static double
score2(const boost::unordered_map<int, double> &m0,
const boost::unordered_map<int, double> &m1)
{
int tmp = 0;
auto i0 = m0.begin();
auto i1 = m1.begin();
while (i0 != m0.end() && i1 != m1.end()) {
if (i0->first < i1->first) ++i0;
else if (i0->first > i1->first) ++i1;
else ++tmp, ++i0, ++i1;
}
return m0.size() + m1.size() - tmp;
}
void IfcTreeWidget::slotObjectsSelected( boost::unordered_map<int, shared_ptr<IfcPPEntity> >& map )
{
if( m_block_selection_signals )
{
return;
}
if( map.size() < 1 )
{
return;
}
// take the first object from map and highlight it
shared_ptr<IfcPPEntity> object = (*(map.begin())).second;
int selected_id = object->m_id;
for( int i=0; i<topLevelItemCount(); ++i )
{
QTreeWidgetItem* toplevel_item = topLevelItem( i );
QTreeWidgetItem* selected_item = ViewerUtil::findItemByIfcId( toplevel_item, selected_id );
if( selected_item != 0 )
{
blockSignals(true);
m_block_selection_signals = true;
setCurrentItem( selected_item, 1, QItemSelectionModel::SelectCurrent );
blockSignals(false);
m_block_selection_signals = false;
break;
}
}
}
void IndexingVariables::set(
boost::unordered_map<std::string, boost::shared_ptr<ArrayStorage> > as,
int offset) {
hm.clear();
hm.insert(as.begin(), as.end());
this->offset = offset;
}
//! Dump Legendre polynomial cache data to stream (table and history).
void dumpLegendrePolynomialCacheData( std::ostream& outputStream,
boost::unordered_map< Point, double > cacheTable,
boost::circular_buffer< Point > cacheHistory )
{
outputStream << "Table:\n";
for ( boost::unordered_map< Point, double >::iterator iteratorCacheTable = cacheTable.begin( );
iteratorCacheTable != cacheTable.end( ); iteratorCacheTable++ )
{
outputStream << "\t" << writeLegendrePolynomialStructureToString(
iteratorCacheTable->first ).c_str( ) << " => "
<< iteratorCacheTable->second << std::endl;
}
outputStream << "History:\n";
for ( boost::circular_buffer< Point >::iterator iteratorCacheHistory = cacheHistory.begin( );
iteratorCacheHistory != cacheHistory.end( ); iteratorCacheHistory++ )
{
outputStream << "\t"
<< writeLegendrePolynomialStructureToString( *iteratorCacheHistory ).c_str( )
<< ", ";
}
outputStream << std::endl;
}
TKmerMapSimple
kmerify_graph_simple(String<TVertexDescriptor const> const & order,
TGraph const & graph,
std::vector<VertexLabels> & vertex_vector,
boost::unordered_set<TVertexDescriptor> const & free_nodes,
boost::unordered_map< std::pair<TVertexDescriptor, TVertexDescriptor>, boost::dynamic_bitset<> > & edge_ids,
int const & kmer_size
)
{
TKmerMapSimple kmer_map;
for (Iterator<String<TVertexDescriptor const> const>::Type it = begin(order) ; it != end(order) ; ++it)
{
TVertexDescriptor const & source_vertex = *it;
if (free_nodes.count(source_vertex) == 0)
{
boost::dynamic_bitset<> id_bits(edge_ids.begin()->second.size());
id_bits.flip();
check_kmers_simple(vertex_vector[source_vertex].dna, graph, source_vertex, vertex_vector, free_nodes, edge_ids, id_bits, kmer_map, static_cast<std::size_t>(kmer_size));
}
}
return kmer_map;
}
TKmerMap
kmerifyGraph(String<TVertexDescriptor const> const & order,
TGraph const & graph,
std::vector<VertexLabels> & vertex_vector,
boost::unordered_set<TVertexDescriptor> const & free_nodes,
boost::unordered_map< std::pair<TVertexDescriptor, TVertexDescriptor>, boost::dynamic_bitset<> > & edge_ids,
int const & kmer_size
)
{
TKmerMap kmer_map;
for (Iterator<String<TVertexDescriptor const> const>::Type it = begin(order) ; it != end(order) ; ++it)
{
TVertexDescriptor const & source_vertex = *it;
if (free_nodes.count(source_vertex) == 0)
{
boost::dynamic_bitset<> id_bits(edge_ids.begin()->second.size());
id_bits.flip();
checkKmers(vertex_vector[source_vertex].dna, source_vertex, source_vertex, graph, vertex_vector, free_nodes, edge_ids, id_bits, kmer_map, static_cast<std::size_t>(kmer_size));
// std::cout << "source_vertex = " << source_vertex << " kmer_map.size() = " << kmer_map.size() << " kmer_size = " << kmer_size;
// std::cout << " vertex_vector[source_vertex].level = " << vertex_vector[source_vertex].level << std::endl;
}
}
return kmer_map;
}
void Streamer::processPickups(Player &player, const std::vector<SharedCell> &cells)
{
static boost::unordered_map<int, Item::SharedPickup> discoveredPickups;
for (std::vector<SharedCell>::const_iterator c = cells.begin(); c != cells.end(); ++c)
{
for (boost::unordered_map<int, Item::SharedPickup>::const_iterator p = (*c)->pickups.begin(); p != (*c)->pickups.end(); ++p)
{
boost::unordered_map<int, Item::SharedPickup>::iterator d = discoveredPickups.find(p->first);
if (d == discoveredPickups.end())
{
if (checkPlayer(p->second->players, player.playerID, p->second->interiors, player.interiorID, p->second->worlds, player.worldID))
