public override BitArray Bits(IndexReader reader)
{
// reader.GetVersion could be used to cache
// Debug.WriteLine(reader.GetVersion()); // could be used to cache
// if (cached reader == reader && _revFirst ==
if (_revFirst == All || _revLast == All) // optimization
{
return(new BitArray(reader.MaxDoc(), true));
}
BitArray last_bits = new BitArray(reader.MaxDoc(), false);
TermEnum t = reader.Terms(new Term(FieldName.RevisionLast, _revFirst.ToString(RevFormat)));
TermDocs d = reader.TermDocs();
//if (t.SkipTo((new Term(FieldName.RevisionLast, revision.ToString(RevFormat))))) // extremely slow
if (t.Term() != null)
{
while (t.Term().Field() == FieldName.RevisionLast)
{
d.Seek(t);
while (d.Next())
{
last_bits[d.Doc()] = true;
}
if (!t.Next())
{
break;
}
}
}
// optimization, skip if we just using the head revision
if (_revLast == Head)
{
return(last_bits);
}
BitArray first_bits = new BitArray(reader.MaxDoc(), true);
t = reader.Terms(new Term("rev_first", (_revLast + 1).ToString(RevFormat)));
//if (t.SkipTo((new Term("rev_first", (revision + 1).ToString(RevFormat))))) // extremely slow
if (t.Term() != null)
{
while (t.Term().Field() == "rev_first")
{
d.Seek(t);
while (d.Next())
{
first_bits[d.Doc()] = false;
}
if (!t.Next())
{
break;
}
}
}
return(last_bits.And(first_bits));
}
private static void IndexFile(IndexWriter writer, TermEnum uidIter, FileInfo file, Operation operation)
{
if (file.FullName.EndsWith(".html") || file.FullName.EndsWith(".htm") || file.FullName.EndsWith(".txt"))
{
// We've found a file we should index.
if (operation == Operation.IncrementalReindex ||
operation == Operation.RemoveStale)
{
// We should only get here with an open uidIter.
Debug.Assert(uidIter != null, "Expected uidIter != null for operation " + operation);
var uid = HTMLDocument.Uid(file); // construct uid for doc
while (uidIter.Term != null && uidIter.Term.Field == "uid" && String.CompareOrdinal(uidIter.Term.Text, uid) < 0)
{
if (operation == Operation.RemoveStale)
{
Console.Out.WriteLine("deleting " + HTMLDocument.Uid2url(uidIter.Term.Text));
writer.DeleteDocuments(uidIter.Term);
}
uidIter.Next();
}
// The uidIter TermEnum should now be pointing at either
// 1) a null term, meaning there are no more uids to check.
// 2) a term matching the current file.
// 3) a term not matching us.
if (uidIter.Term != null && uidIter.Term.Field == "uid" && String.CompareOrdinal(uidIter.Term.Text, uid) == 0)
{
// uidIter points to the current document, we should move one
// step ahead to keep state consistant, and carry on.
uidIter.Next();
}
else if (operation == Operation.IncrementalReindex)
{
// uidIter does not point to the current document, and we're
// currently indexing documents.
var doc = HTMLDocument.Document(file);
Console.Out.WriteLine("adding " + doc.Get("path"));
writer.AddDocument(doc);
}
}
else
{
// We're doing a complete reindexing. We aren't using uidIter,
// but for completeness we assert that it's null (as expected).
