public void SetUpReader()
{
SetWorkingDirectory();
var indexReader = DirectoryReader.Open(indexDirectory);
_maxDoc = indexReader.MaxDoc;
if (debug)
{
Console.WriteLine(_maxDoc);
}
indexReader.Dispose();
indexDirectory.Dispose();
}
/// <summary>
/// Create a new facet store within the specified directory.
/// </summary>
/// <param name="directoryFacetInfo">The directory infomation where the facet files are to be placed.</param>
public void CreateMultiFacetIndex(DirectoryInfo directoryFacetInfo)
{
DirectoryTaxonomyWriter facetWriter = null;
Lucene.Net.Store.Directory directoryFacet = null;
try
{
// Create the facet writer.
directoryFacet = FSDirectory.Open(directoryFacetInfo);
facetWriter = new DirectoryTaxonomyWriter(directoryFacet, IndexWriterConfig.OpenMode_e.CREATE);
// Commit the index.
facetWriter.Commit();
}
catch (Exception)
{
throw;
}
finally
{
if (facetWriter != null)
{
facetWriter.Dispose();
}
if (directoryFacet != null)
{
directoryFacet.Dispose();
}
}
}
public virtual void TestNoExtraFiles()
{
Directory directory = NewDirectory();
IndexWriter writer = new IndexWriter(directory, (IndexWriterConfig)NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random)).SetMaxBufferedDocs(2));
for (int iter = 0; iter < 7; iter++)
{
if (VERBOSE)
{
Console.WriteLine("TEST: iter=" + iter);
}
for (int j = 0; j < 21; j++)
{
Document doc = new Document();
doc.Add(NewTextField("content", "a b c", Field.Store.NO));
writer.AddDocument(doc);
}
writer.Dispose();
TestIndexWriter.AssertNoUnreferencedFiles(directory, "testNoExtraFiles");
// Reopen
writer = new IndexWriter(directory, (IndexWriterConfig)NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random)).SetOpenMode(OpenMode.APPEND).SetMaxBufferedDocs(2));
}
writer.Dispose();
directory.Dispose();
}
public virtual void TestTotalBytesSize()
{
Directory d = NewDirectory();
if (d is MockDirectoryWrapper)
{
((MockDirectoryWrapper)d).Throttling = Throttling.NEVER;
}
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random));
iwc.SetMaxBufferedDocs(5);
iwc.SetMergeScheduler(new TrackingCMS());
if (TestUtil.GetPostingsFormat("id").Equals("SimpleText", StringComparison.Ordinal))
{
// no
iwc.SetCodec(TestUtil.AlwaysPostingsFormat(new Lucene41PostingsFormat()));
}
RandomIndexWriter w = new RandomIndexWriter(Random, d, iwc);
for (int i = 0; i < 1000; i++)
{
Document doc = new Document();
doc.Add(new StringField("id", "" + i, Field.Store.NO));
w.AddDocument(doc);
if (Random.NextBoolean())
{
w.DeleteDocuments(new Term("id", "" + Random.Next(i + 1)));
}
}
Assert.IsTrue(((TrackingCMS)w.IndexWriter.Config.MergeScheduler).TotMergedBytes != 0);
w.Dispose();
d.Dispose();
}
private void button1_Click(object sender, EventArgs e)
{
String indexPath = @"inputFileFS";
if (System.IO.Directory.Exists(indexPath))
{
System.IO.Directory.Delete(indexPath, true);
}
luceneWorkingSet = FSDirectory.Open(indexPath);
StandardAnalyzer analyzer = new StandardAnalyzer(Lucene.Net.Util.Version.LUCENE_30);
IndexWriter writer = new IndexWriter(luceneWorkingSet, analyzer, true, IndexWriter.MaxFieldLength.LIMITED);
StreamReader reader = new StreamReader(@"InputFS\sampleText.txt");
Document doc = new Document();
//Start Here
doc.Add(new Field("lyrics", reader.ReadToEnd(), Field.Store.YES, Field.Index.ANALYZED));
doc.Add(new Field("songName", "Tonight", Field.Store.YES, Field.Index.NOT_ANALYZED));
writer.AddDocument(doc);
writer.Optimize();
writer.Dispose();
luceneWorkingSet.Dispose();
}
/// <summary>
/// Remove text from the existing index.