{
if (boost::geometry::comparable_distance(player.position, p->second->position) <= p->second->streamDistance)
{
boost::unordered_map<int, int>::iterator i = internalPickups.find(p->first);
if (i == internalPickups.end())
{
p->second->worldID = !p->second->worlds.empty() ? player.worldID : -1;
}
discoveredPickups.insert(*p);
}
}
}
}
}
if (processingFinalPlayer)
{
boost::unordered_map<int, int>::iterator i = internalPickups.begin();
while (i != internalPickups.end())
{
boost::unordered_map<int, Item::SharedPickup>::iterator d = discoveredPickups.find(i->first);
if (d == discoveredPickups.end())
{
DestroyPickup(i->second);
i = internalPickups.erase(i);
}
else
{
discoveredPickups.erase(d);
++i;
}
}
for (boost::unordered_map<int, Item::SharedPickup>::iterator d = discoveredPickups.begin(); d != discoveredPickups.end(); ++d)
{
if (internalPickups.size() == visiblePickups)
{
break;
}
int internalID = CreatePickup(d->second->modelID, d->second->type, d->second->position[0], d->second->position[1], d->second->position[2], d->second->worldID);
if (internalID == INVALID_ALTERNATE_ID)
{
break;
}
internalPickups.insert(std::make_pair(d->second->pickupID, internalID));
}
discoveredPickups.clear();
}
}
///Calculates the Gini impurity of a node. The impurity is defined as
///1-sum_j p(j|t)^2
///i.e the 1 minus the sum of the squared probability of observing class j in node t
double CARTTrainer::gini(boost::unordered_map<std::size_t, std::size_t>& countMatrix, std::size_t n){
double res = 0;
boost::unordered_map<std::size_t, std::size_t>::iterator it;
double denominator = n;
for ( it=countMatrix.begin() ; it != countMatrix.end(); it++ ){
res += sqr(it->second/denominator);
}
return 1-res;
}
bool buildNameLookupTable(Kompex::SQLiteStatement * stmt,
boost::unordered_map<std::string,int32_t> &table_names)
{
std::string stmt_insert = "INSERT INTO name_lookup";
stmt_insert += "(name_id,name_lookup) VALUES(@name_id,@name_lookup);";
try {
stmt->BeginTransaction();
stmt->Sql(stmt_insert);
}
catch(Kompex::SQLiteException &exception) {
qDebug() << "ERROR: SQLite exception with insert statement:"
<< QString::fromStdString(exception.GetString());
return false;
}
// keep track of the number of transactions and
// commit after a certain limit
size_t transaction_limit=5000;
size_t transaction_count=0;
// write name lookup info the database
boost::unordered_map<std::string,int32_t>::iterator it;
for(it = table_names.begin(); it != table_names.end(); ++it) {
// prepare sql
try {
//[name_id, name_lookup]
stmt->BindInt(1,it->second);
stmt->BindString(2,it->first);
stmt->Execute();
stmt->Reset();
if(transaction_count > transaction_limit) {
stmt->FreeQuery();
stmt->CommitTransaction();
stmt->BeginTransaction();
stmt->Sql(stmt_insert);
transaction_count=0;
}
transaction_count++;
}
catch(Kompex::SQLiteException &exception) {
qDebug() << "ERROR: SQLite exception writing tile data:"
<< QString::fromStdString(exception.GetString());
qDebug() << "ERROR: name_id:" << it->second;
qDebug() << "ERROR: name_lookup:" << QString::fromStdString(it->first);
return false;
}
}
stmt->FreeQuery();
stmt->CommitTransaction();
return true;
}
void read_db_users()
{
m_read_db_users_time = srv_time();
if (!m_use_sql)
return;
try
{
Csql_query q(m_database, "select @uid");
if (m_read_users_can_leech)
q += ", can_leech";
if (m_read_users_peers_limit)
q += ", peers_limit";
if (m_read_users_torrent_pass)
q += ", torrent_pass";
q += ", torrent_pass_version";
if (m_read_users_wait_time)
q += ", wait_time";
q += " from @users";
Csql_result result = q.execute();
// m_users.reserve(result.size());
for (auto& i : m_users)
i.second.marked = true;
m_users_torrent_passes.clear();
while (Csql_row row = result.fetch_row())
{
t_user& user = m_users[row[0].i()];
user.marked = false;
int c = 0;
user.uid = row[c++].i();
if (m_read_users_can_leech)
user.can_leech = row[c++].i();
if (m_read_users_peers_limit)
user.peers_limit = row[c++].i();
if (m_read_users_torrent_pass)
{
if (row[c].size() == 32)
m_users_torrent_passes[to_array<char, 32>(row[c])] = &user;
c++;
}
user.torrent_pass_version = row[c++].i();
if (m_read_users_wait_time)
user.wait_time = row[c++].i();
}
for (auto i = m_users.begin(); i != m_users.end(); )
{
if (i->second.marked)
m_users.erase(i++);
else
i++;
}
}
catch (bad_query&)
{
}
}
static void serialize(storage_type& s, const boost::unordered_map<T1, T2>& o)
{
uint32_t sz = o.size();
s.serialize(sz);
for (typename boost::unordered_map<T1,T2>::const_iterator it = o.begin(), it_end = o.end(); it != it_end; ++it)
{
s.serialize(it->first);
s.serialize(it->second);
}
}
