1. 引言
相比于Java而言,基于C#的开源爬虫就少了很多。入行这么多年也从未接触过爬虫。出于兴趣今天给大家介绍一个C# 的爬虫工具Abot. 需要使用可以通过Nuget获取。Abot本身就支持多线程的爬取,
内部使用CsQuery来解析爬取到的Html文档。熟悉jQuery的同学肯定能快速上手CsQuery, 它就是一个C#版本的jQuery。
这里以爬取博客园当天的新闻数据为例,看看如何使用Abot。
2. 博客园新闻页面
http://news.cnblogs.com/ 这是博客园的新闻首页。可以看到典型的分页显示。比如 http://news.cnblogs.com/n/page/2/ 这是新闻的第二页。
真正的新闻详细页面 比如:http://news.cnblogs.com/n/544956/, 通过正则表达式可以很轻松的匹配这两种Url 类型。
当然我们可以通过一个 for 循环分别爬取各个page的新闻数据。然后解析出发表于今天的新闻。但是我希望只以http://news.cnblogs.com/ 为种子页面,爬取今天的新闻。
由于博客园新闻分页并不是采用Ajax,对于爬虫而言这非常友好
因此我们定义
/// <summary> /// 种子Url /// </summary> public static readonly Uri FeedUrl = new Uri(@"http://news.cnblogs.com/"); /// <summary> ///匹配新闻详细页面的正则 /// </summary> public static Regex NewsUrlRegex = new Regex("^http://news.cnblogs.com/n/\\d+/$", RegexOptions.Compiled); /// <summary> /// 匹配分页正则 /// </summary> public static Regex NewsPageRegex = new Regex("^http://news.cnblogs.com/n/page/\\d+/$", RegexOptions.Compiled);
3. 实现
Abot 其实已经对爬虫内部实现封装的非常精巧,使用者只需要设置一些Config 参数和爬取页面的一些事件即可。
public static IWebCrawler GetManuallyConfiguredWebCrawler() { CrawlConfiguration config = new CrawlConfiguration(); config.CrawlTimeoutSeconds = 0; config.DownloadableContentTypes = "text/html, text/plain"; config.IsExternalPageCrawlingEnabled = false; config.IsExternalPageLinksCrawlingEnabled = false; config.IsRespectRobotsDotTextEnabled = false; config.IsUriRecrawlingEnabled = false; config.MaxConcurrentThreads = System.Environment.ProcessorCount; config.MaxPagesToCrawl = 1000; config.MaxPagesToCrawlPerDomain = 0; config.MinCrawlDelayPerDomainMilliSeconds = 1000; var crawler = new PoliteWebCrawler(config, null, null, null, null, null, null, null, null); crawler.ShouldCrawlPage(ShouldCrawlPage); crawler.ShouldDownloadPageContent(ShouldDownloadPageContent); crawler.ShouldCrawlPageLinks(ShouldCrawlPageLinks); crawler.PageCrawlStartingAsync += crawler_ProcessPageCrawlStarting; //爬取页面后的回调函数 crawler.PageCrawlCompletedAsync += crawler_ProcessPageCrawlCompletedAsync; crawler.PageCrawlDisallowedAsync += crawler_PageCrawlDisallowed; crawler.PageLinksCrawlDisallowedAsync += crawler_PageLinksCrawlDisallowed; return crawler; }
具体调用非常简单:
public static void Main(string[] args) { var crawler = GetManuallyConfiguredWebCrawler(); var result = crawler.Crawl(FeedUrl); System.Console.WriteLine(result.ErrorException); }
最主要的是PageCrawlCompletedAsync,可以在该事件下获取需要的页面数据。
public static void crawler_ProcessPageCrawlCompletedAsync(object sender, PageCrawlCompletedArgs e) { //判断是否是新闻详细页面 if (NewsUrlRegex.IsMatch(e.CrawledPage.Uri.AbsoluteUri)) { //获取信息标题和发表的时间 var csTitle = e.CrawledPage.CsQueryDocument.Select("#news_title"); var linkDom = csTitle.FirstElement().FirstChild; var newsInfo = e.CrawledPage.CsQueryDocument.Select("#news_info"); var dateString = newsInfo.Select(".time", newsInfo); //判断是不是今天发表的 if (IsPublishToday(dateString.Text())) { var str = (e.CrawledPage.Uri.AbsoluteUri + "\t" + HtmlData.HtmlDecode(linkDom.InnerText) + "\r\n"); System.IO.File.AppendAllText("fake.txt", str); } } } /// <summary> /// "发布于 2016-05-09 11:25" => true /// </summary> public static bool IsPublishToday(string str) { if (string.IsNullOrEmpty(str)) { return false; } const string prefix = "发布于"; int index = str.IndexOf(prefix, StringComparison.OrdinalIgnoreCase); if (index >= 0) { str = str.Substring(prefix.Length).Trim(); } DateTime date; return DateTime.TryParse(str, out date) && date.Date.Equals(DateTime.Today); }
为了提升爬取的效果 比如在首页爬虫抓取到 http://news.cnblogs.com/n/topiclist/, 显然这样的链接我们不需要, 那就可以设置爬取的规则:
/// <summary> /// 如果是Feed页面或者分页或者详细页面才需要爬取 /// </summary> private static CrawlDecision ShouldCrawlPage(PageToCrawl pageToCrawl, CrawlContext context) { if (pageToCrawl.IsRoot || pageToCrawl.IsRetry || FeedUrl == pageToCrawl.Uri || NewsPageRegex.IsMatch(pageToCrawl.Uri.AbsoluteUri) || NewsUrlRegex.IsMatch(pageToCrawl.Uri.AbsoluteUri)) { return new CrawlDecision {Allow = true}; } else { return new CrawlDecision {Allow = false, Reason = "Not match uri"}; } } /// <summary> /// 如果是Feed页面或者分页或者详细页面才需要爬取 /// </summary> private static CrawlDecision ShouldDownloadPageContent(PageToCrawl pageToCrawl, CrawlContext crawlContext) { if (pageToCrawl.IsRoot || pageToCrawl.IsRetry || FeedUrl == pageToCrawl.Uri || NewsPageRegex.IsMatch(pageToCrawl.Uri.AbsoluteUri) || NewsUrlRegex.IsMatch(pageToCrawl.Uri.AbsoluteUri)) { return new CrawlDecision { Allow = true }; } return new CrawlDecision { Allow = false, Reason = "Not match uri" }; } private static CrawlDecision ShouldCrawlPageLinks(CrawledPage crawledPage, CrawlContext crawlContext) { if (!crawledPage.IsInternal) return new CrawlDecision {Allow = false, Reason = "We dont crawl links of external pages"}; if (crawledPage.IsRoot || crawledPage.IsRetry || crawledPage.Uri == FeedUrl || NewsPageRegex.IsMatch(crawledPage.Uri.AbsoluteUri)) { return new CrawlDecision {Allow = true}; } else { return new CrawlDecision {Allow = false, Reason = "We only crawl links of pagination pages"}; } }
最终抓到的数据:
4. 总结
Abot 还是一个非常方便爬虫,如果运用到实际生产环境中,参数配置是首先需要解决的,比如 MaxPagesToCrawl 最大抓取的页面数,还可以设置爬虫内存限制等。
欢迎访问我的个人网站 51zhang.net 网站还在不断开发中…