/// </summary>
/// <param name="directoryIndexInfo">The directory infomation where the index files are located.</param>
/// <param name="names">An array of unique names for the text.</param>
public void RemoveText(DirectoryInfo directoryIndexInfo, string[] names)
{
Lucene.Net.Index.IndexWriter writer = null;
Lucene.Net.Store.Directory directory = null;
try
{
// If exists.
if (names != null && names.Length > 0)
{
// Create the analyzer.
SimpleAnalyzer simpleAnalyzer = new Analyzer.SimpleAnalyzer();
StandardAnalyzer standardAnalyzer = new Analyzer.StandardAnalyzer(simpleAnalyzer);
// Create the index writer.
directory = FSDirectory.Open(directoryIndexInfo);
IndexWriterConfig indexConfig = new IndexWriterConfig(Lucene.Net.Util.LuceneVersion.LUCENE_48, standardAnalyzer);
indexConfig.SetOpenMode(IndexWriterConfig.OpenMode_e.APPEND);
// Open existing or create new.
writer = new IndexWriter(directory, indexConfig);
// Create the query.
List <Query> queries = new List <Query>();
// For each name.
foreach (string name in names)
{
// Create the query.
BooleanQuery query = new BooleanQuery();
query.Add(new TermQuery(new Term("textname", name.ToLower())), BooleanClause.Occur.MUST);
// Add the query.
queries.Add(query);
}
// Delete the text.
writer.DeleteDocuments(queries.ToArray());
// Commit the index.
writer.Commit();
}
}
catch (Exception)
{
throw;
}
finally
{
if (writer != null)
{
writer.Dispose();
}
if (directory != null)
{
directory.Dispose();
}
}
}
public void TearDown()
{
_cwsDir.Dispose();
_pdfDir.Dispose();
_simpleDir.Dispose();
_conventionDir.Dispose();
}
/// <summary>
/// Performs application-defined tasks associated with freeing, releasing, or resetting unmanaged resources.
/// </summary>
public void Dispose()
{
if (_directory != null)
{
_directory.Dispose();
}
}
public override void TestTearDown()
{
base.TestTearDown();
_luceneDir.Dispose();
UmbracoExamineSearcher.DisableInitializationCheck = null;
BaseUmbracoIndexer.DisableInitializationCheck = null;
}
public List <string> Search(string searchword)
{
Lucene.Net.Store.Directory directory = FSDirectory.Open(Settings.Default.IndexDirectory);
Analyzer analyzer = new StandardAnalyzer(Lucene.Net.Util.Version.LUCENE_30);
IndexReader indexReader = IndexReader.Open(directory, true);
Searcher indexSearch = new IndexSearcher(indexReader);
var queryParser = new QueryParser(Lucene.Net.Util.Version.LUCENE_30, "Content", analyzer);
var query = queryParser.Parse(searchword);
List <string> infosearch = new List <string>();
infosearch.Add("Searching for: " + query.ToString().Replace("Content:", ""));
TopDocs resultDocs = indexSearch.Search(query, indexReader.MaxDoc);
infosearch.Add("Results Found: " + resultDocs.TotalHits);
var hits = resultDocs.ScoreDocs;
foreach (var hit in hits)
{
var documentFromSearcher = indexSearch.Doc(hit.Doc);
infosearch.Add(documentFromSearcher.Get("Link"));
}
indexSearch.Dispose();