///Calculates the Gini impurity of a node. The impurity is defined as
///1-sum_j p(j|t)^2
///i.e the 1 minus the sum of the squared probability of observing class j in node t
double RFTrainer::gini(boost::unordered_map<std::size_t, std::size_t> & countMatrix, std::size_t n){
double res = 0;
boost::unordered_map<std::size_t, std::size_t>::iterator it;
if(n){
n = n*n;
for ( it=countMatrix.begin() ; it != countMatrix.end(); it++ ){
res += sqr(it->second)/(double)n;
}
}
return 1-res;
}
int TwoSum::find_range(int lower, int upper) {
int count = 0;
numbers_to_check.resize(upper-lower);
for (int i=lower; i<= upper; i++ ) {
numbers_to_check.push_back(i);
};
for( auto it = numbers.begin(); it != numbers.end(); ++it ) {
if( find_sum(it->first) ) {
++count;
};
};
return count;
};
RealVector CARTTrainer::hist(boost::unordered_map<std::size_t, std::size_t> countMatrix){
//create a normed histogram
std::size_t totalElements = 0;
RealVector normedHistogram(m_maxLabel+1);
normedHistogram.clear();
boost::unordered_map<std::size_t, std::size_t>::iterator it;
for ( it=countMatrix.begin() ; it != countMatrix.end(); it++ ){
normedHistogram(it->first) = it->second;
totalElements += it->second;
}
normedHistogram /= totalElements;
return normedHistogram;
}
void Task::drawKeypoints(const base::samples::frame::Frame &frame2,
const std::vector<cv::KeyPoint> &keypoints1,
const std::vector<cv::KeyPoint> &keypoints2,
const std::vector<cv::DMatch> &matches,
const boost::unordered_map<boost::uuids::uuid, StereoFeature> &hash)
{
/** Convert Images to opencv **/
cv::Mat img2 = frameHelperLeft.convertToCvMat(frame2);
::cv::Mat img_out(img2);
cv::drawKeypoints (img2, keypoints1, img_out, cv::Scalar(0, 255, 0));//green previous
for (std::vector<cv::DMatch>::const_iterator it= matches.begin(); it!= matches.end(); ++it)
{
cv::Point point1 = keypoints1[it->queryIdx].pt;
cv::Point point2 = keypoints2[it->trainIdx].pt;
//cv::circle(img_out, point1, 3, cv::Scalar(0, 255, 0), 1);// green previous
cv::circle(img_out, point2, 3, cv::Scalar(255, 0, 0), 1);// blue current
cv::line (img_out, point1, point2, cv::Scalar(0, 0, 255)); // red line
}
if (_draw_hash_uuid_features.value())
{
/** Draw hash features **/
for (boost::unordered_map<boost::uuids::uuid, StereoFeature>::const_iterator
it = hash.begin(); it != hash.end(); ++it)
{
/** Only current features but with the hash id and the image frame index in the label **/
if (it->second.img_idx == this->fcurrent_left.img_idx)
{
float red = 255.0 - (255.0/(this->frame_window_hash_size)*(this->ffinal_left.img_idx-it->second.img_idx));
cv::circle(img_out, it->second.keypoint_left.pt, 5, cv::Scalar(0, 0, red), 2);
std::vector<std::string> uuid_tokens = this->split(boost::lexical_cast<std::string>(it->first), '-');
cv::putText(img_out, uuid_tokens[uuid_tokens.size()-1]+"["+boost::lexical_cast<std::string>(it->second.img_idx)+"]",
it->second.keypoint_left.pt, cv::FONT_HERSHEY_COMPLEX_SMALL, 1.0, cv::Scalar(0,0,0), 2.5);
}
}
}
::base::samples::frame::Frame *frame_ptr = this->inter_frame_out.write_access();
frameHelperLeft.copyMatToFrame(img_out, *frame_ptr);
frame_ptr->time = this->frame_pair.time;
this->inter_frame_out.reset(frame_ptr);
_inter_frame_samples_out.write(this->inter_frame_out);
return;
}
void save(Archive & ar, const boost::unordered_map<T,U> & t, unsigned int version)
{
// write the size
// TODO: should we handle bucket size as well?
typedef typename boost::unordered_map<T, U>::size_type size_type;
typedef typename boost::unordered_map<T,U>::const_iterator const_iterator;
size_type size = t.size();
ar & BOOST_SERIALIZATION_NVP(size);
unsigned int count = 0;
for (const_iterator it = t.begin(); it != t.end(); it++) {
ar & boost::serialization::make_nvp("pair" + count, *it);
count++;
}
}
RealVector RFTrainer::hist(boost::unordered_map<std::size_t, std::size_t> countMatrix){
RealVector histogram(m_maxLabel+1,0.0);
std::size_t totalElements = 0;
boost::unordered_map<std::size_t, std::size_t>::iterator it;
for ( it=countMatrix.begin() ; it != countMatrix.end(); it++ ){
histogram(it->first) = (double)it->second;
totalElements += it->second;
}
histogram /= totalElements;
return histogram;
}
bool TwoSum::find_sum(long int target) {
long int bucket;
for( auto it = numbers_to_check.begin(); it != numbers_to_check.end(); it++ ) {
if( numbers.find((*it) - target) == numbers.end() ) { continue; };
bucket = numbers.bucket( (*it) - target );
for(auto it_local = numbers.begin(bucket); it_local != numbers.end(bucket); ++it_local) {
if ( ( it_local->first + target ) != (*it) ) { continue; };
if ( it_local->first != target || (it_local->first == target && multi_numbers.find(target) != multi_numbers.end()) ) {
