lucene.net 2.0 src包里自帶了Highlighter.Net-2.0.0,可以用來實現(xiàn)語法高亮。
//定義多條件搜索分析器
BooleanQuery bquery = new BooleanQuery();
//定義分詞器
Analyzer analyzer = new StandardAnalyzer();
//多條件搜索拆分器
MultiFieldQueryParser parser = new MultiFieldQueryParser(new string[] { "title", "content" }, analyzer);
query = parser.Parse(key);
bquery.Add(query, BooleanClause.Occur.MUST);
DateTime now = DateTime.Now;
//初始化搜索器
//實現(xiàn)分布式搜索
List<IndexReader> indexreaders = new List<IndexReader>();
string[] dirs = Directory.GetDirectories(dir);
if (searchtype == SearchType.None)
{
foreach (string item in dirs)
{
//System.Web.HttpContext.Current.Response.Write(item);
indexreaders.Add(IndexReader.Open(Path.Combine(Path.Combine(dir, item), "Save")));
}
}
else
{
//System.Web.HttpContext.Current.Response.Write(searchtype.ToString());
indexreaders.Add(IndexReader.Open(Path.Combine(Path.Combine(dir, searchtype.ToString()), "Save")));
}
MultiReader reader = new MultiReader(indexreaders.ToArray());
indexSearcher = new IndexSearcher(reader);
Hits hits = null;
hits = indexSearcher.Search(bquery);
timer = (DateTime.Now - now).TotalMilliseconds;
int count = hits.Length();
/* 計算顯示的條目 */
int start = (pageNo - 1) * 10;
int end = pageNo * 10 > count ? count : pageNo * 10;
//Response.Write(readerhelper.MyQuery.ToString());
/* 語法高亮顯示設置 */
Highlighter highlighter = new Highlighter(new QueryScorer(query ));
highlighter.SetTextFragmenter(new SimpleFragmenter(100));
for (int i = start; i < end; i++)
{
Lucene.Net.Documents.Document doc = hits.Doc(i);
System.String text = doc.Get("content");
//添加結尾，保證結尾特殊符號不被過濾
string title = doc.Get("title") + "+aaaaaaaaa";
Lucene.Net.Analysis.TokenStream tokenStream = analyzer .TokenStream("content", new System.IO.StringReader(text));
Lucene.Net.Analysis.TokenStream titkeStream = analyzer .TokenStream("title", new System.IO.StringReader(title));
System.String result = highlighter.GetBestFragments(tokenStream, text, 2, "...");
string tresult = highlighter.GetBestFragments(titkeStream, title, 0, "..");
//祛除標題結尾標記
if (tresult.Length > 10)
tresult = tresult.Remove(tresult.Length - 10, 10);
if (string.IsNullOrEmpty(tresult))
tresult = title.Remove(title.Length - 10, 10);
//未標注內容讀取
if (string.IsNullOrEmpty(result))
{
if (text.Length > 100)
result = text.Substring(0, 100);
else
result = text;
}
if (result.Length < text.Length)
result = result + "...";
}
這是使用lucene.net的自帶分詞器StandardAnalyzer，有個弊端是一個字就認為是一個詞?，F(xiàn)在我們要用自己的中文詞義分詞器——MyAnalyzer的話，問題來了。Highlighter一直報錯。為什么會出現(xiàn)這種情況呢？那是因為，中文分詞器，分詞完成一般有分割符號。比如，對于詞“滬江英語”，分割完成后變成"滬江 英語"。也就是說，返回的Token是(滬江,0,2)(英語,3,5)。而如果用Highlighter，它想要取得詞是(滬江,0,2)(英語,2,4)，這個就是因為空格而產生的偏差，引起了Highlighter的報錯。修改Highlighter比較麻煩，要對分詞器做特殊處理也比較難?？梢允褂米值浣忉屍?，分割結果后再高亮。比如，取得"滬江英語"的分詞結果"滬江 英語",然后，把分詞結果傳遞給Highlighter。弊端是，搜索結果會憑空出現(xiàn)很多空格。俺們懶人自然有懶人的解決辦法。那就是用MyAnalyzer實現(xiàn)索引和搜索，用StandardAnalyzer實現(xiàn)高亮：
//定義多條件搜索分析器
BooleanQuery bquery = new BooleanQuery();
//定義分詞器
Analyzer analyzer = new MyAnalyzer();
Analyzer highanalyzer = new StandardAnalyzer();
MultiFieldQueryParser parser = new MultiFieldQueryParser(new string[] { "title", "content" }, highanalyzer );
Query highquery = parser .Parse(key);
//多條件搜索拆分器
MultiFieldQueryParser parser = new MultiFieldQueryParser(new string[] { "title", "content" }, analyzer);
query = parser.Parse(key);
bquery.Add(query, BooleanClause.Occur.MUST);
DateTime now = DateTime.Now;
//初始化搜索器
//實現(xiàn)分布式搜索
List<IndexReader> indexreaders = new List<IndexReader>();
string[] dirs = Directory.GetDirectories(dir);
if (searchtype == SearchType.None)
{
foreach (string item in dirs)
{
//System.Web.HttpContext.Current.Response.Write(item);
indexreaders.Add(IndexReader.Open(Path.Combine(Path.Combine(dir, item), "Save")));
}
}
else
{
//System.Web.HttpContext.Current.Response.Write(searchtype.ToString());
indexreaders.Add(IndexReader.Open(Path.Combine(Path.Combine(dir, searchtype.ToString()), "Save")));
}
MultiReader reader = new MultiReader(indexreaders.ToArray());
indexSearcher = new IndexSearcher(reader);
Hits hits = null;
hits = indexSearcher.Search(bquery);
timer = (DateTime.Now - now).TotalMilliseconds;
int count = hits.Length();
/* 計算顯示的條目 */
int start = (pageNo - 1) * 10;
int end = pageNo * 10 > count ? count : pageNo * 10;
//Response.Write(readerhelper.MyQuery.ToString());
/* 語法高亮顯示設置 */
Highlighter highlighter = new Highlighter(new QueryScorer(highquery));
highlighter.SetTextFragmenter(new SimpleFragmenter(100));
for (int i = start; i < end; i++)
{
Lucene.Net.Documents.Document doc = hits.Doc(i);
System.String text = doc.Get("content");
//添加結尾，保證結尾特殊符號不被過濾
string title = doc.Get("title") + "+aaaaaaaaa";
Lucene.Net.Analysis.TokenStream tokenStream = highanalyzer .TokenStream("content", new System.IO.StringReader(text));
Lucene.Net.Analysis.TokenStream titkeStream = highanalyzer .TokenStream("title", new System.IO.StringReader(title));
System.String result = highlighter.GetBestFragments(tokenStream, text, 2, "...");
string tresult = highlighter.GetBestFragments(titkeStream, title, 0, "..");
//祛除標題結尾標記
if (tresult.Length > 10)
tresult = tresult.Remove(tresult.Length - 10, 10);
if (string.IsNullOrEmpty(tresult))
tresult = title.Remove(title.Length - 10, 10);
//未標注內容讀取
if (string.IsNullOrEmpty(result))
{
if (text.Length > 100)
result = text.Substring(0, 100);
else
result = text;
}
if (result.Length < text.Length)
result = result + "...";
}
這樣的結果還是很友好的，例如：
http://so.yeshj.com/so.aspx?key=%E6%B2%AA%E6%B1%9F%E6%97%A5%E8%AF%AD&h=%E6%B2%AA%E6%B1%9F%E6%97%A5%E8%AF%AD
全文完。