directory.Dispose();
return(infosearch);
}
public void Dispose()
{
isClosing = true;
CloseWriter();
CloseReader();
try
{
if (index != null)
{
index.Dispose();
}
}
catch (Exception ex)
{
Log.ErrorFormat("Failed to dispose Lucene search index.", ex);
}
try
{
if (analyzer != null)
{
analyzer.Close();
analyzer.Dispose();
}
}
catch (Exception ex)
{
Log.ErrorFormat("Failed to close/dispose Lucene search analyzer.", ex);
}
}
private void DoTest(Random random, TextWriter @out, bool useCompoundFiles, int MAX_DOCS)
{
Store.Directory directory = NewDirectory();
Analyzer analyzer = new MockAnalyzer(random);
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, analyzer);
MergePolicy mp = conf.MergePolicy;
mp.NoCFSRatio = useCompoundFiles ? 1.0 : 0.0;
IndexWriter writer = new IndexWriter(directory, conf);
if (Verbose)
{
Console.WriteLine("TEST: now build index MAX_DOCS=" + MAX_DOCS);
}
for (int j = 0; j < MAX_DOCS; j++)
{
Documents.Document d = new Documents.Document();
d.Add(NewTextField(PRIORITY_FIELD, HIGH_PRIORITY, Field.Store.YES));
d.Add(NewTextField(ID_FIELD, Convert.ToString(j), Field.Store.YES));
writer.AddDocument(d);
}
writer.Dispose();
// try a search without OR
IndexReader reader = DirectoryReader.Open(directory);
IndexSearcher searcher = NewSearcher(reader);
Query query = new TermQuery(new Term(PRIORITY_FIELD, HIGH_PRIORITY));
@out.WriteLine("Query: " + query.ToString(PRIORITY_FIELD));
if (Verbose)
{
Console.WriteLine("TEST: search query=" + query);
}
Sort sort = new Sort(SortField.FIELD_SCORE, new SortField(ID_FIELD, SortFieldType.INT32));
ScoreDoc[] hits = searcher.Search(query, null, MAX_DOCS, sort).ScoreDocs;
PrintHits(@out, hits, searcher);
CheckHits(hits, MAX_DOCS, searcher);
// try a new search with OR
searcher = NewSearcher(reader);
hits = null;
BooleanQuery booleanQuery = new BooleanQuery();
booleanQuery.Add(new TermQuery(new Term(PRIORITY_FIELD, HIGH_PRIORITY)), Occur.SHOULD);
booleanQuery.Add(new TermQuery(new Term(PRIORITY_FIELD, MED_PRIORITY)), Occur.SHOULD);
@out.WriteLine("Query: " + booleanQuery.ToString(PRIORITY_FIELD));
hits = searcher.Search(booleanQuery, null, MAX_DOCS, sort).ScoreDocs;
PrintHits(@out, hits, searcher);
CheckHits(hits, MAX_DOCS, searcher);
reader.Dispose();
directory.Dispose();
}
public void ShutDownWriter()
{
_tdIndexWriter.Optimize();
_tdIndexWriter.Dispose();
_tdIndexDirectory.Dispose();
_tdIndexWriter = null;
_tdIndexDirectory = null;
}
public void TestTearDown()
{
//set back to 100
_indexer.OptimizationCommitThreshold = 100;
_indexer.Dispose();
_searcher.Dispose();
_luceneDir.Dispose();
}
protected override void Dispose(bool disposing)
{
if (disposing)
{
isDisposed = true;
cacheDirectory.Dispose();
}
}
public void Close()
{
_analyzer.Close();
_analyzer.Dispose();
_searcher.Dispose();
_directory.Dispose();
_logger.Info("Searcher is closed.");
}
public void mychaxun()
{
//if (string.IsNullOrEmpty(this.txtSearch.Text))