std::cout << target << " + " << it_local->first << " = " << (*it) << " ( " << target + it_local->first << std::endl;
numbers_to_check.erase(it);
return true;
};
};
};
return false;
};
void printIdMap(boost::unordered_map<std::string, long> &my_id_map)
{
// std::cout << my_id_map.size() << std::endl;
std::string max_id;
long max_count = 0;
for (boost::unordered_map<std::string, long>::iterator it = my_id_map.begin() ; it != my_id_map.end() ; ++it)
{
printf("%8s: %4ld\n", it->first.c_str(), it->second);
if (it->second > max_count)
{
max_count = it->second;
max_id = it->first.c_str();
}
}
std::cout << max_id << " wins!" << std::endl;
}
inline void operator<< (object::with_zone& o, const boost::unordered_map<K,V>& v)
{
o.type = type::MAP;
if(v.empty()) {
o.via.map.ptr = NULL;
o.via.map.size = 0;
} else {
object_kv* p = (object_kv*)o.zone->malloc(sizeof(object_kv)*v.size());
object_kv* const pend = p + v.size();
o.via.map.ptr = p;
o.via.map.size = v.size();
typename boost::unordered_map<K,V>::const_iterator it(v.begin());
do {
p->key = object(it->first, o.zone);
p->val = object(it->second, o.zone);
++p;
++it;
} while(p < pend);
}
}
void dump_call_profile()
{
#ifdef TORRENT_PROFILE_CALLS
FILE* out = fopen("blocking_calls.txt", "w+");
std::map<int, std::string> profile;
mutex::scoped_lock l(g_calls_mutex);
for (boost::unordered_map<std::string, int>::const_iterator i = g_blocking_calls.begin()
, end(g_blocking_calls.end()); i != end; ++i)
{
profile[i->second] = i->first;
}
for (std::map<int, std::string>::const_reverse_iterator i = profile.rbegin()
, end(profile.rend()); i != end; ++i)
{
fprintf(out, "\n\n%d\n%s\n", i->first, i->second.c_str());
}
fclose(out);
#endif
}
static std::pair<double, double>
score(const boost::unordered_map<int, double> &real,
const boost::unordered_map<int, double> &fake,
double threshold = 0.0)
{
double res = 0.0;
int cnt = 0;
double e = std::numeric_limits<double>::epsilon();
for (auto i = fake.begin(); i != fake.end(); ++i)
if (i->second >= threshold &&
real.find(i->first) != real.end()) {
++cnt;
double p = real.at(i->first);
double q = fake.at(i->first);
if (p <= e)
res += (1-p) * log((1-p)/(1-q));
else if (1-p <= e || 1-q <= e)
res += p * log(p/q);
else res += p * log(p/q) + (1-p) * log((1-p)/(1-q));
}
return {res, cnt ? res / cnt : 0.0};
}
void Task::cleanHashFeatures (const int current_image_idx,
const int max_number_img,
boost::unordered_map<boost::uuids::uuid, StereoFeature> &hash)
{
if (current_image_idx > max_number_img)
{
const int border_image_idx = (current_image_idx - max_number_img);
/** All the images with idx <= border_image_idx should be eliminated **/
for (boost::unordered_map<boost::uuids::uuid, StereoFeature>::const_iterator
it = hash.begin(); it != hash.end(); ++it)
{
if (it->second.img_idx <= border_image_idx)
{
#ifdef DEBUG_PRINTS
std::cout<<"[CLEAN_HASH_FEATURES] Remove from hash "<<boost::lexical_cast<std::string>(it->first)<<"\n";
#endif
hash.erase(it);
}
}
}
return;
}
void read_db_torrents()
{
m_read_db_torrents_time = srv_time();
if (m_use_sql)
read_db_torrents_sql();
else if (!m_config.m_auto_register)
{
std::set<t_torrent*> new_torrents;
std::ifstream is("xbt_torrents.txt");
for (std::string s; getline(is, s); )
{
s = hex_decode(s);
if (s.size() == 20)
new_torrents.insert(&m_torrents[to_array<char, 20>(s)]);
}
for (auto i = m_torrents.begin(); i != m_torrents.end(); )
{
if (new_torrents.count(&i->second))
i++;
else
m_torrents.erase(i++);
}
}
}
iterator begin() {
return iterator(xmls.begin());
}
void Voxelize::generateUmap(pcl::PointCloud<pcl::PointXYZ> &cloud,double resolution,boost::unordered_map<unordered_map_voxel,un_key> &m)//遍历每个点，生成栅格。然后计算每个栅格的参数
{
double t0 = ros::Time::now().toSec();
un_key tempp; //key键不允许修改，全存在value中
for(int i=0;i<cloud.size();i++)
{
unordered_map_voxel tempv(cloud[i].x,cloud[i].y,cloud[i].z,resolution);
pair<umap::iterator,bool> pair_insert = m.insert(umap::value_type(tempv,tempp));//这个pair就是用来存储插入返回值的，看是否这个栅格已经存在
pair_insert.first->second.vec_index_point.push_back(i);
/*
if(!pair_insert.second)//如果栅格已经存在，那么把value++
{
(pair_insert.first)->second.cout++;
pair_insert.first->second.vec_index_point.push_back(i);
}
else{
pair_insert.first->second.vec_index_point.push_back(i);
}
*/
}
double t2 = ros::Time::now().toSec();
//cout <<"划栅格的时间为"<<(t2-t0)<<endl;
Matrix3d tempmat;
tempmat.setZero();
pcl::PointXYZ temppoint(0,0,0);
for(umap::iterator iter=m.begin();iter!=m.end();)
{
temppoint.x=temppoint.y=temppoint.z=0;
if(iter->second.vec_index_point.size()<10)
{
m.erase(iter++);
continue;
}
for(int j = 0;j<iter->second.vec_index_point.size();j++)
{
double x_ = cloud[iter->second.vec_index_point[j]].x;
