lucene.net简单tutorial

最新推荐文章于 2024-08-25 09:34:14 发布

转载最新推荐文章于 2024-08-25 09:34:14 发布 · 907 阅读

.NET 同时被 2 个专栏收录

201 篇文章

订阅专栏

11 篇文章

订阅专栏

本文详细介绍了如何使用Lucene.NET在C#中创建全文搜索索引，包括从Apache Lucene.Net-2.9.2-incubating.bin.zip文件中提取库，创建文档并将其写入索引。

Step 1 - Create a new Console Application

Then extract the Lucene.Net.dll from the Apache-Lucene.Net-2.9.2-incubating.bin.zip file into yourlib folder.

You'll notice lots of other bits in this zip file. Especially of interest to you later might be the stuff in thecontrib folder. I might get to that in a later tutorial, but for now lets keep it simple.

Step 2 - Add a Reference to the Lucene.net.dll

Your references should look like this

Step 3 - Create a Document

Ok the next step is to create a simple Document with the appropriatefields which you'll want to search on. In our case we're going to create 3 cars. a Ford Fiesta, a Ford Focus and a Vauxhall Astra.

 
   01using  System; 
 
   02using  System.IO; 
 
   03using  Lucene.Net.Analysis; 
 
   04using  Lucene.Net.Analysis.Standard; 
 
   05using  Lucene.Net.Documents; 
 
   06using  Lucene.Net.Index; 
 
   07using  Lucene.Net.Store; 
 
   08using  Directory = Lucene.Net.Store.Directory; 
 
   09using  Version = Lucene.Net.Util.Version; 
 
   10  
 
   11namespace  LuceneNet.App 
 
   12{ 
 
   13    classProgram 
 
   14    {  
 
   15        staticvoid Main(string[] args)
 
   16        {
 
   17            var fordFiesta =new Document();
 
   18            fordFiesta.Add(newField("Id","1", Field.Store.YES, Field.Index.NOT_ANALYZED));
 
   19            fordFiesta.Add(newField("Make","Ford", Field.Store.YES, Field.Index.ANALYZED));
 
   20            fordFiesta.Add(newField("Model","Fiesta", Field.Store.YES, Field.Index.ANALYZED));
 
   21  
 
   22            var fordFocus =new Document();
 
   23            fordFocus.Add(newField("Id","2", Field.Store.YES, Field.Index.NOT_ANALYZED));
 
   24            fordFocus.Add(newField("Make","Ford", Field.Store.YES, Field.Index.ANALYZED));
 
   25            fordFocus.Add(newField("Model","Focus", Field.Store.YES, Field.Index.ANALYZED));
 
   26  
 
   27            var vauxhallAstra =new Document();
 
   28            vauxhallAstra.Add(newField("Id","3", Field.Store.YES, Field.Index.NOT_ANALYZED));
 
   29            vauxhallAstra.Add(newField("Make","Vauxhall", Field.Store.YES, Field.Index.ANALYZED));
 
   30            vauxhallAstra.Add(newField("Model","Astra", Field.Store.YES, Field.Index.ANALYZED));
 
   31  
 
   32  
 
   33  
 
   34            Directory directory = FSDirectory.Open(newDirectoryInfo(Environment.CurrentDirectory + "\\LuceneIndex"));
 
   35            Analyzer analyzer =new StandardAnalyzer(Version.LUCENE_29);
 
   36  
 
   37  
 
   38            var writer =new IndexWriter(directory, analyzer,true, IndexWriter.MaxFieldLength.LIMITED);
 
   39            writer.AddDocument(fordFiesta);
 
   40            writer.AddDocument(fordFocus);
 
   41            writer.AddDocument(vauxhallAstra);
 
   42  
 
   43            writer.Optimize();                        
 
   44            writer.Close();
 
   45              
 
   46  
 
   47              
 
   48        }
 
   49    }  
 
   50}

In the code above you can see we firstly create a series of Document types. For each document we then define a series ofFields. You might think this looks similar to SQL where you create the table schema and eachDocument is like a row in the table! There's an important difference here, documents do not have a schema. If we'd wanted to I could of given the 2 Ford documents an extra field calledSpecialFordField which wouldn't have existed at all on the Vauxhall documents.

