| import gradio as gr |
| import requests |
| import markdownify |
| import readabilipy.simple_json |
| from urllib.parse import urlparse, urlunparse |
| from protego import Protego |
| from typing import Tuple |
|
|
| |
| DEFAULT_USER_AGENT = "Gradio Fetch App/1.0 (https://gradio.app)" |
|
|
| |
| |
| def extract_content_from_html(html: str) -> str: |
| try: |
| ret = readabilipy.simple_json.simple_json_from_html_string(html, use_readability=True) |
| if not ret or not ret.get("content"): |
| return "页面简化失败:无法从HTML中提取主要内容。" |
| content = markdownify.markdownify(ret["content"], heading_style=markdownify.ATX) |
| return content if content.strip() else "页面简化后内容为空。" |
| except Exception as e: |
| return f"处理HTML时出错: {e}" |
|
|
| def get_robots_txt_url(url: str) -> str: |
| parsed = urlparse(url) |
| robots_url = urlunparse((parsed.scheme, parsed.netloc, "/robots.txt", "", "", "")) |
| return robots_url |
|
|
| def check_can_fetch(url: str, user_agent: str) -> Tuple[bool, str]: |
| robots_url = get_robots_txt_url(url) |
| try: |
| response = requests.get(robots_url, headers={"User-Agent": user_agent}, timeout=5) |
| if response.status_code >= 400: |
| return True, f"无法获取 {robots_url} (状态码: {response.status_code}),将继续抓取。" |
| robot_parser = Protego.parse(response.text) |
| if not robot_parser.can_fetch(url, user_agent): |
| return False, f"**抓取被拒绝**:该网站的 `robots.txt` 文件 ({robots_url}) 不允许我们的用户代理 ('{user_agent}') 访问此页面。" |
| return True, f"根据 {robots_url} 的规则,允许抓取。" |
| except requests.exceptions.RequestException as e: |
| return True, f"检查 {robots_url} 时发生网络错误: {e},将继续尝试抓取。" |
|
|
| |
| def fetch_and_process_url( |
| url: str, |
| ignore_robots: bool, |
| force_raw: bool, |
| max_length: int, |
| start_index: int |
| ) -> str: |
| """抓取并处理指定URL的网页内容。 |
| |
| 此函数会访问一个网页,智能地提取其主要正文内容,并将其转换为简洁的Markdown格式。 |
| 它还能处理robots.txt规则、内容截断和返回原始HTML等高级功能。 |
| |
| Args: |
| url (str): 需要抓取的完整网址,必须以 http:// 或 https:// 开头。 |
| ignore_robots (bool): 是否忽略目标网站的 robots.txt 文件中的抓取限制。 |
| force_raw (bool): 是否返回未经简化的原始HTML内容,而不是提取后的Markdown文本。 |
| max_length (int): 返回内容的最大字符数。用于处理超长页面。 |
| start_index (int): 从内容的哪个字符位置开始返回。用于分页获取超长内容。 |
| |
| Returns: |
| str: 处理后的网页内容或错误信息。 |
| """ |
| if not url or not url.strip().startswith(('http://', 'https://')): |
| return "请输入一个有效的URL (以 http:// 或 https:// 开头)。" |
| |
| |
| url = url.strip() |
| user_agent = DEFAULT_USER_AGENT |
| if not ignore_robots: |
| can_fetch, message = check_can_fetch(url, user_agent) |
| if not can_fetch: |
| return message |
| try: |
| response = requests.get(url, headers={"User-Agent": user_agent}, timeout=30, allow_redirects=True) |
| response.raise_for_status() |
| except requests.exceptions.RequestException as e: |
| return f"抓取URL时发生网络错误: {e}" |
| page_raw = response.text |
| content_type = response.headers.get("content-type", "").lower() |
| is_html = "text/html" in content_type |
| prefix = "" |
| if is_html and not force_raw: |
| content = extract_content_from_html(page_raw) |
| else: |
| content = page_raw |
| if not is_html: |
| prefix = f"**注意**: 内容类型是 `{content_type}`,已显示为原始内容。\n\n---\n\n" |
| original_length = len(content) |
| if start_index >= original_length: |
| content_to_show = "错误:起始索引超出了内容总长度。" |
| else: |
| end_index = start_index + max_length |
| content_to_show = content[start_index:end_index] |
| remaining_chars = original_length - (start_index + len(content_to_show)) |
| if remaining_chars > 0: |
| next_start = start_index + len(content_to_show) |
| truncation_message = ( |
| f"\n\n---\n\n**内容被截断**。 " |
| f"已显示字符数: {len(content_to_show)}," |
| f"剩余字符数: {remaining_chars}。" |
| f"要获取更多内容,请将**起始索引**设置为 **{next_start}** 再试。" |
| ) |
| content_to_show += truncation_message |
| return prefix + content_to_show |
|
|
|
|
| |
| with gr.Blocks(theme=gr.themes.Soft()) as app: |
| gr.Markdown("# 网页内容抓取与简化工具") |
| gr.Markdown("输入一个网址,工具会抓取其内容,并默认提取正文转换为Markdown格式。") |
| with gr.Row(): |
| url_input = gr.Textbox(label="URL", placeholder="例如: https://www.google.com/about", scale=4) |
| fetch_button = gr.Button("抓取内容", variant="primary", scale=1) |
| with gr.Accordion("高级选项", open=False): |
| with gr.Row(): |
| ignore_robots_checkbox = gr.Checkbox(label="忽略 robots.txt 限制", value=False) |
| raw_checkbox = gr.Checkbox(label="显示原始HTML (不简化)", value=False) |
| with gr.Row(): |
| max_length_slider = gr.Slider(minimum=500, maximum=50000, value=8000, step=500, label="返回内容的最大长度 (字符数)") |
| start_index_number = gr.Number(label="起始索引", value=0) |
| output_markdown = gr.Markdown(label="抓取结果") |
| fetch_button.click( |
| fn=fetch_and_process_url, |
| inputs=[url_input, ignore_robots_checkbox, raw_checkbox, max_length_slider, start_index_number], |
| outputs=output_markdown |
| ) |
|
|
| if __name__ == "__main__": |
| app.launch(mcp_server=True) |