// MessageBox.Show("请输入搜索的文本");
string selname = "贾宝玉";
StringBuilder sb = new StringBuilder();
Stopwatch sw = new Stopwatch();
sw.Start();
//索引库目录
Lucene.Net.Store.Directory dir = FSDirectory.Open(new System.IO.DirectoryInfo("F:\\doc"), new NoLockFactory());
IndexReader reader = IndexReader.Open(dir, true);
IndexSearcher search = null;
try
{
search = new IndexSearcher(reader);
QueryParser parser = new QueryParser(Lucene.Net.Util.Version.LUCENE_30, "body", new StandardAnalyzer(Lucene.Net.Util.Version.LUCENE_30));
Query query = parser.Parse(LuceneHelper.GetKeyWordSplid(selname));
//执行搜索，获取查询结果集对象
TopDocs ts = search.Search(query, null, 1000);
///获取命中的文档信息对象
ScoreDoc[] docs = ts.ScoreDocs;
sw.Stop();
//this.listBox.Items.Clear();
for (int i = 0; i < docs.Length; i++)
{
int docId = docs[i].Doc;
Document doc = search.Doc(docId);
//this.listBox.Items.Add(doc.Get("number") + "\r\n");
//this.listBox.Items.Add(doc.Get("body") + "\r\n");
//this.listBox.Items.Add("------------------------\r\n");
string a = doc.Get("number") + "\r\n";
string b = doc.Get("body") + "\r\n";
}
}
catch (Exception ex)
{
throw;
}
finally
{
if (search != null)
{
search.Dispose();
}
if (dir != null)
{
dir.Dispose();
}
}
//this.label.Text = "搜索用时:";
//this.timeBox.Text = sw.ElapsedMilliseconds + "毫秒";
}
public void myindexer()
{
#region //盘古分词
//Segment.Init();
//string str = "盘古分词demo2";
//Segment segment = new Segment();
//ICollection<WordInfo> words = segment.DoSegment(str);
#endregion
Stopwatch sw = new Stopwatch();
sw.Start();
IndexWriter writer = null;
Analyzer analyzer = new StandardAnalyzer(Lucene.Net.Util.Version.LUCENE_30);
//Analyzer pganalyzer = new PanGuAnalyzer();
Lucene.Net.Store.Directory dir = FSDirectory.Open(new System.IO.DirectoryInfo("F:\\lucenedata"));
try
{
////IndexReader:对索引进行读取的类。
//该语句的作用：判断索引库文件夹是否存在以及索引特征文件是否存在。
bool isCreate = !IndexReader.IndexExists(dir);
writer = new IndexWriter(dir, analyzer, isCreate, IndexWriter.MaxFieldLength.UNLIMITED);
//添加索引
{
Document doc = new Document();
//string path = System.IO.Directory.GetParent(System.IO.Directory.GetCurrentDirectory()).Parent.FullName + "F:\\doc\\红楼梦.txt";
string path = "F:\\doc\\红楼梦.txt";
string text = System.IO.File.ReadAllText(path, Encoding.Default);
//Field.Store.YES:表示是否存储原值。只有当Field.Store.YES在后面才能用doc.Get("number")取出值来.Field.Index. NOT_ANALYZED:不进行分词保存
doc.Add(new Field("number", 5.ToString(), Field.Store.YES, Field.Index.NOT_ANALYZED));
// Lucene.Net.Documents.Field.TermVector.WITH_POSITIONS_OFFSETS:不仅保存分词还保存分词的距离。
doc.Add(new Field("body", text, Field.Store.YES, Field.Index.ANALYZED, Field.TermVector.WITH_POSITIONS_OFFSETS));
writer.AddDocument(doc);
}
writer.Optimize();
sw.Stop();
}
catch (Exception ex)
{
throw;
}
finally
{
if (writer != null)
{
writer.Dispose();
}
if (dir != null)
{
dir.Dispose();
}
}
}
/// <summary>
/// Remove facet documents from the existing index.