Step 4 - Create a Directory

Now we need to sort out where we're going to store our index:

`1`	`Directory directory = FSDirectory.Open(newDirectoryInfo(Environment.CurrentDirectory +` `"\\LuceneIndex"));`

The above code creates a Directory. this is the place in which we store orwrite our Index to. In this case we use the FSDirectory.Open() factory method to create an Lucene Index on the FileSystem in a folder/directory calledLuceneIndex. We could of equally created new RamDirectory() and just stored our index in RAM for super high performance but Lucene is so fast, that for the most part this is unecessary.

Step 5 - The Analyzer

`1`	`Analyzer analyzer =` `newStandardAnalyzer(Version.LUCENE_29);`

We next create a new Analyzer which essentially turns our text intoTokens which are stored in our Index.

Step 6 - Writing the Documents to the Index

Finally we actually have to write the Documents to our Index

 
   1var writer = newIndexWriter(directory, analyzer, true, IndexWriter.MaxFieldLength.LIMITED);
 
   2            writer.AddDocument(fordFiesta);
 
   3            writer.AddDocument(fordFocus);
 
   4            writer.AddDocument(vauxhallAstra);
 
   5  
 
   6            writer.Optimize();                        
 
   7            writer.Close();

We define an IndexWriter telling it to use our Directory and Analyzer defined above. We then add all the documents to theIndexWriter call the writer.Optimize() which essentally does the semi-equivilent of a defrag on the index and then finallywriter.Close(). At this point we have a perfectly good index to search.

Step 7 - Opening the Index for Searching

Ok now let's actually search our newly created index.

`1`	`IndexReader indexReader = IndexReader.Open(directory,true);`

`2`	`Searcher indexSearch =new` `IndexSearcher(indexReader);`

We firstly need to open an IndexReader here we're passing in theDirectory we created above and wrote our Documents into. AnIndexReader simply reads the index it doesn't do the magic search part. For that we need aSearcher in our case a IndexSearcher which obviously needs to read our index internally.

Step 8 - Creating our Search Query

Prepare to search with a simple search query. The reality is that it's not quite as simple as typing in a Google query, though it's not far off.

`1`	`var queryParser =` `newQueryParser(Version.LUCENE_29,` `"Make", analyzer);`

`2`	`var query = queryParser.Parse("Ford");`

Here we're creating a QueryParser and specifying that we want to search the Make field of our Index. It is totally possible to search more than one field but for simplicity I'm not going to do that here. We also use the sameStandardAnalyzer we use above to apply the same tokenization to our query.

Finally we parse our raw query Ford. So hopefully we'll be able to find some Ford cars in our Index specifically in theMake field.

Step 9 - Performing the Search

 
   1Console.WriteLine("Searching for: "+ query.ToString()); 
 
   2            TopDocs resultDocs = indexSearch.Search(query, indexReader.MaxDoc());
 
   3              
 
   4            Console.WriteLine("Results Found: "+ resultDocs.totalHits);

Above we perform our search and return the TopDocs that match. Hopefully this should be2 results if yours is anything like mine. There are indeed 2 Fords in our index.

Step 10 - Displaying our Search Results

Last part. we just quickly loop through and print out the cars that match.

 
   1var hits = resultDocs.scoreDocs;  
 
   2            foreach(var hit inhits)  
 
   3            {                
 
   4                var documentFromSearcher = indexSearch.Doc(hit.doc);
 
   5                Console.WriteLine(documentFromSearcher.Get("Make") +" " + documentFromSearcher.Get("Model"));
 
   6            }
 
   7  
 
   8            indexSearch.Close();
 
   9            directory.Close();