source: gsdl/trunk/lib/gdbmclass.cpp@ 15557

/**********************************************************************
 *
 * gdbmclass.cpp -- 
 * Copyright (C) 1999-2008  The New Zealand Digital Library Project
 *
 * A component of the Greenstone digital library software
 * from the New Zealand Digital Library Project at the
 * University of Waikato, New Zealand.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *
 *********************************************************************/

#include "gdbmclass.h"
#include "unitool.h"
#include "gsdlunicode.h"
#include "OIDtools.h"


gdbmclass::~gdbmclass()
{
  closedatabase();
}

// returns true if opened
bool gdbmclass::opendatabase (const text_t &filename, int mode, int num_retrys, 
#ifdef __WIN32__
                  bool need_filelock
#else
                              bool
#endif
                  )
{
  text_t data_location;
  int block_size = 512;
  
  if (gdbmfile != NULL) {
    if (openfile == filename) return true;
    else closedatabase ();
  }

  openfile = filename;

  // Map the DB mode values into GDBM mode values
  int gdbm_mode = GDBM_READER;
  if (mode == DB_WRITER)
  {
    gdbm_mode = GDBM_WRITER;
  }
  else if (mode == DB_WRITER_CREATE)
  {
    gdbm_mode = GDBM_WRCREAT;
  }

  char *namebuffer = filename.getcstr();
  do {
#ifdef __WIN32__
    gdbmfile = gdbm_open (namebuffer, block_size, gdbm_mode, 00664, NULL, (need_filelock) ? 1 : 0);
#else
    gdbmfile = gdbm_open (namebuffer, block_size, gdbm_mode, 00664, NULL);
#endif
    --num_retrys;
  } while (num_retrys>0 && gdbmfile==NULL &&
       (gdbm_errno==GDBM_CANT_BE_READER || gdbm_errno==GDBM_CANT_BE_WRITER));
  delete []namebuffer;
  
  if (gdbmfile == NULL && logout != NULL) {
    outconvertclass text_t2ascii;
    (*logout) << text_t2ascii << "database open failed on: " << filename << "\n";
  }

  return (gdbmfile != NULL);
}


void gdbmclass::closedatabase ()
{
  if (gdbmfile == NULL) return;
  
  gdbm_close (gdbmfile);
  gdbmfile = NULL;
  openfile.clear();
}


// returns true on success
bool gdbmclass::setinfo (const text_t &key, const infodbclass &info)
{
  if (gdbmfile == NULL) return false;

  text_t subkey;
  text_t data;

  // get all the keys and values
  infodbclass::const_iterator info_here = info.begin();
  infodbclass::const_iterator info_end = info.end();
  while (info_here != info_end) {
    // add the key
    subkey.clear();
    subkey.push_back('<');
    text_t::const_iterator subkey_here = (*info_here).first.begin();
    text_t::const_iterator subkey_end = (*info_here).first.end();
    while (subkey_here != subkey_end) {
      if (*subkey_here == '>') {
    subkey.push_back('\\'); subkey.push_back('>');
      } else if (*subkey_here == '\n') {
    subkey.push_back('\\'); subkey.push_back('n');
      } else if (*subkey_here == '\r') {
    subkey.push_back('\\'); subkey.push_back('r');
      } else if (*subkey_here == '\\') {
    subkey.push_back('\\'); subkey.push_back('\\');
      } else {
    subkey.push_back (*subkey_here);
      }
      ++subkey_here;
    }
    subkey.push_back('>');

    // add the values
    text_tarray::const_iterator subvalue_here = (*info_here).second.begin();
    text_tarray::const_iterator subvalue_end = (*info_here).second.end();
    while (subvalue_here != subvalue_end) {
      data += subkey;
      
      text_t::const_iterator thissubvalue_here = (*subvalue_here).begin();
      text_t::const_iterator thissubvalue_end = (*subvalue_here).end();
      while (thissubvalue_here != thissubvalue_end) {
    if (*thissubvalue_here == '>') {
      data.push_back('\\'); data.push_back('>');
    } else if (*thissubvalue_here == '\n') {
      data.push_back('\\'); data.push_back('n');
    } else if (*thissubvalue_here == '\r') {
      data.push_back('\\'); data.push_back('r');
    } else if (*thissubvalue_here == '\\') {
      data.push_back('\\'); data.push_back('\\');
    } else {
      data.push_back (*thissubvalue_here);
    }
    
    ++thissubvalue_here;
      }
      
      data.push_back('\n');
      ++subvalue_here;
    }

    ++info_here;
  }

  // store the value
  datum key_data;
  datum data_data;