Debug.Assert(uidIter == null, "Expected uidIter == null for operation == " + operation);
var doc = HTMLDocument.Document(file);
Console.Out.WriteLine("adding " + doc.Get("path"));
writer.AddDocument(doc);
}
}
}
//public IEnumerable<Article> LeaderResults(int DocID, List<DocumentClusters> Clusters)
//{
// IEnumerable<Article> Articles = ArticleReader.ReadArticles(@"Data\cacm.all");
// List<Article> Results = new List<Article>();
//// foreach (DocumentClusters cluster in Clusters)
//// {
//// if (cluster.LeaderDocumentID == DocID)
//// {
//// foreach (DocumentRank ClusteredDoc in cluster)
//// {
//// foreach (Article article in Articles)
//// {
//// if (ClusteredDoc.DocumentID == article.Id)
//// {
//// Results.Add(article);
//// }
//// }
//// //return Results;
//// }
//// }
//// }
// return Results;
//}
public IEnumerable <TermData> GetTermFrequency()
{
List <TermData> termlist = new List <TermData>();
if (System.IO.Directory.Exists(indexPath))
{
luceneIndexDirectory = FSDirectory.Open(indexPath);
// writer = new IndexWriter(luceneIndexDirectory, analyzer, true, IndexWriter.MaxFieldLength.UNLIMITED);
//IndexReader reader = writer.GetReader();
IndexReader reader = IndexReader.Open(luceneIndexDirectory, true);
TermEnum terms = reader.Terms();
while (terms.Next())
{
TermData td = new TermData();
Term term = terms.Term;
td.TermText = term.Text;
td.TermFrequency = reader.DocFreq(term);
termlist.Add(td);
}
reader.Dispose();
return(termlist);
}
else
{
throw new NullReferenceException("Index does not exist.");
}
}
/// <summary>Increments the enumeration to the next element. True if one exists. </summary>
public override bool Next()
{
if (actualEnum == null)
{
return(false); // the actual enumerator is not initialized!
}
currentTerm = null;
while (currentTerm == null)
{
if (EndEnum())
{
return(false);
}
if (actualEnum.Next())
{
Term term = actualEnum.Term;
if (TermCompare(term))
{
currentTerm = term;
return(true);
}
}
else
{
return(false);
}
}
currentTerm = null;
return(false);
}
public void AddWithAnalyzerSuccess()
{
TestObject t = new TestObject()
{
Number = 9876,
String = "Test Object 9876",
};
Assert.AreEqual(0, writer.NumDocs());
writer.Add(t, new KeywordAnalyzer());
writer.Commit();
Assert.AreEqual(1, writer.NumDocs());
IndexReader reader = IndexReader.Open(dir, true);
TermEnum terms = reader.Terms();
int nTerms = 0;
while (terms.Next())
{
if (String.Equals("String", terms.Term.Field))
{
Assert.AreEqual("Test Object 9876", terms.Term.Text);
nTerms++;
}
}
Assert.AreEqual(1, nTerms);
}
internal void InitOverview()
{
// populate analyzers
searchTabPage.Init();
string indexName = indexPath + (_readOnly ? " (R)" : "");
statusBar.Panels[0].Text = resources.GetString("StatusIndexName") + indexName;
List <string> fieldNames = new List <string>(indexReader.GetFieldNames(IndexReader.FieldOption.ALL));
SetFieldNames(fieldNames);
searchTabPage.SetFieldnames(fieldNames);
overviewTabPage.Init(indexName);
TermEnum termsEnum = indexReader.Terms(); //TODO: Duplicated
int i = 0;
while (termsEnum.Next())
{
i++;
}
termsEnum.Close();
documentsTabPage.Init(i, indexReader.NumDocs());
}
/// <summary>
/// 查找指定数目的Term
/// </summary>
/// <param name="num"></param>
/// <returns></returns>
public TermModel[] FindTerms(int num)
{
num++;
TermInfoQueue queue = new TermInfoQueue(num);
TermEnum enum2 = open.Reader.Terms();
int count = 0;
while (enum2.Next())
{
string str = enum2.Term().Field();
if ((currentField != null) && (!str.Equals(currentField)))