/// </summary>
/// <param name="directoryIndexInfo">The directory infomation where the index files are located.</param>
/// <param name="directoryFacetInfo">The directory infomation where the facet files are to be placed.</param>
/// <param name="textNames">The array of names for text data.</param>
/// <param name="filePaths">The array of full paths (without root 'C:\'. e.g. 'temp/http/index.html') for file documents.</param>
public void RemoveMultiFacetDocuments(DirectoryInfo directoryIndexInfo, DirectoryInfo directoryFacetInfo, string[] textNames, string[] filePaths)
{
Lucene.Net.Index.IndexWriter writer = null;
DirectoryTaxonomyWriter facetWriter = null;
Lucene.Net.Store.Directory directory = null;
Lucene.Net.Store.Directory directoryFacet = null;
try
{
// Create the analyzer.
SimpleAnalyzer simpleAnalyzer = new Analyzer.SimpleAnalyzer();
StandardAnalyzer standardAnalyzer = new Analyzer.StandardAnalyzer(simpleAnalyzer);
// Create the index writer.
directory = FSDirectory.Open(directoryIndexInfo);
IndexWriterConfig indexConfig = new IndexWriterConfig(Lucene.Net.Util.LuceneVersion.LUCENE_48, standardAnalyzer);
indexConfig.SetOpenMode(IndexWriterConfig.OpenMode_e.APPEND);
// Open existing or create new.
writer = new IndexWriter(directory, indexConfig);
// Create the facet writer.
directoryFacet = FSDirectory.Open(directoryFacetInfo);
facetWriter = new DirectoryTaxonomyWriter(directoryFacet, IndexWriterConfig.OpenMode_e.APPEND);
// Create the delet query.
FacetFilter filter = new FacetFilter();
Query[] queries = filter.RemoveDocuments(textNames, filePaths);
writer.DeleteDocuments(queries);
// Commit the index.
writer.Commit();
facetWriter.Commit();
}
catch (Exception)
{
throw;
}
finally
{
if (writer != null)
{
writer.Dispose();
}
if (directory != null)
{
directory.Dispose();
}
}
}
private void button2_Click(object sender, EventArgs e)
{
if (string.IsNullOrEmpty(this.txtSearch.Text))
{
MessageBox.Show("请输入搜索的文本");
}
StringBuilder sb = new StringBuilder();
Stopwatch sw = new Stopwatch();
sw.Start();
//索引库目录
Lucene.Net.Store.Directory dir = FSDirectory.Open(new System.IO.DirectoryInfo("IndexDir"), new NoLockFactory());
IndexReader reader = IndexReader.Open(dir, true);
IndexSearcher search = null;
try
{
search = new IndexSearcher(reader);
QueryParser parser = new QueryParser(Lucene.Net.Util.Version.LUCENE_30, "body", new PanGuAnalyzer());
Query query = parser.Parse(LuceneHelper.GetKeyWordSplid(this.txtSearch.Text));
//执行搜索，获取查询结果集对象
TopDocs ts = search.Search(query, null, 1000);
///获取命中的文档信息对象
ScoreDoc[] docs = ts.ScoreDocs;
sw.Stop();
this.listBox1.Items.Clear();
for (int i = 0; i < docs.Length; i++)
{
int docId = docs[i].Doc;
Document doc = search.Doc(docId);
this.listBox1.Items.Add(doc.Get("number") + "\r\n");
this.listBox1.Items.Add(doc.Get("body") + "\r\n");
this.listBox1.Items.Add("------------------------\r\n");
}
}
catch (Exception ex)
{
throw;
}
finally
{
if (search != null)
{
search.Dispose();
}
if (dir != null)
{
dir.Dispose();
}
}
this.label1.Text = "搜索用时:" + sw.ElapsedMilliseconds + "毫秒";
}
public void Dispose()
{
if (mIndexWriter != null)
{
mIndexWriter.Dispose();
