今天ModestMT.Zou發布了DotnetSpider爬蟲第二章節,內容簡單明了,基本看懂了,於是想自己試試看,直接就拿博客園開刀了。
這裡有最基本的使用方式,本文章不介紹
這裡我已經從https://github.com/zlzforever/DotnetSpider上下載代碼並編譯通過
這裡用的是VS2015,因為此項目有些C#6.0語法糖
public class Cnblog { public string Title { get; set; } public string Url { get; set; } public string Author { get; set; } public string Conter { get; set; } }
先引用兩個Dll類庫
Java2Dotnet.Spider.Core.dll
Newtonsoft.Json.dll
如果你編譯DotnetSpider成功的話,可以在output目錄中找到
/// <summary> /// 頁面列表處理器 /// </summary> public class PageListProcessor : IPageProcessor { public Site Site{get; set; } public void Process(Page page) { var totalCnblogElements = page.Selectable.SelectList(Selectors.XPath("//div[@class='post_item']")).Nodes(); List<Cnblog> results = new List<Cnblog>(); foreach (var cnblogElement in totalCnblogElements) { var cnblog = new Cnblog(); cnblog.Title = cnblogElement.Select(Selectors.XPath(".//div[@class='post_item_body']/h3/a")).GetValue(); cnblog.Url = cnblogElement.Select(Selectors.XPath(".//div[@class='post_item_body']/h3")).Links().GetValue(); cnblog.Author = cnblogElement.Select(Selectors.XPath(".//div[@class='post_item_foot']/a[1]")).GetValue(); results.Add(cnblog); } page.AddResultItem("Result", results); } }
關於XPath,可以到這裡學習http://www.w3school.com.cn/xpath/,我也是下午剛看了一遍,因為有XML/HTML基礎,基本沒壓力
關於XPath表達式如何寫,我覺得用谷歌審核元素就足夠了,可以復制XPath。也有一款谷歌XPath插件,因我翻不了牆,就沒安裝。
如下圖://*[@id="post_list"]/div[20]/div[2]/h3/a,然後再按需改改
需要實現 IPipeline這個接口,然後你想保存到文件或數據庫就自己選擇
public class ListPipeline : IPipeline { private string _path; public ListPipeline(string path) { if (string.IsNullOrEmpty(path)) { throw new Exception("文件名不能為空!"); } _path = path; if (!File.Exists(_path)) { File.Create(_path); } } public void Dispose() { } public void Process(ResultItems resultItems, ISpider spider) { lock (this) { foreach (Cnblog entry in resultItems.Results["Result"]) { File.AppendAllText(_path, JsonConvert.SerializeObject(entry)); } } }
接下來在Program的Main方法中寫運行代碼
class Program { static void Main(string[] args) { var site = new Site() { EncodingName = "UTF-8" }; for (int i = 1; i <= 30; i++)//30頁 { site.AddStartUrl( $"http://www.cnblogs.com/p{i}");//已更正去掉#號,本來是"http://www.cnblogs.com/#p{i}",這樣發現請求的是http://www.cnblogs.com } Spider spider = Spider.Create(site, new PageListProcessor(), new QueueDuplicateRemovedScheduler()).AddPipeline(new ListPipeline("test.json")).SetThreadNum(2);//兩個線程 spider.Run(); Console.Read(); } }
這樣每一頁信息就被保存起來了,但到這裡還沒完,一般情況不僅僅是采集列表頁,也會采集詳細頁,於是我又添加了兩個類,暫時我是這樣實現的,但感覺有點慢
添加頁面詳細數據處理器
/// <summary> /// 頁面詳細處理器 /// </summary> public class PageDetailProcessor : IPageProcessor { private Cnblog cnblog; public PageDetailProcessor(Cnblog _cnblog) { cnblog = _cnblog; } public Site Site { get; set; } public void Process(Page page) { cnblog.Conter=page.Selectable.Select(Selectors.XPath("//*[@id='cnblogs_post_body']")).GetValue(); page.AddResultItem("detail",cnblog); } }
再添加頁面詳細數據保存
public class DetailPipeline : IPipeline { private string path; public DetailPipeline(string _path) { if (string.IsNullOrEmpty(_path)) { throw new Exception("路徑不能為空!"); } path = _path; if (!Directory.Exists(_path)) { Directory.CreateDirectory(_path); } } public void Dispose() { } public void Process(ResultItems resultItems, ISpider spider) { Cnblog cnblog=resultItems.Results["detail"]; FileStream fs=File.Create(path + "\\" + cnblog.Title + ".txt"); byte[] bytes=UTF8Encoding.UTF8.GetBytes("Url:"+cnblog.Url+Environment.NewLine+cnblog.Conter); fs.Write(bytes,0,bytes.Length); fs.Flush(); fs.Close(); } }
修改ListPipeline這個類RequestDetail方法,我的想法是列表數據保存一次就請求一次詳細頁,然後再保存詳細頁
所有詳細頁都保存在details這個目錄下
public class ListPipeline : IPipeline { private string _path; public ListPipeline(string path) { if (string.IsNullOrEmpty(path)) { throw new Exception("文件名不能為空!"); } _path = path; if (!File.Exists(_path)) { File.Create(_path); } } public void Dispose() { } public void Process(ResultItems resultItems, ISpider spider) { lock (this) { foreach (Cnblog entry in resultItems.Results["Result"]) { File.AppendAllText(_path, JsonConvert.SerializeObject(entry)); RequestDetail(entry); } } } /// <summary> /// 請求詳細頁 /// </summary> /// <param name="entry"></param> private static void RequestDetail(Cnblog entry) { ISpider spider; var site = new Site() {EncodingName = "UTF-8"}; site.AddStartUrl(entry.Url); spider = Spider.Create(site, new PageDetailProcessor(entry), new QueueDuplicateRemovedScheduler()) .AddPipeline(new DetailPipeline("details")) .SetThreadNum(1); spider.Run(); } }
其它代碼保持不變,運行程序,現在已經能保存詳細頁內容了
最後,程序運行下來沒什麼大問題,但就是在采集詳細頁時比較慢,我的想法是把所有詳細頁一起加到調度中心,然後開多個線程去運行,這個有待學習。
今天把上面的問題解決了,修改ListPipeline類,這樣就可一次把所有詳細頁都加到調度中心,然後開多個線程去請求。
public void Process(ResultItems resultItems, ISpider spider) { lock (this) { var site = new Site() { EncodingName = "UTF-8" }; foreach (Cnblog entry in resultItems.Results["Result"]) { File.AppendAllText(_path, JsonConvert.SerializeObject(entry)); site.AddStartUrl(entry.Url); } RequestDetail(site); } } /// <summary> /// 請求詳細頁 /// </summary> /// <param name="site"></param> private static void RequestDetail(Site site) { ISpider spider = Spider.Create(site, new PageDetailProcessor(), new QueueDuplicateRemovedScheduler()) .AddPipeline(new DetailPipeline("details")) .SetThreadNum(3); spider.Run(); }
PageDetailProcessor類也更改了,加入標題、url獲取
public void Process(Page page) { Cnblog cnblog=new Cnblog(); cnblog.Title = page.Selectable.Select(Selectors.XPath("//a[@id='cb_post_title_url']")).GetValue(); cnblog.Conter=page.Selectable.Select(Selectors.XPath("//*[@id='cnblogs_post_body']")).GetValue(); cnblog.Url = page.Url; page.AddResultItem("detail",cnblog); }
Demo下載