double y_ = cloud[iter->second.vec_index_point[j]].y;
double z_ = cloud[iter->second.vec_index_point[j]].z;
tempmat += (Vector3d(x_,y_,z_) * RowVector3d(x_,y_,z_));//存储pipiT的和
temppoint.x += x_;
temppoint.y += y_;
temppoint.z += z_;
}
int n = iter->second.vec_index_point.size();
//cout <<"栅格大小为"<<n<<endl;
Vector3d v(temppoint.x/n,temppoint.y/n,temppoint.z/n);//存储栅格重心
RowVector3d vT(temppoint.x/n,temppoint.y/n,temppoint.z/n);
tempmat /= n;
tempmat -= v*vT;//S
iter->second.matS = tempmat;
//cout <<"S 矩阵= "<< tempmat <<endl;
vector<eigen_sort> vector1(3);//用于排序
//cout <<"tempmat="<<endl<<tempmat<<endl;
EigenSolver<MatrixXd> es(tempmat);
VectorXcd eivals = es.eigenvalues();
MatrixXcd eigenvectors = es.eigenvectors();
//cout << "特征值="<<eivals<<endl;
vector1[0]=eigen_sort(eivals(0).real(),es.eigenvectors().col(0));
vector1[1]=eigen_sort(eivals(1).real(),es.eigenvectors().col(1));
vector1[2]=eigen_sort(eivals(2).real(),es.eigenvectors().col(2));
sort(vector1.begin(),vector1.end());
double lamada1 = vector1[0].eigen_value;
double lamada2 = vector1[1].eigen_value;
double lamada3 = vector1[2].eigen_value;
//cout <<"lamada="<<lamada1<<" "<<lamada2<<" "<<lamada3<<endl;
//
if(vector1[0].eigen_vector(2).real()<0)
{
vector1[0].eigen_vector(2).real() = -vector1[0].eigen_vector(2).real();
}
if(vector1[2].eigen_vector(2).real()<0)
{
vector1[2].eigen_vector(2).real() = -vector1[2].eigen_vector(2).real();
}
iter->second.c = (lamada3-lamada2)/(lamada1+lamada2+lamada3);
iter->second.p = 2*(lamada2-lamada1)/(lamada1+lamada2+lamada3);
iter->second.u = v;
iter->second.v1 = vector1[0].eigen_vector;
iter->second.v3 = vector1[2].eigen_vector;
double c = iter->second.c;
double p = iter->second.p;
iter->second.vector_9D << v(0),v(1),v(2),p*vector1[0].eigen_vector(0).real(),p*vector1[0].eigen_vector(1).real(),
p*vector1[0].eigen_vector(2).real(),c*vector1[2].eigen_vector(0).real(),c*vector1[2].eigen_vector(1).real(),
p*vector1[2].eigen_vector(2).real();
//cout << "9D向量="<<iter->second.vector_9D<<endl;
//ArrayXd ad = iter->second.vector_9D;
//cout <<"ad = "<<ad<<endl;
//cout <<"ad square="<<ad.square()<<endl;
tempmat.setZero();
iter++;
}
double tn = ros::Time::now().toSec();
//cout << "对栅格计算处理的时间"<<(tn - t2)<<endl;
//cout <<"栅格数量为" <<m.size()<<endl;
}
//------------------------------------------------------------------------------
bool MzIDIO::insertMZIDValues(boost::unordered_map<PercolatorOutFeatures, string, PercolatorOutFeatures> &pout_values,
vector <string> &filenames, bool multiplemzidfiles) {
ifstream fpr;
ofstream fpw;
ostream *wout;
string mzidname,s1,psmid;
int i1,vi1,n,xmlindent;
n=0;
psmid="";
try {
for (vi1=0; vi1<filenames.size(); vi1++) {
mzidname=boost::lexical_cast<boost::filesystem::path>(filenames[vi1]).stem().string();
if (!multiplemzidfiles)
mzidname="";
fpr.open(filenames[vi1].c_str());
if (outputfileending.length()>0) {
fpw.open(setOutputFileName(filenames[vi1]).c_str());
wout=&fpw;
}
else
wout=&cout;
while (getline(fpr,s1)) {
if (s1.find(MZID_PARAM::END_INSERT_TAG)!=string::npos) {
if (psmid.length()==0)
THROW_ERROR_VALUE(PRINT_TEXT::BAD_XML,filenames[vi1]);
for (i1=0; i1<ARRAYSIZE(MZID_PARAM::ELEMENT_DATA::ELEMENTS); i1++) {
if (pout_values.find(PercolatorOutFeatures(mzidname,psmid,i1))==pout_values.end())
continue;
n++;
switch (MZID_PARAM::ELEMENT_DATA::ELEMENTS[i1]) {
case MZID_PARAM::CVPARAM: {
*wout << boost::format(MZID_PARAM::CVPARAM_TAG) % string(xmlindent,' ')
% MZID_PARAM::ELEMENT_DATA::ACCESSIONS[i1]
% MZID_PARAM::ELEMENT_DATA::CVREFS[i1]
% MZID_PARAM::ELEMENT_DATA::NAMES[i1]
% pout_values[PercolatorOutFeatures(mzidname,psmid,i1)];
break;
}
case MZID_PARAM::USERPARAM: {
*wout << boost::format(MZID_PARAM::USERPARAM_TAG) % string(xmlindent,' ')
% MZID_PARAM::ELEMENT_DATA::NAMES[i1]
% pout_values[PercolatorOutFeatures(mzidname,psmid,i1)];
break;
}
}
pout_values.erase(PercolatorOutFeatures(mzidname,psmid,i1));
}
psmid="";
}
xmlindent=s1.find_first_not_of(" ");
*wout << s1 << endl;
if (s1.find(MZID_PARAM::PSMID_TAG)!=string::npos && s1.find(MZID_PARAM::START_INSERT_TAG)!=string::npos) {
i1=s1.find(MZID_PARAM::PSMID_TAG)+strlen(MZID_PARAM::PSMID_TAG);
psmid=s1.substr(i1,s1.find("\"",i1)-i1);
}
}
fpr.close();
fpw.close();
}
clog << boost::format(PRINT_TEXT::INSERTED) % n << endl;
for (boost::unordered_map<PercolatorOutFeatures, string, PercolatorOutFeatures>::iterator
it=pout_values.begin(); it!=pout_values.end(); it++)
cerr << boost::format(PRINT_TEXT::PSM_NOT_ENTERED) % it->first.filename % it->first.psmid << endl;
return true;
}
catch (exception &e) {
cerr << e.what() << endl;
fpr.close();
fpw.close();
return false;
}
}
//-----------------------------------------------------------------------------
