schoolNews/schoolNewsCrawler/test.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "948be230",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "项目根目录: f:\\Project\\schoolNews\n",
      "✓ 已启用自动重载模块功能 - 修改 .py 文件后会自动生效\n"
     ]
    }
   ],
   "source": [
    "# 自动重载模块（当文件修改后自动刷新）\n",
    "%reload_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import sys\n",
    "import os\n",
    "\n",
    "# 先添加项目根目录到路径（必须在导入之前）\n",
    "project_root = os.path.abspath(os.path.join(os.getcwd(), '..'))\n",
    "if project_root not in sys.path:\n",
    "    sys.path.insert(0, project_root)\n",
    "\n",
    "# 然后再导入模块\n",
    "from crawler.xhw.XhwCrawler import XhwCrawler\n",
    "from crawler.BaseCrawler import NewsItem\n",
    "from loguru import logger\n",
    "import json\n",
    "from pprint import pprint\n",
    "\n",
    "print(f\"项目根目录: {project_root}\")\n",
    "print(\"✓ 已启用自动重载模块功能 - 修改 .py 文件后会自动生效\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "31a8a0dd",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m2025-11-20 14:48:38.587\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mcrawler.BaseCrawler\u001b[0m:\u001b[36m__init__\u001b[0m:\u001b[36m71\u001b[0m - \u001b[1m初始化爬虫: XhwCrawler\u001b[0m\n",
      "\u001b[32m2025-11-20 14:48:39.615\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mcrawler.xhw.XhwCrawler\u001b[0m:\u001b[36m_init_driver\u001b[0m:\u001b[36m113\u001b[0m - \u001b[1mChrome浏览器初始化成功\u001b[0m\n",
      "\u001b[32m2025-11-20 14:48:39.615\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mcrawler.xhw.XhwCrawler\u001b[0m:\u001b[36m_init_driver\u001b[0m:\u001b[36m122\u001b[0m - \u001b[1m访问主页获取初始Cookie\u001b[0m\n",
      "\u001b[32m2025-11-20 14:48:39.616\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mcrawler.xhw.XhwCrawler\u001b[0m:\u001b[36m_init_driver\u001b[0m:\u001b[36m123\u001b[0m - \u001b[1m准备访问URL: https://xhsz.news.cn/\u001b[0m\n",
      "\u001b[32m2025-11-20 14:48:41.227\u001b[0m | \u001b[1mINFO    \u001b[0m | \u001b[36mcrawler.xhw.XhwCrawler\u001b[0m:\u001b[36m_init_driver\u001b[0m:\u001b[36m126\u001b[0m - \u001b[1m成功访问URL: https://xhsz.news.cn/\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "crawler = XhwCrawler()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e5a6e91c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#crawler.search(\"大学\", 1)\n",
    "# crawler.search(\"中国\", 10, \"xhsz\")\n",
    "# crawler.search(\"中国\", 10, \"news\")\n",
    "# crawler.search(\"中国\", 10, \"xhsz\")\n",
    "# crawler.search(\"中国\", 10, \"news\")\n",
    "# crawler.search(\"中国\", 10, \"news\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "7e0f56fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "# crawler.parse_xhsz_news_detail(\"https://xhsz.news.cn/focus_news/detail?id=9752\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "47327ebf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "NewsItem(title='微纪录片｜习近平的“三农”情', contentRows=[{'tag': 'video', 'content': \"<video src='https://vodpub6.v.news.cn/yqfbzx-original/20240207/202402072819fe60663140eab9599581dcae8c1e_73db5fe0318c44469be3ea83adfc730d.mp4' />\"}, {'tag': 'p', 'content': '<p style=\"text-align: center;\">习近平对农民有着深厚的感情</p>'}, {'tag': 'p', 'content': '<p style=\"text-align: center;\">对“三农”问题有深入的思考</p>'}, {'tag': 'p', 'content': '<p style=\"text-align: center;\">一路走来</p>'}, {'tag': 'p', 'content': '<p style=\"text-align: center;\">他经常和农民在一起</p>'}, {'tag': 'p', 'content': '<p style=\"text-align: center;\">从小小山村的党支部书记</p>'}, {'tag': 'p', 'content': '<p style=\"text-align: center;\">到党的总书记</p>'}, {'tag': 'p', 'content': '<p style=\"text-align: center;\">寸寸光阴</p>'}, {'tag': 'p', 'content': '<p style=\"text-align: center;\">见证着他的“三农”情</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003总策划：刘健</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003策划：李拯宇</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003监制：孙志平</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003制片：樊华</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003统筹：韩珅、王志斌</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003编导：陈晓宇</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003记者：陈晓宇、范世辉、岳文婷、邹尚伯、张晨俊、王怿文、李涛、胡友松、朱晓光</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003报道员：刘鹏飞、李树锋、张伟、朱海亮、徐涛、王盟、王静</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003海报：韩彤（实习）</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003鸣谢：中共延川县委宣传部</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003中共石家庄市委宣传部</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003中共曹县县委宣传部</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003新华社音视频部制作</p>'}, {'tag': 'p', 'content': '<p>\\u2003\\u2003新华通讯社出品</p>'}], url='https://www.news.cn/politics/leaders/20240207/2819fe60663140eab9599581dcae8c1e/c.html', viewCount=None, publishTime='2024-02-07 12:43:29', author=None, source='新华社', category=None, executeStatus=0, executeMessage=None)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#crawler.parse_xh_news_detail(\"https://www.news.cn/politics/leaders/20250224/5384be3d47c643b3a68e3bb724656152/c.html\")\n",
    "crawler.parse_xh_news_detail(\"https://www.news.cn/politics/leaders/20240207/2819fe60663140eab9599581dcae8c1e/c.html\") #视频"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fa359d5b",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "schoolNewsCrawler",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}