mIndexWriter = null;
}
if (mIndex != null)
{
mIndex.Dispose();
mIndex = null;
}
}
public void Dispose()
{
LogManager.GetCurrentClassLogger().Info("Stopping Lucene indexing services.");
if (PackageFileSystemWatcher != null)
{
PackageFileSystemWatcher.Dispose();
}
PackageIndexer.Dispose();
Provider.Dispose();
LuceneDirectory.Dispose();
}
public virtual void Dispose()
{
if (m_searcherMgr != null)
{
m_searcherMgr.Dispose();
m_searcherMgr = null;
}
if (writer != null)
{
writer.Dispose();
dir.Dispose();
writer = null;
}
}
public void Dispose()
{
if (searcherMgr != null)
{
searcherMgr.Dispose();
searcherMgr = null;
}
if (writer != null)
{
writer.Dispose();
dir.Dispose();
writer = null;
}
}
public void TestRecreateTaxonomy()
{
replicator.Publish(CreateRevision(1));
client.UpdateNow();
// recreate index and taxonomy
Directory newTaxo = NewDirectory();
new DirectoryTaxonomyWriter(newTaxo).Dispose();
publishTaxoWriter.ReplaceTaxonomy(newTaxo);
publishIndexWriter.DeleteAll();
replicator.Publish(CreateRevision(2));
client.UpdateNow();
newTaxo.Dispose();
}
public void BuildIndex(IEnumerable <SampleDataFileRow> dataToIndex)
{
foreach (var sampleDataFileRow in dataToIndex)
{
Document doc = new Document();
doc.Add(new Field("LineNumber",
sampleDataFileRow.LineNumber.ToString(),
Field.Store.YES,
Field.Index.NOT_ANALYZED));
doc.Add(new Field("LineText",
sampleDataFileRow.LineText,
Field.Store.YES,
Field.Index.ANALYZED));
doc.Add(new Field("ID",
sampleDataFileRow.Id.ToString(),
Field.Store.YES,
Field.Index.NOT_ANALYZED));
writer.AddDocument(doc);
}
writer.Optimize();
writer.Commit();
writer.Dispose();
luceneIndexDirectory.Dispose();
}
public void Dispose()
{
CloseWriter();
CloseReader();
if (index != null)
{
index.Dispose();
}
if (analyzer != null)
{
analyzer.Close();
analyzer.Dispose();
}
}
protected virtual void Dispose(bool disposing)
{
if (!_isDisposed)
{
if (disposing)
{
_writer.Commit();
_writer.Dispose();
_directory.Dispose();
_analyzer.Dispose();
}
_isDisposed = true;
}
}
public static List <Post> getPitanjaByTag(int TAGID)
{
Directory directoryPronadjenaPitanjaTagovi = Data.Lucene.Indexing.GetDirectoryPitanjaTagovi();
Analyzer analyzer = new StandardAnalyzer(Version.LUCENE_30);
IndexReader indexReader = IndexReader.Open(directoryPronadjenaPitanjaTagovi, true);
Searcher searcher = new IndexSearcher(indexReader);
//var queryParser = new QueryParser(Version.LUCENE_30, "Naslov", analyzer);
var queryParser = new MultiFieldQueryParser(Version.LUCENE_30, new[] { "PostID", "TagID", "DatumKreiranja" }, analyzer);
var query = queryParser.Parse(Convert.ToString(TAGID)); // Rastavljanje rečenice na rijeci
TopDocs pronadjeno = searcher.Search(query, indexReader.MaxDoc);
List <Post> PitanjaTagovi = new List <Post>();
if (pronadjeno != null)
{
var hits = pronadjeno.ScoreDocs;
foreach (var hit in hits)
{
var documentFromSearcher = searcher.Doc(hit.Doc);
using (TriglavBL temp = new TriglavBL())
{
PitanjaTagovi.Add(temp.getPitanjeByID(Convert.ToInt32(documentFromSearcher.Get("PostID"))));
}
}
searcher.Dispose();
directoryPronadjenaPitanjaTagovi.Dispose();
return(PitanjaTagovi);
}
else
