---

FlattextDocMgr.hpp

Go to the documentation of this file.

/*==========================================================================
 *
 *  Original source copyright (c) 2001, Carnegie Mellon University.
 *  See copyright.cmu for details.
 *  Modifications copyright (c) 2002, University of Massachusetts.
 *  See copyright.umass for details.
 *
 *==========================================================================
*/

#ifndef _FLATTEXTDOCMGR_HPP
#define _FLATTEXTDOCMGR_HPP

#include "common_headers.hpp"
#include "TrecParser.hpp"
#include "WebParser.hpp"
#include "ReutersParser.hpp"
#include "ChineseParser.hpp"
#include "ChineseCharParser.hpp"
#include "ArabicParser.hpp"
#include "Exception.hpp"
#include "DocumentManager.hpp"

#define FT_SUFFIX ".flat"
#define FT_LOOKUP ".lookup"
#define FT_FID    ".fid"

class FlattextDocMgr : public DocumentManager, public TextHandler {
public:
  enum ParseModes {TREC=0, WEB=1, CHINESE=2, CHINESECHAR=3, ARABIC=4};

  struct lookup_e {
    int fid;
    int offset;
    int bytes;
  };

  struct abc {
    bool operator() (char* s1, char* s2) const {
      return strcmp(s1, s2) < 0;
    }
  };

  FlattextDocMgr(char* name, ParseModes mode, char* source);
  
  FlattextDocMgr(const char* name);

  ~FlattextDocMgr();

  bool open(const char*manname);

  const char* getMyID();

  char* getDoc(const char* docID);

  void buildMgr();

  char* handleDoc(char * docno);

  static Parser* createParser(ParseModes mode);

private:
  bool readinSources(char* fn);

  void writeTOC();

  bool loadTOC(char* fn);
  bool loadFTLookup(const char* fn);
  bool loadFTFiles(const char* fn, int num);

  int numdocs;              // how many docs we have
  ParseModes pm;           // what type of parser we have
  long prevpos;              // pos of previous doc beginning
  string IDname;            // my name
  vector<string> sources;   // list of all source files
  int fileid, lastid;       // fileid of current/last file being processed
  ofstream writefpos;       // stream for writing out file positions
  map<char*, lookup_e, abc> table; 
};

#endif

---