{
continue;
}
if (enum2.DocFreq() > count)
{
queue.Put(new TermModel(enum2.Term(), enum2.DocFreq()));
if (queue.Size() < num)
{
continue;
}
queue.Pop();
count = ((TermModel)queue.Top()).Count;
}
}
enum2.Close();
TermModel[] modleArray = new TermModel[queue.Size()];
for (int i = 0; i < modleArray.Length; i++)
{
modleArray[(modleArray.Length - i) - 1] = (TermModel)queue.Pop();
}
return(modleArray);
}
protected internal override System.Object CreateValue(IndexReader reader, Entry entryKey)
{
System.String field = StringHelper.Intern(entryKey.field);
System.String[] retArray = new System.String[reader.MaxDoc];
TermDocs termDocs = reader.TermDocs();
TermEnum termEnum = reader.Terms(new Term(field));
try
{
do
{
Term term = termEnum.Term;
if (term == null || (System.Object)term.Field != (System.Object)field)
{
break;
}
System.String termval = term.Text;
termDocs.Seek(termEnum);
while (termDocs.Next())
{
retArray[termDocs.Doc] = termval;
}
}while (termEnum.Next());
}
finally
{
termDocs.Close();
termEnum.Close();
}
return(retArray);
}
public void AddSuccess()
{
TestObject t = new TestObject()
{
Number = 1234,
String = "Test Object 1234",
};
Assert.AreEqual(0, writer.NumDocs());
writer.Add(t);
writer.Commit();
Assert.AreEqual(1, writer.NumDocs());
IndexReader reader = IndexReader.Open(dir, true);
TermEnum terms = reader.Terms();
HashSet <string> expectedTerms = new HashSet <string>(new string[] { "test", "object", "1234" });
while (terms.Next())
{
if (String.Equals("String", terms.Term.Field))
{
Assert.True(expectedTerms.Contains(terms.Term.Text));
expectedTerms.Remove(terms.Term.Text);
}
}
Assert.AreEqual(0, expectedTerms.Count);
}
public IEnumerable <TermInfo> GetTerms()
{
var directory = _openIndexModel.Directory;
IndexReader indexReader = null;
TermEnum terms = null;
try
{
indexReader = IndexReader.Open(directory, true); // ToDo should i open this only once
terms = indexReader.Terms();
while (terms.Next())
{
System.Threading.Thread.Sleep(2);
var term = terms.Term();
yield return(new TermInfo {
Term = term.Text(), Field = term.Field(), Frequency = terms.DocFreq()
});
}
}
finally
{
if (indexReader != null)
{
indexReader.Close();
}
if (terms != null)
{
terms.Close();
}
}
yield break;
}
public virtual void TestTerms()
{
TermEnum terms = _reader.Terms();
Assert.IsTrue(terms != null);
while (terms.Next() == true)
{
Term term = terms.Term;
Assert.IsTrue(term != null);
//System.out.println("Term: " + term);
System.String fieldValue = (System.String)DocHelper.NameValues[term.Field];
Assert.IsTrue(fieldValue.IndexOf(term.Text) != -1);
}
TermDocs termDocs = _reader.TermDocs();
Assert.IsTrue(termDocs != null);
termDocs.Seek(new Term(DocHelper.TextField1Key, "field"));
Assert.IsTrue(termDocs.Next() == true);
termDocs.Seek(new Term(DocHelper.NoNormsKey, DocHelper.NoNormsText));
Assert.IsTrue(termDocs.Next() == true);
TermPositions positions = _reader.TermPositions();
positions.Seek(new Term(DocHelper.TextField1Key, "field"));
Assert.IsTrue(positions != null);
Assert.IsTrue(positions.Doc == 0);
Assert.IsTrue(positions.NextPosition() >= 0);
}
protected internal override System.Object CreateValue(IndexReader reader, Entry entryKey)
{
Entry entry = entryKey;
System.String field = entry.field;
FloatParser parser = (FloatParser)entry.custom;
if (parser == null)
{
try
{
return(wrapper.GetFloats(reader, field, Lucene.Net.Search.FieldCache_Fields.DEFAULT_FLOAT_PARSER));
}
catch (System.FormatException)