{
searcher.Dispose();
directoryPronadjenaPitanjaTagovi.Dispose();
return(PitanjaTagovi);
}
}
public virtual void TestNegativeQueryBoost()
{
Query q = new TermQuery(new Term("foo", "bar"));
q.Boost = -42f;
Assert.AreEqual(-42f, q.Boost, 0.0f);
Store.Directory directory = NewDirectory();
try
{
Analyzer analyzer = new MockAnalyzer(Random);
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, analyzer);
IndexWriter writer = new IndexWriter(directory, conf);
try
{
Documents.Document d = new Documents.Document();
d.Add(NewTextField("foo", "bar", Field.Store.YES));
writer.AddDocument(d);
}
finally
{
writer.Dispose();
}
IndexReader reader = DirectoryReader.Open(directory);
try
{
IndexSearcher searcher = NewSearcher(reader);
ScoreDoc[] hits = searcher.Search(q, null, 1000).ScoreDocs;
Assert.AreEqual(1, hits.Length);
Assert.IsTrue(hits[0].Score < 0, "score is not negative: " + hits[0].Score);
Explanation explain = searcher.Explain(q, hits[0].Doc);
Assert.AreEqual(hits[0].Score, explain.Value, 0.001f, "score doesn't match explanation");
Assert.IsTrue(explain.IsMatch, "explain doesn't think doc is a match");
}
finally
{
reader.Dispose();
}
}
finally
{
directory.Dispose();
}
}
public void GRandom()
{
int numDocs = TestUtil.NextInt(Random(), (10 * RANDOM_MULTIPLIER),
(100 * RANDOM_MULTIPLIER));
Directory dir = null;
RandomIndexWriter writer = null;
IndexReader ir = null;
try
{
dir = NewDirectory();
writer = new RandomIndexWriter(Random(), dir, new MockAnalyzer(Random(),
MockTokenizer.WHITESPACE, false), Similarity, TimeZone);
int maxLength = TestUtil.NextInt(Random(), 5, 50);
List<string> originals = new List<string>(numDocs);
List<string[]> breaks = new List<string[]>(numDocs);
for (int i = 0; i < numDocs; i++)
{
string orig = "";
if (Random().nextBoolean())
{
while (!GoodTestString(orig))
{
orig = TestUtil.RandomSimpleString(Random(), maxLength);
}
}
else
{
while (!GoodTestString(orig))
{
orig = TestUtil.RandomUnicodeString(Random(), maxLength);
}
}
originals.Add(orig);
int totalLength = orig.CodePointCount(0, orig.Length);
int breakAt = orig.OffsetByCodePoints(0,
TestUtil.NextInt(Random(), 1, totalLength - 1));
string[] broken = new string[2];
broken[0] = orig.Substring(0, breakAt - 0);
broken[1] = orig.Substring(breakAt);
breaks.Add(broken);
Document doc = new Document();
doc.Add(NewTextField("random_break", broken[0] + " " + broken[1],
Field.Store.NO));
doc.Add(NewTextField("random_combine", orig, Field.Store.NO));
writer.AddDocument(doc);
}
writer.Commit();
writer.Dispose();
ir = DirectoryReader.Open(dir);
WordBreakSpellChecker wbsp = new WordBreakSpellChecker();
wbsp.MaxChanges = (1);
wbsp.MinBreakWordLength = (1);
wbsp.MinSuggestionFrequency = (1);
wbsp.MaxCombineWordLength = (maxLength);
for (int i = 0; i < originals.size(); i++)
{
string orig = originals[i];
string left = breaks[i][0];
string right = breaks[i][1];
{
Term term = new Term("random_break", orig);
SuggestWord[][] sw = wbsp.SuggestWordBreaks(term, originals.size(),
ir, SuggestMode.SUGGEST_ALWAYS,
WordBreakSpellChecker.BreakSuggestionSortMethod.NUM_CHANGES_THEN_MAX_FREQUENCY);
bool failed = true;
