Пример #1
0
		private int[] docMap; // use getDocMap()
		
		internal SegmentMergeInfo(int b, TermEnum te, IndexReader r)
		{
			base_Renamed = b;
			reader = r;
			termEnum = te;
			term = te.Term();
		}
Пример #2
0
		public virtual void  Seek(TermEnum termEnum)
		{
			TermInfo ti;
			Term term;
			
			// use comparison of fieldinfos to verify that termEnum belongs to the same segment as this SegmentTermDocs
			if (termEnum is SegmentTermEnum && ((SegmentTermEnum) termEnum).fieldInfos == parent.core.fieldInfos)
			{
				// optimized case
				SegmentTermEnum segmentTermEnum = ((SegmentTermEnum) termEnum);
				term = segmentTermEnum.Term();
				ti = segmentTermEnum.TermInfo();
			}
			else
			{
				// punt case
				term = termEnum.Term();
				ti = parent.core.GetTermsReader().Get(term);
			}
			
			Seek(ti, term);
		}
Пример #3
0
			public override bool Next()
			{
				if (termEnum == null)
					return false;
				
				// another term in this field?
				if (termEnum.Next() && (System.Object) termEnum.Term().Field() == (System.Object) field)
					return true; // yes, keep going
				
				termEnum.Close(); // close old termEnum
				
				// find the next field with terms, if any
				if (fieldIterator == null)
				{
                    System.Collections.Comparer comparer = System.Collections.Comparer.Default;
                    System.Collections.SortedList newList = new System.Collections.SortedList();
                    if (Enclosing_Instance.fieldToReader != null)
                    {
                        if (Enclosing_Instance.fieldToReader.Count > 0)
                        {
                            int index = 0;
                            while (comparer.Compare(Enclosing_Instance.fieldToReader.GetKey(index), field) < 0)
                                index++;
                            for (; index < Enclosing_Instance.fieldToReader.Count; index++)
                            {
                                newList.Add(Enclosing_Instance.fieldToReader.GetKey(index), Enclosing_Instance.fieldToReader[Enclosing_Instance.fieldToReader.GetKey(index)]);
                            }
                        }
                    }

                    fieldIterator = newList.Keys.GetEnumerator();
                    fieldIterator.MoveNext();
					System.Object generatedAux = fieldIterator.Current; // Skip field to get next one
				}
				while (fieldIterator.MoveNext())
				{
					field = ((System.String) fieldIterator.Current);
					termEnum = ((IndexReader) Enclosing_Instance.fieldToReader[field]).Terms(new Term(field));
					Term term = termEnum.Term();
					if (term != null && (System.Object) term.Field() == (System.Object) field)
						return true;
					else
						termEnum.Close();
				}
				
				return false; // no more fields
			}
Пример #4
0
			public ParallelTermEnum(ParallelReader enclosingInstance, Term term)
			{
				InitBlock(enclosingInstance);
				field = term.Field();
				IndexReader reader = ((IndexReader) Enclosing_Instance.fieldToReader[field]);
				if (reader != null)
					termEnum = reader.Terms(term);
			}
Пример #5
0
			public ParallelTermEnum(ParallelReader enclosingInstance)
			{
				InitBlock(enclosingInstance);
				try
				{
					field = ((System.String) Enclosing_Instance.fieldToReader.GetKey(0));
				}
				catch (ArgumentOutOfRangeException e)
				{
					// No fields, so keep field == null, termEnum == null
					return;
				}
				if (field != null)
					termEnum = ((IndexReader) Enclosing_Instance.fieldToReader[field]).Terms();
			}
Пример #6
0
		/// <summary> Not implemented.</summary>
		/// <throws>  UnsupportedOperationException </throws>
		public virtual void  Seek(TermEnum termEnum)
		{
			throw new System.NotSupportedException();
		}
Пример #7
0
			public virtual void  Seek(TermEnum termEnum)
			{
				Seek(termEnum.Term());
				if (termEnum is MultiTermEnum)
				{
					tenum = (MultiTermEnum) termEnum;
					if (topReader != tenum.topReader)
						tenum = null;
				}
			}
Пример #8
0
			public virtual void  Seek(TermEnum termEnum)
			{
				in_Renamed.Seek(termEnum);
			}
Пример #9
0
        private void  MergeTermInfos(FormatPostingsFieldsConsumer consumer)
        {
            int base_Renamed = 0;
            int readerCount  = readers.Count;

            for (int i = 0; i < readerCount; i++)
            {
                IndexReader      reader   = (IndexReader)readers[i];
                TermEnum         termEnum = reader.Terms();
                SegmentMergeInfo smi      = new SegmentMergeInfo(base_Renamed, termEnum, reader);
                int[]            docMap   = smi.GetDocMap();
                if (docMap != null)
                {
                    if (docMaps == null)
                    {
                        docMaps   = new int[readerCount][];
                        delCounts = new int[readerCount];
                    }
                    docMaps[i]   = docMap;
                    delCounts[i] = smi.reader.MaxDoc() - smi.reader.NumDocs();
                }

                base_Renamed += reader.NumDocs();

                System.Diagnostics.Debug.Assert(reader.NumDocs() == reader.MaxDoc() - smi.delCount);

                if (smi.Next())
                {
                    queue.Add(smi);
                }
                // initialize queue
                else
                {
                    smi.Close();
                }
            }

            SegmentMergeInfo[] match = new SegmentMergeInfo[readers.Count];

