1Using#region Using
  2
  3using System;
  4using System.IO;
  5using System.Text;
  6using System.Web;
  7using System.Collections.ObjectModel;
  8using System.Collections.Generic;
  9using System.Collections.Specialized;
 10using LiveBlog.Core;
 11using System.Text.RegularExpressions;
 12
 13#endregion
 14
 15namespace LiveBlog.Core
 16...{
 17  /**//// <summary>
 18  /// Searches the post collection and returns a result based on a search term.
 19  /// <remarks>It is used for related posts and the in-site search feature.</remarks>
 20  /// </summary>
 21  public static class Search
 22  ...{
 23
 24    static Search()
 25    ...{
 26      BuildCatalog();
 27      Post.Saved += new EventHandler<SavedEventArgs>(Post_Saved);
 28      Page.Saved += new EventHandler<SavedEventArgs>(Page_Saved);
 29      BlogSettings.Changed += delegate ...{ BuildCatalog(); };
 30      Post.CommentAdded += new EventHandler<EventArgs>(Post_CommentAdded);
 31      Post.CommentRemoved += delegate ...{ BuildCatalog(); };
 32      Comment.Approved += new EventHandler<EventArgs>(Post_CommentAdded);
 33    }
 34    
 35    Event handlers#region Event handlers
 36
 37    /**//// <summary>
 38    /// Adds a post to the catalog when it is added.
 39    /// </summary>
 40    private static void Post_Saved(object sender, SavedEventArgs e)
 41    ...{
 42      lock (_SyncRoot)
 43      ...{
 44        if (e.Action == SaveAction.Insert)
 45        ...{
 46          AddItem(sender as Post);
 47        }
 48        else
 49        ...{
 50          BuildCatalog();
 51        }
 52      }
 53    }
 54
 55    private static void Page_Saved(object sender, SavedEventArgs e)
 56    ...{
 57      lock (_SyncRoot)
 58      ...{
 59        if (e.Action == SaveAction.Insert)
 60        ...{
 61          AddItem(sender as Page);
 62        }
 63        else
 64        ...{
 65          BuildCatalog();
 66        }
 67      }
 68    }
 69
 70        static void Post_CommentAdded(object sender, EventArgs e)
 71        ...{
 72            if (BlogSettings.Instance.EnableCommentSearch)
 73            ...{
 74                Comment comment = (Comment)sender;
 75                if (comment.IsVisible)
 76                    AddItem(comment);
 77            }
 78        }
 79
 80    #endregion
 81
 82    Search#region Search
 83
 84    /**//// <summary>
 85    /// Searches all the posts and returns a ranked result set.
 86    /// </summary>
 87    /// <param name="searchTerm">The term to search for</param>
 88    /// <param name="includeComments">True to include a post's comments and their authors in search</param>
 89    public static List<IPublishable> Hits(string searchTerm, bool includeComments)
 90    ...{
 91      lock (_SyncRoot)
 92      ...{
 93        List<Result> results = BuildResultSet(searchTerm, includeComments);
 94        List<IPublishable> items = results.ConvertAll(new Converter<Result, IPublishable>(ResultToPost));
 95                results.Clear();
 96        OnSearcing(searchTerm);
 97        return items;
 98      }
 99    }
100
101    /**//// <summary>
102    /// Returns a list of posts that is related to the specified post.
103    /// </summary>
104    public static List<IPublishable> FindRelatedItems(IPublishable post)
105    ...{
106      string term = post.Title;
107
108      foreach (Category cat in post.Categories)
109      ...{
110        term += " " + cat.Title;
111      }
112
113      term = CleanContent(term, false);
114      return Hits(term, false);
115    }
116
117    /**//// <summary>
118    /// Builds the results set and ranks it.
119    /// </summary>
120    private static List<Result> BuildResultSet(string searchTerm, bool includeComments)
121    ...{
122      List<Result> results = new List<Result>();
123      string term = CleanContent(searchTerm.ToLowerInvariant().Trim(), false);
124      string[] terms = term.Split(new char[] ...{ ' ' }, StringSplitOptions.RemoveEmptyEntries);
125      string regex = string.Format(System.Globalization.CultureInfo.InvariantCulture, "({0})", string.Join("|", terms));
126
127      foreach (Entry entry in _Catalog)
128      ...{
129        Result result = new Result();
130        if (!(entry.Item is Comment))
131        ...{
132          int titleMatches = Regex.Matches(entry.Title, regex).Count;
133          result.Rank = titleMatches * 20;
134
135          int postMatches = Regex.Matches(entry.Content, regex).Count;
136          result.Rank += postMatches;
137        }
138        else if (includeComments)
139        ...{
140          int commentMatches = Regex.Matches(entry.Content + entry.Title, regex).Count;
141          result.Rank += commentMatches;
142        }
143
144        if (result.Rank > 0)
145        ...{
146          result.Item = entry.Item;
147          results.Add(result);
148        }
149      }
150
151      results.Sort();
152      return results;
153    }
154
155    /**//// <summary>
156    /// A converter delegate used for converting Results to Posts.