foreach (SuggestWord[] sw1 in sw)
{
assertTrue(sw1.Length == 2);
if (sw1[0].String.equals(left) && sw1[1].String.equals(right))
{
failed = false;
}
}
assertFalse("Failed getting break suggestions\n >Original: "
+ orig + "\n >Left: " + left + "\n >Right: " + right, failed);
}
{
Term[] terms = {new Term("random_combine", left),
new Term("random_combine", right)};
CombineSuggestion[] cs = wbsp.SuggestWordCombinations(terms,
originals.size(), ir, SuggestMode.SUGGEST_ALWAYS);
bool failed = true;
foreach (CombineSuggestion cs1 in cs)
{
assertTrue(cs1.OriginalTermIndexes.Length == 2);
if (cs1.Suggestion.String.equals(left + right))
{
failed = false;
}
}
assertFalse("Failed getting combine suggestions\n >Original: "
+ orig + "\n >Left: " + left + "\n >Right: " + right, failed);
}
}
}
catch (Exception e)
{
throw e;
}
finally
{
try
{
ir.Dispose();
}
catch (Exception /*e1*/) { }
try
{
writer.Dispose();
}
catch (Exception /*e1*/) { }
try
{
dir.Dispose();
}
catch (Exception /*e1*/) { }
}
}
private RAMDirectory(Directory dir, bool closeDir, IOContext context)
: this()
{
foreach (string file in dir.ListAll())
{
dir.Copy(this, file, file, context);
}
if (closeDir)
{
dir.Dispose();
}
}
/// <summary>
/// Perform search
/// </summary>
/// <param name="query"></param>
/// <param name="startIndex"></param>
/// <param name="blockSize"></param>
/// <param name="indexDirEs"></param>
/// <param name="indexDirEn"></param>
/// <param name="sortBy"></param>
/// <returns></returns>
public List<IssueDocument> MedesSearch(Query query, int startIndex, int blockSize, Directory indexDirEs, Directory indexDirEn, Directory indexDirHe, string sortBy)
{
#if DEBUG
T.TraceMessage(string.Format("Begin search , query: '{0}'", query.ToString()));
#endif
List<IssueDocument> result = new List<IssueDocument>();
try
{
// build a multi searcher across the 2 indexes
MultiSearcher mSearcher = CombineSearchers(indexDirEs, indexDirEn, indexDirHe);
TopDocs tDocs = null;
int iterateLast = startIndex + blockSize;
string customScoreField = "article_id";
FieldScoreQuery dateBooster = new FieldScoreQuery(customScoreField, FieldScoreQuery.Type.FLOAT);
CustomScoreQuery customQuery = new CustomScoreQuery(query, dateBooster);
tDocs = mSearcher.Search(customQuery, 1000);
//ScoreDoc[] hits = tpDcs.scoreDocs;
if (startIndex + blockSize > tDocs.TotalHits) iterateLast = tDocs.TotalHits;
for (int i = startIndex; i < iterateLast; i++)
{
// Document hitDoc = mSearcher.Doc(hits[i].doc);
Document hitDoc = mSearcher.Doc(i);
result.Add(new IssueDocument() { Id = Int32.Parse(hitDoc.Get("issue_id").ToString())});
}
// close the searcher and indexes
mSearcher.Dispose();
indexDirEs.Dispose();
indexDirEn.Dispose();
indexDirHe.Dispose();
}
catch (Exception ex)
{
T.TraceError("Error MedesSearch, query '{0}'", query.ToString());
T.TraceError(ex);
throw ex;
}
return result;
}
// LUCENE-1468
private void CheckDirectoryFilter(Directory dir)
{
string name = "file";
try
{
dir.CreateOutput(name, NewIOContext(Random())).Dispose();
Assert.IsTrue(SlowFileExists(dir, name));
Assert.IsTrue(Arrays.AsList(dir.ListAll()).Contains(name));
}
finally
{
dir.Dispose();
}
}
