tdevelop/lib/interfaces/hashedstring.cpp

/***************************************************************************
   copyright            : (C) 2006 by David Nolden
   email                : david.nolden.tdevelop@art-master.de
***************************************************************************/

/***************************************************************************
 *                                                                         *
 *   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.                                   *
 *                                                                         *
 ***************************************************************************/

#include "hashedstring.h"
#include <kdatastream.h>
#include <sstream>
#include <algorithm>
#include <iterator>
#include<ext/hash_set>
#include<set>
#include<algorithm>

//It needs to be measured whether this flag should be turned on or off. It seems just to move the complexity from one position to the other, without any variant being really better.
#define USE_HASHMAP
    
size_t fastHashString( const TQString& str );

size_t hashStringSafe( const TQString& str ) {
  size_t hash = 0;
  int len = str.length();
  for( int a = 0; a < len; a++ ) {
    hash = str[a].tqunicode() + (hash * 17);
  }
  return hash;
}

size_t HashedString::hashString(  const TQString& str )
{
	return fastHashString( str );
}

size_t fastHashString( const TQString& str ) {
  size_t hash = 0;
  if( !str.isEmpty() ) {
    const TQChar* curr = str.tqunicode();
    const TQChar* end = curr + str.length();
    TQChar c;
    for(; curr < end ;) {
      c = *curr;
      hash = c.tqunicode() + ( hash * 17 );
      ++curr;
    }
  }
  return hash;
}

void HashedString::initHash() {
    m_hash = hashString( m_str );
}


class HashedStringSetData : public KShared {
  public:
#ifdef USE_HASHMAP
      typedef __gnu_cxx::hash_set<HashedString> StringSet;
#else
      typedef std::set<HashedString> StringSet; //must be a set, so the set-algorithms work
#endif
      StringSet m_files;
  mutable bool m_hashValid;
      mutable size_t m_hash;
      HashedStringSetData() : m_hashValid( false ) {
      }
  inline void invalidateHash() {
    m_hashValid = false;
  }

  void computeHash() const;
};

void HashedStringSetData::computeHash() const {
  int num = 1;
  m_hash = 0;
  for( StringSet::const_iterator it = m_files.begin(); it != m_files.end(); ++it ) {
    num *= 7;
    m_hash += num * (*it).hash();
  }
  m_hashValid = true;
}

HashedStringSet::HashedStringSet() {}

HashedStringSet::~HashedStringSet() {}

HashedStringSet::HashedStringSet( const HashedString& file ) {
  insert( file );
}

HashedStringSet::HashedStringSet( const HashedStringSet& rhs ) : m_data( rhs.m_data ) {}

HashedStringSet operator + ( const HashedStringSet& lhs, const HashedStringSet& rhs ) {
  HashedStringSet ret = lhs;
  ret += rhs;

  return ret;
}

int HashedStringSet::size() const {
    if( !m_data ) return 0;
    return m_data->m_files.size();
}

HashedStringSet& HashedStringSet::operator = ( const HashedStringSet& rhs ) {
  m_data = rhs.m_data;
  return *this;
}

HashedStringSet& HashedStringSet::operator +=( const HashedStringSet& rhs ) {
  if ( !rhs.m_data )
    return * this;

#ifndef USE_HASHMAP
  KSharedPtr<HashedStringSetData> oldData = m_data;
  if( !oldData ) oldData = new HashedStringSetData();
  m_data = new HashedStringSetData();
  std::set_union( oldData->m_files.begin(), oldData->m_files.end(), rhs.m_data->m_files.begin(), rhs.m_data->m_files.end(), std::insert_iterator<HashedStringSetData::StringSet>( m_data->m_files, m_data->m_files.end() ) );
#else
  makeDataPrivate();
  m_data->m_files.insert( rhs.m_data->m_files.begin(), rhs.m_data->m_files.end() );
  /*HashedStringSetData::StringSet::const_iterator end = rhs.m_data->m_files.end();
  HashedStringSetData::StringSet& mySet( m_data->m_files );
  for( HashedStringSetData::StringSet::const_iterator it = rhs.m_data->m_files.begin(); it != end; ++it ) {
    mySet.insert( *it );
  }*/
  