{
return(wrapper.GetFloats(reader, field, Lucene.Net.Search.FieldCache_Fields.NUMERIC_UTILS_FLOAT_PARSER));
}
}
float[] retArray = null;
TermDocs termDocs = reader.TermDocs();
TermEnum termEnum = reader.Terms(new Term(field));
try
{
do
{
Term term = termEnum.Term;
if (term == null || (System.Object)term.Field != (System.Object)field)
{
break;
}
float termval = parser.ParseFloat(term.Text);
if (retArray == null)
{
// late init
retArray = new float[reader.MaxDoc];
}
termDocs.Seek(termEnum);
while (termDocs.Next())
{
retArray[termDocs.Doc] = termval;
}
}while (termEnum.Next());
}
catch (StopFillCacheException)
{
}
finally
{
termDocs.Close();
termEnum.Close();
}
if (retArray == null)
{
// no values
retArray = new float[reader.MaxDoc];
}
return(retArray);
}
public override DocIdSet GetDocIdSet(IndexReader reader)
{
TermEnum enumerator = query.GetEnum(reader);
try
{
// if current term in enum is null, the enum is empty -> shortcut
if (enumerator.Term == null)
{
return(DocIdSet.EMPTY_DOCIDSET);
}
// else fill into an OpenBitSet
OpenBitSet bitSet = new OpenBitSet(reader.MaxDoc);
int[] docs = new int[32];
int[] freqs = new int[32];
TermDocs termDocs = reader.TermDocs();
try
{
int termCount = 0;
do
{
Term term = enumerator.Term;
if (term == null)
{
break;
}
termCount++;
termDocs.Seek(term);
while (true)
{
int count = termDocs.Read(docs, freqs);
if (count != 0)
{
for (int i = 0; i < count; i++)
{
bitSet.Set(docs[i]);
}
}
else
{
break;
}
}
} while (enumerator.Next());
query.IncTotalNumberOfTerms(termCount); // {{Aroush-2.9}} is the use of 'temp' as is right?
}
finally
{
termDocs.Close();
}
return(bitSet);
}
finally
{
enumerator.Close();
}
}
public void UpdateWithKindSuccess()
{
TestObject t = new TestObject()
{
Number = 1234,
String = "Test Object 1234",
};
Assert.AreEqual(0, writer.NumDocs());
writer.Add <object>(t);
writer.Commit();
Assert.AreEqual(1, writer.NumDocs());
TestObject t2 = new TestObject()
{
Number = 2345,
String = "Something Else 2345",
};
writer.Update(t2, DocumentObjectTypeKind.Static, new TermQuery(new Term("String", "1234")));
writer.Commit();
Assert.AreEqual(2, writer.NumDocs());
writer.DeleteDocuments <object>(new MatchAllDocsQuery());
writer.Commit();
Assert.AreEqual(1, writer.NumDocs());
TestObject t3 = new TestObject()
{
Number = 3456,
String = "Completely Different 3456",
};
writer.Update(t3, DocumentObjectTypeKind.Actual, new TermQuery(new Term("String", "2345")));
writer.Commit();
Assert.AreEqual(1, writer.NumDocs());
IndexReader reader = IndexReader.Open(dir, true);
TermEnum terms = reader.Terms();
HashSet <string> expectedTerms = new HashSet <string>(new string[] { "completely", "different", "3456" });
while (terms.Next())
{
if (String.Equals("String", terms.Term.Field))
{
if (expectedTerms.Contains(terms.Term.Text))
{
expectedTerms.Remove(terms.Term.Text);
}
}
}
Assert.AreEqual(0, expectedTerms.Count);
}
public void UpdateWithKindWithAnalyzerSuccess()
{
TestObject t = new TestObject()
{
Number = 1234,
String = "Test Object 1234",
};
Assert.AreEqual(0, writer.NumDocs());
writer.Add <object>(t, new KeywordAnalyzer());
writer.Commit();
Assert.AreEqual(1, writer.NumDocs());
TestObject t2 = new TestObject()
{
Number = 2345,
String = "Something Else 2345",
};
writer.Update(t2, DocumentObjectTypeKind.Static, new TermQuery(new Term("String", "Test Object 1234")), new KeywordAnalyzer());
writer.Commit();
Assert.AreEqual(2, writer.NumDocs());
writer.DeleteDocuments <object>(new MatchAllDocsQuery());