            System.String currentField = null;
            FormatPostingsTermsConsumer termsConsumer = null;

            while (queue.Size() > 0)
            {
                int matchSize = 0;                 // pop matching terms
                match[matchSize++] = (SegmentMergeInfo)queue.Pop();
                Term             term = match[0].term;
                SegmentMergeInfo top  = (SegmentMergeInfo)queue.Top();

                while (top != null && term.CompareTo(top.term) == 0)
                {
                    match[matchSize++] = (SegmentMergeInfo)queue.Pop();
                    top = (SegmentMergeInfo)queue.Top();
                }

                if ((System.Object)currentField != (System.Object)term.field)
                {
                    currentField = term.field;
                    if (termsConsumer != null)
                    {
                        termsConsumer.Finish();
                    }
                    FieldInfo fieldInfo = fieldInfos.FieldInfo(currentField);
                    termsConsumer            = consumer.AddField(fieldInfo);
                    omitTermFreqAndPositions = fieldInfo.omitTermFreqAndPositions;
                }

                int df = AppendPostings(termsConsumer, match, matchSize);                 // add new TermInfo

                checkAbort.Work(df / 3.0);

                while (matchSize > 0)
                {
                    SegmentMergeInfo smi = match[--matchSize];
                    if (smi.Next())
                    {
                        queue.Add(smi);
                    }
                    // restore queue
                    else
                    {
                        smi.Close();                         // done with a segment
                    }
                }
            }
        }
Пример #10
0
 public void Seek(TermEnum termEnum)
 {
     throw new NotSupportedException();
 }
Пример #11
0
        /// <summary> Test the term index.</summary>
        private Status.TermIndexStatus TestTermIndex(SegmentInfo info, SegmentReader reader)
        {
            Status.TermIndexStatus status = new Status.TermIndexStatus();

            try
            {
                if (infoStream != null)
                {
                    infoStream.Write("    test: terms, freq, prox...");
                }

                TermEnum      termEnum      = reader.Terms();
                TermPositions termPositions = reader.TermPositions();

                // Used only to count up # deleted docs for this term
                MySegmentTermDocs myTermDocs = new MySegmentTermDocs(reader);

                int maxDoc = reader.MaxDoc();

                while (termEnum.Next())
                {
                    status.termCount++;
                    Term term    = termEnum.Term();
                    int  docFreq = termEnum.DocFreq();
                    termPositions.Seek(term);
                    int lastDoc = -1;
                    int freq0   = 0;
                    status.totFreq += docFreq;
                    while (termPositions.Next())
                    {
                        freq0++;
                        int doc  = termPositions.Doc();
                        int freq = termPositions.Freq();
                        if (doc <= lastDoc)
                        {
                            throw new System.SystemException("term " + term + ": doc " + doc + " <= lastDoc " + lastDoc);
                        }
                        if (doc >= maxDoc)
                        {
                            throw new System.SystemException("term " + term + ": doc " + doc + " >= maxDoc " + maxDoc);
                        }

                        lastDoc = doc;
                        if (freq <= 0)
                        {
                            throw new System.SystemException("term " + term + ": doc " + doc + ": freq " + freq + " is out of bounds");
                        }

                        int lastPos = -1;
                        status.totPos += freq;
                        for (int j = 0; j < freq; j++)
                        {
                            int pos = termPositions.NextPosition();
                            if (pos < -1)
                            {
                                throw new System.SystemException("term " + term + ": doc " + doc + ": pos " + pos + " is out of bounds");
                            }
                            if (pos < lastPos)
                            {
                                throw new System.SystemException("term " + term + ": doc " + doc + ": pos " + pos + " < lastPos " + lastPos);
                            }
                        }
                    }

                    // Now count how many deleted docs occurred in
                    // this term:
                    int delCount;
                    if (reader.HasDeletions())
                    {
                        myTermDocs.Seek(term);
                        while (myTermDocs.Next())
                        {
                        }
                        delCount = myTermDocs.delCount;
                    }
                    else
                    {
                        delCount = 0;
                    }

                    if (freq0 + delCount != docFreq)
                    {
                        throw new System.SystemException("term " + term + " docFreq=" + docFreq + " != num docs seen " + freq0 + " + num docs deleted " + delCount);
                    }
                }

                Msg("OK [" + status.termCount + " terms; " + status.totFreq + " terms/docs pairs; " + status.totPos + " tokens]");
            }
            catch (System.Exception e)
            {
                Msg("ERROR [" + System.Convert.ToString(e.Message) + "]");
                status.error = e;
                if (infoStream != null)
                {
                    infoStream.WriteLine(e.StackTrace);
                }
            }

            return(status);
        }
Пример #12
0
 public virtual void  Seek(TermEnum termEnum)
 {
     in_Renamed.Seek(termEnum);
 }
Пример #13
0
 public FilterTermEnum(TermEnum in_Renamed)
 {
     this.in_Renamed = in_Renamed;
 }
Пример #14
0
 public void Seek(TermEnum termEnum)
 {
     throw new NotSupportedException();
 }
Пример #15
0
			public virtual void  Seek(TermEnum termEnum)
			{
				Seek(termEnum.Term());
			}
Пример #16
0
			public FilterTermEnum(TermEnum in_Renamed)
			{
				this.in_Renamed = in_Renamed;
			}
Пример #17
0
 public virtual void  Seek(TermEnum termEnum)
 {
     Seek(termEnum.Term());
 }