ImageUriParser.cs

using System;
using System.Text;
using System.Collections.Specialized;
using System.Text.RegularExpressions;

namespace Common
{
        public class ImageUriParser
        {

                public ImageUriParser() {}

                public static HTTPRequest[] GenerateImageRequests(MessageLogger msgLog, HTTPRequest originalRequest, HTTPResponse htmlResponse){
                        Uri[] imageUris = GetImageUris(msgLog, originalRequest.URI, htmlResponse.Body.data);
                        HTTPRequest [] requests = new HTTPRequest[imageUris.Length];
                        for(int i = 0; i < requests.Length; i++){
                                // do some processing
                                try {
                                        if (imageUris[i] != null) {
                                                requests[i] = new HTTPRequest(originalRequest, imageUris[i]);
                                        }
                                } catch (Exception ex) {
                                        msgLog.LogError("Unable to create image request for {0} : {1}", imageUris[i], ex.Message);
                                        requests[i] = null;
                                }
                        }
                        return requests;
                }

                public static Uri[] GetImageUris(MessageLogger msgLog, Uri baseUri, byte[] htmlAsASCIIBytes) {
                        string htmlString = Encoding.ASCII.GetString(htmlAsASCIIBytes);
                        StringCollection imageStrings = new StringCollection();
                        Regex r;
                        Match m;
                        // The regular expression to match against
                        r = new Regex("<\\s*?img.*?src\\s*?=\\s*?(?:\"(?<1>[^\"]*)\"|(?<1>\\S+)).*?>", RegexOptions.IgnoreCase|RegexOptions.IgnorePatternWhitespace|RegexOptions.Compiled|RegexOptions.Singleline);
                        // match strings
                        for (m = r.Match(htmlString); m.Success; m = m.NextMatch()) {
                                string s = m.Groups[1].Value;
                                s.Trim(new char[] {'\n','\r','\f'}); // remove newlines and the like
                                imageStrings.Add(s);
                        }
                        msgLog.Log("Parse found {0} images before duplicate removal", imageStrings.Count);
                        // add to unique string collection, removing duplicates
                        StringCollection uniqueStrings = new StringCollection();
                        for (int i = 0; i < imageStrings.Count; i++) {
                                if (!uniqueStrings.Contains(imageStrings[i])) { 
                                        uniqueStrings.Add(imageStrings[i]);
                                }
                        }
                        // create URIs from unique strings
                        Uri[] result = new Uri[uniqueStrings.Count];
                        for (int i = 0; i < uniqueStrings.Count; i++) {
                                try {
                                        result[i] = new Uri(baseUri, uniqueStrings[i], true); 
                                        /* if imageStrings contains full URI, this will
                                         * ignore the base uri in 'request */
                                } catch (UriFormatException) {
                                        msgLog.LogError("Dropping this parse and push IMG - bad URI");
                                }
                        }
                        return result;
                }
        }
}

---