writer.Commit();
Assert.AreEqual(1, writer.NumDocs());
TestObject t3 = new TestObject()
{
Number = 3456,
String = "Completely Different 3456",
};
writer.Update(t3, DocumentObjectTypeKind.Actual, new TermQuery(new Term("String", "Something Else 2345")), new KeywordAnalyzer());
writer.Commit();
Assert.AreEqual(1, writer.NumDocs());
IndexReader reader = IndexReader.Open(dir, true);
TermEnum terms = reader.Terms();
int nTerms = 0;
while (terms.Next())
{
if (String.Equals("String", terms.Term.Field))
{
if (String.Equals("Completely Different 3456", terms.Term.Text))
{
nTerms++;
}
}
}
Assert.AreEqual(1, nTerms);
}
public void DisplayInternalIndex()
{
Directory mainIndexDir = SearchFactory.GetSearchFactory(sessions).GetDirectoryProvider(typeof(Book)).Directory;
IndexReader reader = IndexReader.Open(mainIndexDir);
TermEnum terms = reader.Terms();
while (terms.Next())
{
Term term = terms.Term();
log.Debug("In " + term.Field() + ": " + term.Text());
}
}
public Term Next()
{
//in Lucene 3, a call to reader.terms(term) is already pre-positioned, you don't call next first
if (initialState)
{
initialState = false;
return(Term());
}
else
{
return(termEnum.Next() ? Term() : null);
}
}
internal bool Next()
{
if (termEnum.Next())
{
term = termEnum.Term();
return(true);
}
else
{
term = null;
return(false);
}
}
static void CheckIsHeadOnly(IndexSearcher searcher)
{
TermEnum t = searcher.Reader.Terms(new Term(FieldName.RevisionLast, "0"));
Assert.IsNotNull(t);
Assert.AreEqual(FieldName.RevisionLast, t.Term().Field());
while (t.Term().Field() == FieldName.RevisionLast)
{
Assert.AreEqual(Revision.HeadString, t.Term().Text());
if (t.Next())
{
continue;
}
}
}
/// <summary>
/// Converts TermEnum to an enumerable of Term instances.
/// </summary>
/// <param name="termEnum">TermEnum instance which should be converted to an IEnumerable</param>
/// <returns>Enumerable of Term objects.</returns>
public static IEnumerable <Term> ToEnumerable(this TermEnum termEnum)
{
if (termEnum != null)
{
using ( termEnum )
{
do
{
if (termEnum.Term != null)
{
yield return(termEnum.Term);
}
}while(termEnum.Next());
}
}
}
public HashSet <String> GetListOfItemsNameFromIndex()
{
IndexReader reader = IndexReader.Open(DirectoryFs, true);
TermEnum terms = reader.Terms();
HashSet <String> uniqueTerms = new HashSet <String>();
while (terms.Next())
{
Term term = terms.Term;
if (term.Field.Equals("LineText"))
{
uniqueTerms.Add(term.Text);
}
}
return(uniqueTerms);
}
protected int GetDictValueCount(IndexReader reader, string field)
{
int ret = 0;
using (TermEnum termEnum = reader.Terms(new Term(field, "")))
{
do
{
Term term = termEnum.Term;
if (term == null || string.CompareOrdinal(term.Field, field) != 0)
{
break;
}
ret++;
} while (termEnum.Next());
}
return(ret);
}
/// <summary>
/// 查找指定字段Term
/// </summary>
/// <param name="field"></param>
/// <param name="text"></param>
/// <param name="current"></param>
/// <returns></returns>
public TermModel FindTerm(string field, string text, bool current)
{
TermEnum enum2 = open.Reader.Terms();
if (enum2.SkipTo(new Term(field, text)))
{
TermModel modle2 = null;
while ((!current && enum2.Next() && field.Equals(enum2.Term().Field())) ||
current)
{
modle2 = new TermModel(enum2.Term(), enum2.DocFreq());
break;
}
enum2.Close();
return(modle2);
}
return(null);
}
/// <summary>
/// Converts TermEnum to an enumerable of Term instances. Stops enumeration when the Term does not belong to the specified field.
