爬虫如何解析HTML源码中script中的window.__NUXT__=后面的数据

• 1、先用xpath定位到该script标签，并且抽取出文本内容

from lxml.html import soupparser

url = 'https://bingx.com/ja-jp/events/'
res = requests.get(url, impersonate="chrome101")
pattern = '//script[contains(text(), "window.__NUXT__=")]/text()'
tree = soupparser.fromstring(res.text)
result = tree.xpath(pattern)
if result:
    js_code = result[0]
    print(js_code)

• 2、给解析出的原始文本在头部加上一行window = {};\n

如果不加, 将会报window undifined的错误，因为确实没有定义这个变量名；我们先要给初始化，定义下

    js_code = 'window = {};\n' + js_code

• 3、用execjs解析该数据

import execjs

# 编译并执行 JavaScript 代码
ctx = execjs.compile(js_code)

# 获取 window.__NUXT__ 的值
nuxt_data = ctx.eval('window.__NUXT__')

# 提取 fetch 中的 items 数组
items = nuxt_data['fetch']['data-v-d36bc1ac:0']['items']

• 4、完整代码示例

import execjs
from curl_cffi import requests
from lxml.html import soupparser
from dateutil import parser


url = 'https://bingx.com/ja-jp/events/'
res = requests.get(url, impersonate="chrome101")
pattern = '//script[contains(text(), "window.__NUXT__=")]/text()'
tree = soupparser.fromstring(res.text)
result = tree.xpath(pattern)
if result:
    js_code = result[0]
    print(js_code)
    js_code = 'window = {};\n' + js_code
    # 编译并执行 JavaScript 代码
    ctx = execjs.compile(js_code)

    # 获取 window.__NUXT__ 的值
    nuxt_data = ctx.eval('window.__NUXT__')

    # 提取 fetch 中的 items 数组
    items = nuxt_data['fetch']['data-v-d36bc1ac:0']['items']
    for item in items:
        print(item)
        print(f"标题: {item['title']}")
        print(f"内容: {item['content']}")
        print(f"活动URL: {item['activityUrl']}")
        # 发布时间
        onTime = item['onTime']
        dt = parser.isoparse(onTime)
        formatted_time = dt.strftime("%Y-%m-%d %H:%M:%S")

---

解决方案来自Grok3 Think模式；对比了chat-gpt4.5, deepseek，Grok3的回答效果最好