 | from urllib.parse import urljoin |
| 4 | |
| 5 | class WebParserClient: |
| 6 | def __init__(self, base_url: str = "http://localhost:8000"): |
| 7 | """ |
| 8 | 初始化Web解析器客户端 |
| 9 | |
| 10 | Args: |
| 11 | base_url: API服务器的基础URL，默认为本地测试服务器 |
| 12 | """ |
| 13 | self.base_url = base_url.rstrip('/') |
| 14 | |
| 15 | def parse_urls(self, urls: List[str]) -> List[Dict[str, Union[str, bool]]]: |
| 16 | """ |
| 17 | 发送URL列表到解析服务器并获取解析结果 |
| 18 | |
| 19 | Args: |
| 20 | urls: 需要解析的URL列表 |
| 21 | |
| 22 | Returns: |
| 23 | 解析结果列表 |
| 24 | |
| 25 | Raises: |
| 26 | requests.exceptions.RequestException: 当API请求失败时 |
| 27 | """ |
| 28 | endpoint = urljoin(self.base_url, "/parse_urls") |
| 29 | response = requests.post(endpoint, json={"urls": urls}) |
| 30 | response.raise_for_status() # 如果响应状态码不是200，抛出异常 |
| 31 | |
| 32 | return response.json()["results"] |
| 33 | |
| 34 | |
| 35 | # 使用示例 |