std::size_t DofMapBuilder::build_constrained_vertex_indices(
const Mesh& mesh,
const std::map<unsigned int,
std::pair<unsigned int, unsigned int> >& slave_to_master_vertices,
std::vector<std::size_t>& modified_global_indices)
{
// MPI communicator
const MPI_Comm mpi_comm = mesh.mpi_comm();
// Get vertex sharing information (local index, [(sharing process p,
// local index on p)])
const boost::unordered_map<unsigned int, std::
vector<std::pair<unsigned int, unsigned int> > >
shared_vertices = DistributedMeshTools::compute_shared_entities(mesh, 0);
// Mark shared vertices
std::vector<bool> vertex_shared(mesh.num_vertices(), false);
boost::unordered_map<unsigned int, std::vector<std::pair<unsigned int,
unsigned int> > >::const_iterator shared_vertex;
for (shared_vertex = shared_vertices.begin();
shared_vertex != shared_vertices.end(); ++shared_vertex)
{
dolfin_assert(shared_vertex->first < vertex_shared.size());
vertex_shared[shared_vertex->first] = true;
}
// Mark slave vertices
std::vector<bool> slave_vertex(mesh.num_vertices(), false);
std::map<unsigned int, std::pair<unsigned int,
unsigned int> >::const_iterator slave;
for (slave = slave_to_master_vertices.begin();
slave != slave_to_master_vertices.end(); ++slave)
{
dolfin_assert(slave->first < slave_vertex.size());
slave_vertex[slave->first] = true;
}
// MPI process number
const std::size_t proc_num = MPI::rank(mesh.mpi_comm());
// Communication data structures
std::vector<std::vector<std::size_t> >
new_shared_vertex_indices(MPI::size(mesh.mpi_comm()));
// Compute modified global vertex indices
std::size_t new_index = 0;
modified_global_indices
= std::vector<std::size_t>(mesh.num_vertices(),
std::numeric_limits<std::size_t>::max());
for (VertexIterator vertex(mesh); !vertex.end(); ++vertex)
{
const std::size_t local_index = vertex->index();
if (slave_vertex[local_index])
{
// Do nothing, will get new master index later
}
else if (vertex_shared[local_index])
{
// If shared, let lowest rank process number the vertex
boost::unordered_map<unsigned int, std::vector<std::pair<unsigned int, unsigned int> > >::const_iterator
it = shared_vertices.find(local_index);
dolfin_assert(it != shared_vertices.end());
const std::vector<std::pair<unsigned int, unsigned int> >& sharing_procs
= it->second;
// Figure out if this is the lowest rank process sharing the vertex
std::vector<std::pair<unsigned int, unsigned int> >::const_iterator
min_sharing_rank = std::min_element(sharing_procs.begin(),
sharing_procs.end());
std::size_t _min_sharing_rank = proc_num + 1;
if (min_sharing_rank != sharing_procs.end())
_min_sharing_rank = min_sharing_rank->first;
if (proc_num <= _min_sharing_rank)
{
// Re-number vertex
modified_global_indices[vertex->index()] = new_index;
// Add to list to communicate
std::vector<std::pair<unsigned int, unsigned int> >::const_iterator p;
for (p = sharing_procs.begin(); p != sharing_procs.end(); ++p)
{
dolfin_assert(p->first < new_shared_vertex_indices.size());
// Local index on remote process
new_shared_vertex_indices[p->first].push_back(p->second);
// Modified global index
new_shared_vertex_indices[p->first].push_back(new_index);
}
new_index++;
}
}
else
modified_global_indices[vertex->index()] = new_index++;
}
// Send number of owned entities to compute offeset
std::size_t offset = MPI::global_offset(mpi_comm, new_index, true);
// Add process offset to modified indices
for (std::size_t i = 0; i < modified_global_indices.size(); ++i)
modified_global_indices[i] += offset;
// Add process offset to shared vertex indices before sending
for (std::size_t p = 0; p < new_shared_vertex_indices.size(); ++p)
for (std::size_t i = 1; i < new_shared_vertex_indices[p].size(); i += 2)
new_shared_vertex_indices[p][i] += offset;
// Send/receive new indices for shared vertices
std::vector<std::vector<std::size_t> > received_vertex_data;
MPI::all_to_all(mesh.mpi_comm(), new_shared_vertex_indices,
received_vertex_data);
// Set index for shared vertices that have been numbered by another
// process
for (std::size_t p = 0; p < received_vertex_data.size(); ++p)
{
const std::vector<std::size_t>& received_vertex_data_p
= received_vertex_data[p];
for (std::size_t i = 0; i < received_vertex_data_p.size(); i += 2)
{
const unsigned int local_index = received_vertex_data_p[i];
const std::size_t recv_new_index = received_vertex_data_p[i + 1];
dolfin_assert(local_index < modified_global_indices.size());
modified_global_indices[local_index] = recv_new_index;
}
}
// Request master vertex index from master owner
std::vector<std::vector<std::size_t> >
master_send_buffer(MPI::size(mpi_comm));
std::vector<std::vector<std::size_t> >
