---

KeyfileDocMgr.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 _LEMUR_KEYFILE_DOCMGR_HPP
#define _LEMUR_KEYFILE_DOCMGR_HPP

#include "common_headers.hpp"
#include "DocumentManager.hpp"
#include "RVLCompress.hpp"
#include "TextHandlerManager.hpp"
#include "Match.hpp"
#include "Keyfile.hpp"

// array of byte offsets, indexed by token for each doc.
#define BT_POSITIONS ".btp"
// source file start, length
#define BT_LOOKUP ".btl"
// TOC
#define BT_TOC ".bdm"
// source files.
#define BT_FID ".bfi"


class KeyfileDocMgr : public DocumentManager, public TextHandler {
public:
  KeyfileDocMgr() {  myDoc = NULL;  numdocs = 0; }

  KeyfileDocMgr(const char *name);

  KeyfileDocMgr(string name, string mode, string source);  

  virtual ~KeyfileDocMgr();

  char* getDoc(const char* docID);
  virtual char* handleDoc(char * docno);
  virtual void handleEndDoc();
  virtual char *handleWord(char * word) {
    if (word != NULL) {
      int end = myparser->fileTell() - 1;
      int start = (end - strlen(word)) + 1;
      Match m;
      m.start = start - docEntry.offset;
      m.end = end - docEntry.offset;    
      offsets.push_back(m);
    }
    return word;
  }
  virtual void setParser(Parser *p) {
    myparser = p;
  }

  virtual void buildMgr();
  virtual const char *getMyID() {
    return IDnameext.c_str();
  }

  vector<Match> getOffsets(char *docID);

  virtual bool open(const char* manname) {
    string tmp(manname);
    IDname = tmp.substr(0, tmp.length() - 4);
    return loadTOC();
  }

protected:
  struct btl {
    int fid;
    long offset;
    long bytes;
  };

  virtual void writeTOC();
  virtual bool loadTOC();
  bool loadFTFiles(const char* fn, int num);
  vector <Match> offsets;
  int numdocs;              // how many docs we have
  string pm;  // parse mode

  Keyfile poslookup; // btree for lookup to positions list.
  Keyfile doclookup; // btree for lookup to doc start.
  int dbcache;
  
  btl docEntry;
  char *myDoc;
  int doclen;
  string IDname;            // my name
  string IDnameext;                     // my name w/ extension
  vector<string> sources;   // list of all source files
  int fileid;       // fileid of current/last file being processed
};

#endif // _LEMUR_KEYFILE_DOCMGR_HPP

---