  // get a utf-8 encoded c string of the unicode key
  key_data.dptr = (to_utf8(key)).getcstr();
  if (key_data.dptr == NULL) {
    if (logout != NULL) (*logout) << "gdbmclass: out of memory\n";
    return false;
  }
  key_data.dsize = strlen (key_data.dptr);

  data_data.dptr = (to_utf8(data)).getcstr();
  if (data_data.dptr == NULL) {
    if (logout != NULL) (*logout) << "gdbmclass: out of memory\n";
    delete []key_data.dptr;
    return false;
  }
  data_data.dsize = strlen (data_data.dptr);

  int ret = gdbm_store (gdbmfile, key_data, data_data, GDBM_REPLACE);
  delete []key_data.dptr;
  delete []data_data.dptr;

  return (ret == 0);
}


//returns true on success
bool gdbmclass::setinfo (const text_t &key, const text_t &data)
{
  if (gdbmfile == NULL) return false;
  
  // store the value
  datum key_data;
  datum data_data;

  // get a utf-8 encoded c string of the unicode key
  key_data.dptr = (to_utf8(key)).getcstr();
  if (key_data.dptr == NULL) {
    if (logout != NULL) (*logout) << "gdbmclass: out of memory\n";
    return false;
  }
  key_data.dsize = strlen (key_data.dptr);

  data_data.dptr = (to_utf8(data)).getcstr();
  if (data_data.dptr == NULL) {
    if (logout != NULL) (*logout) << "gdbmclass: out of memory\n";
    delete []key_data.dptr;
    return false;
  }
  data_data.dsize = strlen (data_data.dptr);

  int ret = gdbm_store (gdbmfile, key_data, data_data, GDBM_REPLACE);
  delete []key_data.dptr;
  delete []data_data.dptr;

  return (ret == 0);
}


void gdbmclass::deletekey (const text_t &key)
{
  if (gdbmfile == NULL) return;

  // get a utf-8 encoded c string of the unicode key
  datum key_data;
  key_data.dptr = (to_utf8(key)).getcstr();
  if (key_data.dptr == NULL) return;
  key_data.dsize = strlen (key_data.dptr);

  // delete the key
  gdbm_delete (gdbmfile, key_data);

  // free up the key memory
  delete []key_data.dptr;
}


// getfirstkey and getnextkey are used for traversing the database
// no insertions or deletions should be carried out while traversing
// the database. when there are no keys left to visit in the database
// an empty string is returned.
text_t gdbmclass::getfirstkey ()
{
  if (gdbmfile == NULL) return g_EmptyText;

  // get the first key
  datum firstkey_data = gdbm_firstkey (gdbmfile);
  if (firstkey_data.dptr == NULL) return g_EmptyText;

  // convert it to text_t
  text_t firstkey;
  firstkey.setcarr (firstkey_data.dptr, firstkey_data.dsize);
  free (firstkey_data.dptr);
  return to_uni(firstkey);  // convert to unicode
}


text_t gdbmclass::getnextkey (const text_t &key)
{
  if (gdbmfile == NULL || key.empty()) return g_EmptyText;

  // get a utf-8 encoded c string of the unicode key
  datum key_data;
  key_data.dptr = (to_utf8(key)).getcstr();
  if (key_data.dptr == NULL) return g_EmptyText;
  key_data.dsize = strlen (key_data.dptr);
  
  // get the next key
  datum nextkey_data = gdbm_nextkey (gdbmfile, key_data);
  if (nextkey_data.dptr == NULL) {
    delete []key_data.dptr;
    return g_EmptyText;
  }

  // convert it to text_t
  text_t nextkey;
  nextkey.setcarr (nextkey_data.dptr, nextkey_data.dsize);
  free (nextkey_data.dptr);
  delete []key_data.dptr;
  return to_uni(nextkey);  // convert to unicode
}


// replaces the .fc, .lc, .pr, .rt, .ns and .ps syntax (first child, 
// last child, parent, root, next sibling, previous sibling) 
// it expects child, parent, etc. to exist if syntax has been used
// so you should test before using
text_t gdbmclass::translate_OID (const text_t &inOID, infodbclass &info)
{
  if (inOID.size() < 4) return inOID;
  if (findchar (inOID.begin(), inOID.end(), '.') == inOID.end()) return inOID;

  text_t OID = inOID;
  text_tarray tailarray;
  text_t tail = substr (OID.end()-3, OID.end());
  if (tail == ".rt") {
    get_top (inOID, OID);
    return OID;
  }
  while (tail == ".fc" || tail == ".lc" || tail == ".pr" || 
     tail == ".ns" || tail == ".ps") {
    tailarray.push_back(tail);
    OID.erase (OID.end()-3, OID.end());
    tail = substr (OID.end()-3, OID.end());
    if (tail == ".rt") {
      get_top (inOID, OID);
      return OID;
    }
  }

  if (tailarray.empty()) return inOID;
  text_tarray::const_iterator begin = tailarray.begin();
  text_tarray::const_iterator here = tailarray.end() - 1;

  while (here >= begin) {

    if (*here == ".fc")
      get_first_child (OID, info);
    else if (*here == ".lc")
      get_last_child (OID, info);
    else if (*here == ".pr")
      OID = get_parent (OID);
    else if (*here == ".ns")
      get_next_sibling (OID, info);
    else if (*here == ".ps")
      get_previous_sibling (OID, info);
    