local_slave_index(MPI::size(mpi_comm));
std::map<unsigned int,
std::pair<unsigned int, unsigned int> >::const_iterator master;
for (master = slave_to_master_vertices.begin();
master != slave_to_master_vertices.end(); ++master)
{
const unsigned int local_index = master->first;
const unsigned int master_proc = master->second.first;
const unsigned int remote_master_local_index = master->second.second;
dolfin_assert(master_proc < local_slave_index.size());
dolfin_assert(master_proc < master_send_buffer.size());
local_slave_index[master_proc].push_back(local_index);
master_send_buffer[master_proc].push_back(remote_master_local_index);
}
// Send/receive master local indices for slave vertices
std::vector<std::vector<std::size_t> > received_slave_vertex_indices;
MPI::all_to_all(mpi_comm, master_send_buffer,
received_slave_vertex_indices);
// Send back new master vertex index
std::vector<std::vector<std::size_t> >
master_vertex_indices(MPI::size(mpi_comm));
for (std::size_t p = 0; p < received_slave_vertex_indices.size(); ++p)
{
const std::vector<std::size_t>& local_master_indices
= received_slave_vertex_indices[p];
for (std::size_t i = 0; i < local_master_indices.size(); ++i)
{
std::size_t master_local_index = local_master_indices[i];
dolfin_assert(master_local_index < modified_global_indices.size());
master_vertex_indices[p].push_back(modified_global_indices[master_local_index]);
}
}
// Send/receive new global master indices for slave vertices
std::vector<std::vector<std::size_t> > received_new_slave_vertex_indices;
MPI::all_to_all(mpi_comm, master_vertex_indices,
received_new_slave_vertex_indices);
// Set index for slave vertices
for (std::size_t p = 0; p < received_new_slave_vertex_indices.size(); ++p)
{
const std::vector<std::size_t>& new_indices
= received_new_slave_vertex_indices[p];
const std::vector<std::size_t>& local_indices = local_slave_index[p];
for (std::size_t i = 0; i < new_indices.size(); ++i)
{
const std::size_t local_index = local_indices[i];
const std::size_t new_global_index = new_indices[i];
dolfin_assert(local_index < modified_global_indices.size());
modified_global_indices[local_index] = new_global_index;
}
}
// Send new indices to process that share a vertex but were not
// responsible for re-numbering
return MPI::sum(mpi_comm, new_index);
}
void HDBScan::get_clusters(vector<CondensedTree*>& tree,
boost::unordered_map<int, double>& stability,
vector<int>& out_labels,
vector<double>& out_probs,
vector<double>& out_stabilities,
int cluster_selection_method,
bool allow_single_cluster,
bool match_reference_implementation)
{
vector<int> node_list;
boost::unordered_map<int, double>::iterator sit;
for (sit = stability.begin(); sit!=stability.end(); sit++) {
node_list.push_back(sit->first);
}
std::sort(node_list.begin(), node_list.end(), std::greater<int>());
if (!allow_single_cluster) {
// exclude root
node_list.pop_back();
}
vector<CondensedTree*> cluster_tree;
for (int i=0; i<tree.size(); i++) {
if (tree[i]->child_size > 1) {
cluster_tree.push_back(tree[i]);
}
}
boost::unordered_map<int, bool> is_cluster;
for (int i=0; i<node_list.size(); i++) {
is_cluster[node_list[i]] = true;
}
int num_points = DBL_MIN;
for (int i=0; i<tree.size(); i++) {
if (tree[i]->child_size == 1) {
if (tree[i]->child > num_points) {
num_points = tree[i]->child;
}
}
}
num_points += 1;
double max_lambda = tree[0]->lambda_val;
for (int i=1; i<tree.size(); i++) {
if (tree[i]->lambda_val > max_lambda) {
max_lambda = tree[i]->lambda_val;
}
}
if (cluster_selection_method == 0) {
// eom
for (int i=0; i<node_list.size(); i++) {
int node = node_list[i];
double subtree_stability = 0;
for (int j=0; j<cluster_tree.size(); j++) {
if (cluster_tree[j]->parent == node) {
//child_selection.push_back(node);
int child = cluster_tree[j]->child;
subtree_stability += stability[child];
}
}
if (subtree_stability > stability[node]) {
is_cluster[node] = false;
stability[node] = subtree_stability;
} else {
vector<int> sub_nodes = bfs_from_cluster_tree(cluster_tree, node);
for (int j=0; j<sub_nodes.size(); j++) {
int sub_node = sub_nodes[j];
if (sub_node != node) {
is_cluster[sub_node] = false;
}
}
}
}
} else if (cluster_selection_method == 1) {
// leaf
int parent_min = tree[0]->parent;
for (int i=1; i<tree.size(); i++) {
if (tree[i]->parent < parent_min) {
parent_min = tree[i]->parent;
}
}
boost::unordered_map<int, bool>::iterator it;
vector<int> leaves_ = get_cluster_tree_leaves(cluster_tree);
set<int> leaves;
for (int i=0; i<leaves_.size(); i++) {
leaves.insert(leaves_[i]);
}
if ( leaves.empty()) {
for (it=is_cluster.begin(); it!=is_cluster.end(); it++) {
is_cluster[it->first] = false;
}
