Main Page | Namespace List | Class Hierarchy | Class List | File List | Namespace Members | Class Members | File Members | Related Pages

CtfIndexer.hpp

Go to the documentation of this file.
00001 /*==========================================================================
00002  * Copyright (c) 2000-2004 Carnegie Mellon University.  All Rights Reserved.
00003  *
00004  * Use of the Lemur Toolkit for Language Modeling and Information Retrieval
00005  * is subject to the terms of the software license set forth in the LICENSE
00006  * file included with this software (and below), and also available at
00007  * http://www.lemurproject.org/license.html
00008  *
00009  *==========================================================================
00010 */
00011 
00012 #include "TextHandler.hpp"
00013 #include "PushIndex.hpp"
00014 #include "InvFPTermList.hpp"
00015 #include "Parser.hpp"
00016 #include "WordSet.hpp"
00017 
00018 #include <stdio.h>
00019 
00020 #ifndef _CTFINDEXER_HPP
00021 #define _CTFINDEXER_HPP
00022 
00023 namespace lemur 
00024 {
00025   namespace distrib 
00026   {
00027     
00028     class CtfIndexer : public lemur::api::TextHandler {
00029 
00030     public:
00031       CtfIndexer(const string &csName, int bufferSize, 
00032                  bool countStopWords = false);
00033       ~CtfIndexer();
00034 
00035       char * handleWord(char * word);
00036 
00037       void newDb(const string &name);
00038 
00039     private:
00040       int ctfCount;
00041       bool first;
00042       lemur::index::PushIndex * collsel;
00043       lemur::parse::DocumentProps * csdp;
00044       lemur::index::InvFPTerm * term;
00045       bool countStopWds;
00046     };
00047   }
00048 }
00049 
00050 #endif

Generated on Tue Jun 15 11:02:53 2010 for Lemur by doxygen 1.3.4