    if (here == begin)
      break;
    --here;
  }

  return OID;
}


void gdbmclass::get_first_child (text_t &OID, infodbclass &info)
{
  text_t firstchild;
  if (getinfo (OID, info)) {
    text_t &contains = info["contains"];
    if (!contains.empty()) {
      text_t parent = OID;
      getdelimitstr (contains.begin(), contains.end(), ';', firstchild);
      if (firstchild.empty()) OID = contains;
      else OID = firstchild;
      if (*(OID.begin()) == '"') translate_parent (OID, parent);
    }
  }
}


void gdbmclass::get_last_child (text_t &OID, infodbclass &info)
{
  text_tarray children;
  if (getinfo (OID, info)) {
    text_t &contains = info["contains"];
    if (!contains.empty()) {
      text_t parent = OID;
      splitchar (contains.begin(), contains.end(), ';', children);
      OID = children.back();
      if (*(OID.begin()) == '"') translate_parent (OID, parent);
    }
  }
}

  
void gdbmclass::get_next_sibling (text_t &OID, infodbclass &info)
{
  text_tarray siblings;
  text_t parent = get_parent (OID);
         
  if (getinfo (parent, info)) {
    text_t &contains = info["contains"];
    if (!contains.empty()) {
      splitchar (contains.begin(), contains.end(), ';', siblings);
      text_tarray::const_iterator here = siblings.begin();
      text_tarray::const_iterator end = siblings.end();
      text_t shrunk_OID = OID;
      shrink_parent (shrunk_OID);
      while (here != end) {
    if (*here == shrunk_OID && (here+1 != end)) {
      OID = *(here+1);
      if (*(OID.begin()) == '"') translate_parent (OID, parent);
      break;
    }
    ++here;
      }
    }
  }
}


void gdbmclass::get_previous_sibling (text_t &OID, infodbclass &info)
{
  text_tarray siblings;
  text_t parent = get_parent (OID);

  if (getinfo (parent, info)) {
    text_t &contains = info["contains"];
    if (!contains.empty()) {
      splitchar (contains.begin(), contains.end(), ';', siblings);
      text_tarray::const_iterator here = siblings.begin();
      text_tarray::const_iterator end = siblings.end();
      text_t shrunk_OID = OID;
      shrink_parent (shrunk_OID);
      while (here != end) {
    if (*here == shrunk_OID && (here != siblings.begin())) {
      OID = *(here-1);
      if (*(OID.begin()) == '"') translate_parent (OID, parent);
      break;
    }
    ++here;
      }
    }
  }
}


// returns true on success
bool gdbmclass::getinfo (const text_t& key, infodbclass &info)
{
  text_t data;
  
  if (!getkeydata (key, data)) return false;
  text_t::iterator here = data.begin ();
  text_t::iterator end = data.end ();
  
  text_t ikey, ivalue;
  info.clear (); // reset info
  
  while (getinfoline (here, end, ikey, ivalue)) {
    info.addinfo (ikey, ivalue);
  }
  
  return true;
}


// returns true if exists
bool gdbmclass::exists (const text_t& key)
{
  text_t data;
  return getkeydata (key, data);
}


// returns true on success
bool gdbmclass::getkeydata (const text_t& key, text_t &data)
{
  datum key_data;
  datum return_data;

  if (gdbmfile == NULL) return false;
  
  // get a utf-8 encoded c string of the unicode key
  key_data.dptr = (to_utf8(key)).getcstr();
  if (key_data.dptr == NULL) {
    if (logout != NULL) (*logout) << "gdbmclass: out of memory\n";
    return false;
  }
  key_data.dsize = strlen (key_data.dptr);
  
  // fetch the result
  return_data = gdbm_fetch (gdbmfile, key_data);
  delete []key_data.dptr;
  
  if (return_data.dptr == NULL) return false;

  data.setcarr (return_data.dptr, return_data.dsize);
  free (return_data.dptr);
  data = to_uni(data);  // convert to unicode

  return true;
}


// returns true on success
bool gdbmclass::getinfoline (text_t::iterator &here, text_t::iterator end, 
                 text_t &key, text_t &value)
{
  key.clear();
  value.clear();

  // ignore white space
  while (here != end && is_unicode_space (*here)) ++here;

  // get the '<'
  if (here == end || *here != '<') return false;
  ++here;
  
  // get the key
  while (here != end && *here != '>') {
    key.push_back(*here);
    ++here;
  }
  
  // get the '>'
  if (here == end || *here != '>') return false;
  ++here;
  
  // get the value
  while (here != end && *here != '\n') {
    if (*here == '\\') { 
      // found escape character
      ++here;
      if (here != end) {
    if (*here == 'n') value.push_back ('\n');
    else if (*here == 'r') value.push_back ('\r');
    else value.push_back(*here);
      }

    } else {
      // a normal character
      value.push_back(*here);
    }

    ++here;
  }

  return true;
}