#endif
  return *this;
}

HashedStringSet& HashedStringSet::operator -=( const HashedStringSet& rhs ) {
  if( !m_data ) return *this;
  if( !rhs.m_data ) return *this;
#ifndef USE_HASHMAP
  KSharedPtr<HashedStringSetData> oldData = m_data;
  m_data = new HashedStringSetData();
  std::set_difference( oldData->m_files.begin(), oldData->m_files.end(), rhs.m_data->m_files.begin(), rhs.m_data->m_files.end(), std::insert_iterator<HashedStringSetData::StringSet>( m_data->m_files, m_data->m_files.end() ) );
#else
  makeDataPrivate();
  HashedStringSetData::StringSet::const_iterator end = rhs.m_data->m_files.end();
  HashedStringSetData::StringSet::const_iterator myEnd = m_data->m_files.end();
  HashedStringSetData::StringSet& mySet( m_data->m_files );
  for( HashedStringSetData::StringSet::const_iterator it = rhs.m_data->m_files.begin(); it != end; ++it ) {
    mySet.erase( *it );
  }
  
#endif
  return *this;
}


void HashedStringSet::makeDataPrivate() {
  if ( !m_data ) {
    m_data = new HashedStringSetData();
    return ;
  }
  if ( m_data->_KShared_count() != 1 )
    m_data = new HashedStringSetData( *m_data );
}

bool HashedStringSet::operator[] ( const HashedString& rhs ) const {
  //if ( rhs.str() == "*" )
    //return true; /// * stands for "any file"
  if ( !m_data )
    return false;
  return m_data->m_files.find( rhs ) != m_data->m_files.end();
}

void HashedStringSet::insert( const HashedString& str ) {
  if( str.str().isEmpty() ) return;
  makeDataPrivate();
  m_data->m_files.insert( str );
  m_data->invalidateHash();
}

bool HashedStringSet::operator <= ( const HashedStringSet& rhs ) const {
  if ( !m_data )
    return true;
  if ( m_data->m_files.empty() )
    return true;
  if ( !rhs.m_data )
    return false;
#ifndef USE_HASHMAP
  return std::includes( rhs.m_data->m_files.begin(), rhs.m_data->m_files.end(), m_data->m_files.begin(), m_data->m_files.end() );
#else
  const HashedStringSetData::StringSet& otherSet( rhs.m_data->m_files );
  HashedStringSetData::StringSet::const_iterator end = rhs.m_data->m_files.end();
  HashedStringSetData::StringSet::const_iterator myEnd = m_data->m_files.end();

  for( HashedStringSetData::StringSet::const_iterator it = m_data->m_files.begin(); it != myEnd; ++it ) {
    HashedStringSetData::StringSet::const_iterator i = otherSet.find( *it );
    if( i == end ) return false;
  }
  return true;
#endif
}

bool HashedStringSet::operator == ( const HashedStringSet& rhs ) const {
  if( hash() != rhs.hash() ) return false;
  
  bool empty1 = false;
  if ( !m_data )
    empty1 = true;
  else if ( m_data->m_files.empty() )
    empty1 = true;
  bool empty2 = false;
  if ( !rhs.m_data )
    empty2 = true;
  else if ( rhs.m_data->m_files.empty() )
    empty2 = true;

  if ( empty1 && empty2 )
    return true;
  if ( empty1 || empty2 )
    return false;

  return m_data->m_files == rhs.m_data->m_files;
}

size_t HashedStringSet::hash() const {
  if( !m_data ) return 0;
  if( !m_data->m_hashValid ) m_data->computeHash();
  return m_data->m_hash;
}

void HashedStringSet::read( TQDataStream& stream ) {
    bool b;
    stream >> b;
    if( b ) {
        m_data = new HashedStringSetData();
        int cnt;
        stream >> cnt;
        HashedString s;
        for( int a = 0; a < cnt; a++ ) {
            stream >> s;
            m_data->m_files.insert( s );
        }
    } else {
        m_data = 0;
    }
}

void HashedStringSet::write( TQDataStream& stream ) const {
    bool b = m_data;
    stream << b;
    if( b ) {
        int cnt = m_data->m_files.size();
        stream << cnt;
        for( HashedStringSetData::StringSet::const_iterator it = m_data->m_files.begin(); it != m_data->m_files.end(); ++it ) {
            stream << *it;
        }
    }
}

std::string HashedStringSet::print() const {
  std::ostringstream s;
  if( m_data ) {
        for( HashedStringSetData::StringSet::const_iterator it = m_data->m_files.begin(); it != m_data->m_files.end(); ++it ) {
	  s << (*it).str().ascii() << "\n";
	}
  }
  return s.str();
}