157    /// </summary>
158    private static IPublishable ResultToPost(Result result)
159    ...{
160      return result.Item;
161    }
162
163    #endregion
164
165    Properties and private fields#region Properties and private fields
166
167    private readonly static object _SyncRoot = new object();
168    private readonly static Regex _StripHtml = new Regex("<[^>]*>", RegexOptions.Compiled);
169    private readonly static StringCollection _StopWords = GetStopWords();
170    private static Collection<Entry> _Catalog = new Collection<Entry>();
171
172    #endregion
173
174    BuildCatalog#region BuildCatalog
175
176    private static void BuildCatalog()
177    ...{
178      OnIndexBuilding();
179
180      lock (_SyncRoot)
181      ...{
182        _Catalog.Clear();
183        foreach (Post post in Post.Posts)
184        ...{
185          if (!post.IsVisible)
186            continue;
187
188          AddItem(post);
189          if (BlogSettings.Instance.EnableCommentSearch)
190          ...{
191            foreach (Comment comment in post.Comments)
192            ...{
193              if (comment.IsVisible)
194                AddItem(comment);
195            }
196          }
197        }
198
199        foreach (Page page in Page.Pages)
200        ...{
201          if (page.IsVisible)
202            AddItem(page);
203        }
204      }
205
206      OnIndexBuild();
207    }
208
209    /**//// <summary>
210    /// Adds an IPublishable item to the search catalog. 
211    /// That will make it immediately searchable.
212    /// </summary>
213    public static void AddItem(IPublishable item)
214    ...{
215      Entry entry = new Entry();
216      entry.Item = item;
217      entry.Title = CleanContent(item.Title, false);
218      entry.Content = HttpUtility.HtmlDecode(CleanContent(item.Content, true));
219      if (item is Comment)
220      ...{
221        entry.Content += HttpUtility.HtmlDecode(CleanContent(item.Author, false));
222      }
223      _Catalog.Add(entry);
224    }
225
226    /**//// <summary>
227    /// Removes stop words and HTML from the specified string.
228    /// </summary>
229    private static string CleanContent(string content, bool removeHtml)
230    ...{
231      if (removeHtml)
232        content = _StripHtml.Replace(content, string.Empty);
233
234            content = content
235                                            .Replace("\\", string.Empty)
236                                            .Replace("|", string.Empty)
237                                            .Replace("(", string.Empty)
238                                            .Replace(")", string.Empty)
239                                            .Replace("[", string.Empty)
240                                            .Replace("]", string.Empty)
241                                            .Replace("+", string.Empty);
242
243      string[] words = content.Split(new char[] ...{ ' ', '\n', '\r' }, StringSplitOptions.RemoveEmptyEntries);
244      StringBuilder sb = new StringBuilder();
245      for (int i = 0; i < words.Length; i++)
246      ...{
247        string word = words[i].ToLowerInvariant().Trim();
248        if (word.Length > 1 && !_StopWords.Contains(word))
249          sb.Append(word + " ");
250      }
251
252      return sb.ToString();
253    }
254
255    /**//// <summary>
256    /// Retrieves the stop words from the stopwords.txt file located in App_Data.
257    /// </summary>
258    private static StringCollection GetStopWords()
259    ...{
260      using (StreamReader reader = new StreamReader(System.Web.HttpContext.Current.Server.MapPath(BlogSettings.Instance.StorageLocation) + "stopwords.txt"))
261      ...{
262        string file = reader.ReadToEnd();
263        string[] words = file.Split(new char[] ...{ '\n', '\r' }, StringSplitOptions.RemoveEmptyEntries);
264
265        StringCollection col = new StringCollection();
266        col.AddRange(words);
267
268        return col;
269      }
270    }
271
272    #endregion
273
274