• C# 网页图片采集


    调用方法 GatherPic g = new GatherPic(“http://www.baidu.com”,"E:/XXX");
                g.start();

    using System;
    using System.Collections.Generic;
    using System.Linq;
    using System.Text;
    using System.Text.RegularExpressions;
    using System.Net;
    using System.IO;
    using System.Windows.Forms;
    namespace ImageCollect
    {
        public class GatherPic
        {
            private string savePath;
            private string getUrl;
            private WebBrowser wb;
            private int iImgCount;
            //初始化参数
            public GatherPic(string sWebUrl, string sSavePath)
            {
                this.getUrl = sWebUrl;
                this.savePath = sSavePath;
            }
            //开始采集
            public bool start()
            {
                if (getUrl.Trim().Equals(""))
                {
                    MessageBox.Show("哪来的虾米连网址都没输!");
                    return false;
                }
                this.wb = new WebBrowser();
                this.wb.Navigate(getUrl);
                //委托事件
                this.wb.DocumentCompleted += new System.Windows.Forms.WebBrowserDocumentCompletedEventHandler(DocumentCompleted);
                return true;
            }
            //WebBrowser.DocumentCompleted委托事件
            private void DocumentCompleted(object sender, WebBrowserDocumentCompletedEventArgs e)
            {
                //页面里框架iframe加载完成不掉用SearchImgList()
                if (e.Url != wb.Document.Url) return;
                SearchImgList();
            }
            //检查出所有图片并采集到本地
            public void SearchImgList()
            {
                string sImgUrl;
                //取得所有图片地址
                HtmlElementCollection elemColl = this.wb.Document.GetElementsByTagName("img");
                this.iImgCount = elemColl.Count;
                foreach (HtmlElement elem in elemColl)
                {
                    sImgUrl = elem.GetAttribute("src");
                    //调用保存远程图片函数
                    SaveImageFromWeb(sImgUrl, this.savePath);
                }
            }
            //保存远程图片函数
            public int SaveImageFromWeb(string imgUrl, string path)
            {
                string imgName = imgUrl.ToString().Substring(imgUrl.ToString().LastIndexOf("/") + 1);
                path = path + "//" + imgName;
                string defaultType = ".jpg";
                string[] imgTypes = new string[] { ".jpg", ".jpeg", ".png", ".gif", ".bmp" };
                string imgType = imgUrl.ToString().Substring(imgUrl.ToString().LastIndexOf("."));
                foreach (string it in imgTypes)
                {
                    if (imgType.ToLower().Equals(it))
                        break;
                    if (it.Equals(".bmp"))
                        imgType = defaultType;
                }
                try
                {
                    HttpWebRequest request = (HttpWebRequest)WebRequest.Create(imgUrl);
                    request.UserAgent = "Mozilla/6.0 (MSIE 6.0; Windows NT 5.1; Natas.Robot)";
                    request.Timeout = 10000;
                    WebResponse response = request.GetResponse();
                    Stream stream = response.GetResponseStream();
                    if (response.ContentType.ToLower().StartsWith("image/"))
                    {
                        byte[] arrayByte = new byte[1024];
                        int imgLong = (int)response.ContentLength;
                        int l = 0;
                        // CreateDirectory(path);
                        FileStream fso = new FileStream(path, FileMode.Create);
                        while (l < imgLong)
                        {
                            int i = stream.Read(arrayByte, 0, 1024);
                            fso.Write(arrayByte, 0, i);
                            l += i;
                        }
                        fso.Close();
                        stream.Close();
                        response.Close();
                        return 1;
                    }
                    else
                    {
                        return 0;
                    }
                }
                catch (WebException)
                {
                    return 0;
                }
                catch (UriFormatException)
                {
                    return 0;
                }
            }
        }
    }

  • 相关阅读:
    [ css 计数器 counter ] css中counter计数器实例演示三
    [ css 计数器 counter ] css中counter计数器实例演示二
    Leetcode 15. 三数之和
    Leetcode 13. 罗马数字转整数
    Leetcode 19. 删除链表的倒数第 N 个结点
    Leetcode 12. 整数转罗马数字
    Leetcode 11. 盛最多水的容器 双指针
    剑指 Offer 68
    剑指 Offer 68
    面试题 04.02. 最小高度树
  • 原文地址:https://www.cnblogs.com/blsong/p/2661879.html
Copyright © 2020-2023  润新知