TQDataStream& operator << ( TQDataStream& stream, const HashedString& str ) {
    stream << str.m_str;
    stream << str.m_hash;
    return stream;
}

TQDataStream& operator >> ( TQDataStream& stream, HashedString& str ) {
    stream >> str.m_str;
    stream >> str.m_hash;
    return stream;
}

void HashedStringSetGroup::addSet( size_t id, const HashedStringSet& set ) {
  if( set.m_data && !set.m_data->m_files.empty() ) {
    m_sizeMap[ id ] = set.size();
    for( HashedStringSetData::StringSet::const_iterator it = set.m_data->m_files.begin(); it != set.m_data->m_files.end(); ++it ) {
      GroupMap::iterator itr = m_map.find( *it );
      if( itr == m_map.end() ) {
        itr = m_map.insert( std::make_pair( *it, ItemSet() ) ).first;
      }
      itr->second.insert( id );
    }
  } else {
    m_global.insert( id );
  }
}

void HashedStringSetGroup::disableSet( size_t id ) {
  m_disabled.insert( id );
}

void HashedStringSetGroup::enableSet( size_t id ) {
  m_disabled.erase( id );
}

bool HashedStringSetGroup::isDisabled( size_t id ) const {
  return m_disabled.find( id ) != m_disabled.end();
}

void HashedStringSetGroup::removeSet( size_t id ) {
  m_disabled.erase( id );
  m_global.erase( id );
  m_sizeMap.erase( id );
  for( GroupMap::iterator it = m_map.begin(); it != m_map.end(); ++it ) {
    it->second.erase( id );
  }
}

void HashedStringSetGroup::findGroups( HashedStringSet strings, ItemSet& target ) const {
  target.clear();
  if( !strings.m_data ) {
    std::set_difference( m_global.begin(), m_global.end(), m_disabled.begin(), m_disabled.end(), std::insert_iterator<ItemSet>( target, target.end() ) );
    return;
  }
  //This might yet be optimized by sorting the sets according to their size, and starting the intersectioning with the smallest ones.
  __gnu_cxx::hash_map<size_t, int> hitCounts;
  
  for( HashedStringSetData::StringSet::const_iterator it = strings.m_data->m_files.begin(); it != strings.m_data->m_files.end(); ++it ) {
    GroupMap::const_iterator itr = m_map.find( *it );
      if( itr == m_map.end() ) {
        //There are no string-sets that include the currently searched for string
        continue;
      }

      for( ItemSet::const_iterator it2 = itr->second.begin(); it2 != itr->second.end(); ++it2 ) {
        __gnu_cxx::hash_map<size_t, int>::iterator v = hitCounts.find( *it2 );
        if( v != hitCounts.end() ) {
          ++(*v).second;
        } else {
          hitCounts[*it2] = 1;
        }
      }
  }

  //Now count together all groups that are completely within the given string-set(their hitCount equals their size)
  ItemSet found;
  for( __gnu_cxx::hash_map<size_t, int>::const_iterator it = hitCounts.begin(); it != hitCounts.end(); ++it ) {
    if( (*it).second == (*m_sizeMap.find( (*it).first )).second )
      found.insert( (*it).first );
  }

  
  std::set_union( found.begin(), found.end(), m_global.begin(), m_global.end(), std::insert_iterator<ItemSet>( target, target.end() ) );
  
  target.swap( found );
  target.clear();
  std::set_difference( found.begin(), found.end(), m_disabled.begin(), m_disabled.end(), std::insert_iterator<ItemSet>( target, target.end() ) );
}