is_cluster[parent_min] = true;
}
for (it=is_cluster.begin(); it!=is_cluster.end(); it++) {
int c = it->first;
if (leaves.find(c) != leaves.end()) {
is_cluster[c] = true;
} else {
is_cluster[c] = false;
}
}
}
set<int> clusters;
boost::unordered_map<int, bool>::iterator it;
for (it = is_cluster.begin(); it!= is_cluster.end(); it++) {
int c = it->first;
if (it->second) {
clusters.insert(c);
}
}
vector<int> _clusters;
set<int>::iterator _it;
for (_it =clusters.begin(); _it != clusters.end(); _it++) {
_clusters.push_back(*_it);
}
std::sort(_clusters.begin(), _clusters.end());
boost::unordered_map<int, int> cluster_map, reverse_cluster_map;
for (int i=0; i<_clusters.size(); i++) {
cluster_map[_clusters[i]] = i;
reverse_cluster_map[i] = _clusters[i];
}
// do labeling (tree, clusters, cluster_map, False, False
out_labels = do_labelling(tree, clusters, cluster_map, allow_single_cluster, match_reference_implementation);
// probs = get_probabilities(tree, reverse_cluster_map, labels)
out_probs = get_probabilities(tree, reverse_cluster_map, out_labels);
// stabilities = get_stability_scores(labels, clusters, stability, max_lambda)
out_stabilities = get_stability_scores(out_labels, clusters, stability, max_lambda);
}
void Task::featuresOut(const int current_image_idx, const boost::unordered_map<boost::uuids::uuid, StereoFeature> &hash)
{
ExteroFeatures features_samples;
base::samples::Pointcloud features_points;
double const &baseline(cameracalib.extrinsic.tx);
cv::Mat cv_image1;
/** Uncertainty in the images (left and right) planes **/
Eigen::Matrix4d px_var;
px_var << this->pxleftVar, Eigen::Matrix2d::Zero(),
Eigen::Matrix2d::Zero(), this->pxrightVar;
#ifdef DEBUG_PRINTS
std::cout<<"[FEATURES_OUT] Current hash.size(): "<<hash.size()<<"\n";
#endif
/** Parse the hash table of features **/
for (boost::unordered_map<boost::uuids::uuid, StereoFeature>::const_iterator
it = hash.begin(); it != hash.end(); ++it)
{
//std::cout<<"[FEATURES_OUT] img_idx: "<<it->second.img_idx<<" current_idx: "<<current_image_idx<<"\n";
if (it->second.img_idx == current_image_idx)
{
Feature feature;
/** Get the uuid **/
feature.index = it->first;
/** Take the "Stereo" 2d point **/
cv::Point const &left_pt(it->second.keypoint_left.pt);
cv::Point const &right_pt(it->second.keypoint_right.pt);
feature.stereo_point = base::Vector3d(left_pt.x, right_pt.x, left_pt.y);
/** Compute the 3d point **/
double disparity = std::min(right_pt.x - left_pt.x, -1);
base::Vector4d image_point (left_pt.x, left_pt.y, disparity, 1);
base::Vector4d homogeneous_point = Q * image_point;
feature.point_3d = base::Vector3d (homogeneous_point(0)/homogeneous_point(3),
homogeneous_point(1)/homogeneous_point(3),
homogeneous_point(2)/homogeneous_point(3));
#ifdef DEBUG_PRINTS
std::cout<<"[FEATURES_OUT] Image point[left]: "<<left_pt.x<<" "<<left_pt.y<<"\n";
std::cout<<"[FEATURES_OUT] Image point[right]: "<<right_pt.x<<" "<<right_pt.y<<"\n";
std::cout<<"[FEATURES_OUT] 3D point:\n"<<feature.point_3d<<"\n";
#endif
if (_output_debug.value())
{
/** 3D point coordinates in base point cloud structure **/
features_points.points.push_back(base::Vector3d (
homogeneous_point(0)/homogeneous_point(3),
homogeneous_point(1)/homogeneous_point(3),
homogeneous_point(2)/homogeneous_point(3)));
/** Color **/
cv_image1 = frame_helper::FrameHelper::convertToCvMat(this->left_color_frame);
cv::Vec3b color = cv_image1.at<cv::Vec3b>(left_pt.y, left_pt.x);
base::Vector4d color4d;
color4d[0] = color[0]/255.0;//R
color4d[1] = color[1]/255.0;//G
color4d[2] = color[2]/255.0;//B
color4d[3] = 1.0;//Alpha
features_points.colors.push_back(color4d);
}
/** Compute the covariance **/
Eigen::Matrix<double, 3, 4> noise_jacobian; /** Jacobian Matrix for the triangulation noise model */
double disparity_power = pow(disparity,2);
noise_jacobian << -(baseline*right_pt.x)/disparity_power, 0.00, (baseline*left_pt.x)/disparity_power, 0.00,
-(baseline*left_pt.y)/disparity_power, baseline/disparity, (baseline*left_pt.y)/disparity_power, 0.00,
-(baseline*cameracalib.camLeft.fx)/disparity_power, 0.00, (baseline*cameracalib.camLeft.fx)/disparity_power, 0.00;
feature.cov_3d = noise_jacobian * px_var * noise_jacobian.transpose();
/** Store the feature in the vector **/
features_samples.features.push_back(feature);
}
}
#ifdef DEBUG_PRINTS
std::cout<<"[FEATURES_OUT] Sent "<<features_samples.features.size()<<" vector of uuids\n";
#endif
features_samples.time = this->frame_pair.time;
features_samples.img_idx = current_image_idx;
_features_samples_out.write(features_samples);
if (_output_debug.value())
{
features_points.time = this->frame_pair.time;
_features_point_samples_out.write(features_points);
}
return;
}