/// </summary>
/// <param name="termEnum">TermEnum instance which should be converted to an IEnumerable</param>
/// <param name="field">Field which should be enumerated.</param>
/// <returns>Enumerable of Term objects.</returns>
public static IEnumerable <Term> ToEnumerable(this TermEnum termEnum, string field)
{
if (termEnum != null)
{
do
{
Term term = termEnum.Term;
if (term != null)
{
if (term.Field != field)
{
break;
}
yield return(term);
}
}while(termEnum.Next());
}
}
private OpenBitSet FastBits(IndexReader reader)
{
OpenBitSet bits = new OpenBitSet(reader.MaxDoc);
bits.Set(0, reader.MaxDoc); //assume all are valid
Term startTerm = new Term(fieldName);
TermEnum te = reader.Terms(startTerm);
if (te != null)
{
Term currTerm = te.Term;
while ((currTerm != null) && (currTerm.Field == startTerm.Field)) //term fieldnames are interned
{
if (te.DocFreq() > 1)
{
int lastDoc = -1;
//unset potential duplicates
TermDocs td = reader.TermDocs(currTerm);
td.Next();
if (keepMode == KM_USE_FIRST_OCCURRENCE)
{
td.Next();
}
do
{
lastDoc = td.Doc;
bits.Clear(lastDoc);
} while (td.Next());
if (keepMode == KM_USE_LAST_OCCURRENCE)
{
//restore the last bit
bits.Set(lastDoc);
}
}
if (!te.Next())
{
break;
}
currTerm = te.Term;
}
}
return(bits);
}
private void buttonFirstTerm_Click(object sender, System.EventArgs e)
{
if (_luke.IndexReader == null)
{
_luke.ShowStatus(_luke.resources.GetString("NoIndex"));
return;
}
try
{
TermEnum te = _luke.IndexReader.Terms();
te.Next();
Term t = te.Term();
_ShowTerm(t);
}
catch (Exception exc)
{
_luke.ShowStatus(exc.Message);
}
}
public IEnumerable <string> Terms(string field)
{
if (!index.Storage.Exists)
{
yield break;
}
IndexReader reader = index.Storage.OpenReader();
TermEnum terms = reader.Terms(new Term(field));
do
{
if (terms.Term.Field != field)
{
yield break;
}
yield return(terms.Term.Text);
} while (terms.Next());
}
protected internal override System.Object CreateValue(IndexReader reader, Entry entryKey)
{
Entry entry = entryKey;
System.String field = entry.field;
ShortParser parser = (ShortParser)entry.custom;
if (parser == null)
{
return(wrapper.GetShorts(reader, field, Lucene.Net.Search.FieldCache_Fields.DEFAULT_SHORT_PARSER));
}
short[] retArray = new short[reader.MaxDoc];
TermDocs termDocs = reader.TermDocs();
TermEnum termEnum = reader.Terms(new Term(field));
try
{
do
{
Term term = termEnum.Term;
if (term == null || (System.Object)term.Field != (System.Object)field)
{
break;
}
short termval = parser.ParseShort(term.Text);
termDocs.Seek(termEnum);
while (termDocs.Next())
{
retArray[termDocs.Doc] = termval;
}
}while (termEnum.Next());
}
catch (StopFillCacheException)
{
}
finally
{
termDocs.Close();
termEnum.Close();
}
return(retArray);
}
internal void Init(string indexName)
{
TermEnum termsEnum = _luke.IndexReader.Terms(); //TODO: Duplicated
int i = 0;
while (termsEnum.Next())
{
i++;
}
termsEnum.Close();
TermsNumber = i;
IndexName = indexName;
IndexVersion = IndexReader.GetCurrentVersion(_luke.Directory).ToString();
HasDeletions = _luke.IndexReader.HasDeletions().ToString();
DocumentsNumber = _luke.IndexReader.NumDocs();
LastModified = IndexReader.LastModified(_luke.Directory);
ShowTopTerms();
}
private IEnumerable <string> GetIndexTerms()
{
List <string> termsList = new List <string>();
using (Directory luceneIndexDirectory = FSDirectory.Open(indexPath))
{
using (IndexReader reader = IndexReader.Open(luceneIndexDirectory, true))
{
TermEnum terms = reader.Terms();
while (terms.Next())
{
Term term = terms.Term;
string termText = term.Text;
termsList.Add(termText);
}
return(termsList);
}
}
}
