mirror of https://github.com/CGAL/cgal
478 lines
14 KiB
C++
Executable File
478 lines
14 KiB
C++
Executable File
// ======================================================================
|
|
//
|
|
// Copyright (c) 2002 The CGAL Consortium
|
|
//
|
|
// This software and related documentation is part of an INTERNAL release
|
|
// of the Computational Geometry Algorithms Library (CGAL). It is not
|
|
// intended for general use.
|
|
//
|
|
// ----------------------------------------------------------------------
|
|
//
|
|
// release :
|
|
// release_date :
|
|
//
|
|
// file : include/CGAL/Orthogonal_priority_search.h
|
|
// package : ASPAS
|
|
// revision : 2.4
|
|
// revision_date : 2002/16/08
|
|
// authors : Hans Tangelder (<hanst@cs.uu.nl>)
|
|
// maintainer : Hans Tangelder (<hanst@cs.uu.nl>)
|
|
// coordinator : Utrecht University
|
|
//
|
|
// ======================================================================
|
|
|
|
#ifndef ORTHOGONAL_PRIORITY_SEARCH
|
|
#define ORTHOGONAL_PRIORITY_SEARCH
|
|
#include <cstring>
|
|
#include <list>
|
|
#include <queue>
|
|
#include <memory>
|
|
#include <CGAL/Kd_tree_node.h>
|
|
|
|
namespace CGAL {
|
|
|
|
template <class Traits, class Query_item, class Distance>
|
|
class Orthogonal_priority_search {
|
|
|
|
public:
|
|
|
|
typedef typename Traits::Item Item;
|
|
typedef typename Traits::NT NT;
|
|
typedef Item** Item_iterator;
|
|
typedef Kd_tree_node<Traits> Node;
|
|
typedef Kd_tree<Traits> Tree;
|
|
|
|
|
|
typedef std::pair<Item*,NT> Item_with_distance;
|
|
typedef std::pair<Node*,NT> Node_with_distance;
|
|
|
|
|
|
|
|
// this forward declaration may cause problems for g++
|
|
class iterator;
|
|
|
|
|
|
|
|
typedef std::vector<Node_with_distance*> Node_with_distance_vector;
|
|
|
|
typedef std::vector<Item_with_distance*> Item_with_distance_vector;
|
|
|
|
typedef std::vector<NT> Distance_vector;
|
|
|
|
iterator *start;
|
|
iterator *past_the_end;
|
|
|
|
public:
|
|
|
|
// constructor
|
|
Orthogonal_priority_search(Tree& tree,
|
|
Query_item& q, Distance& tr, NT Eps = NT(0.0),
|
|
bool search_nearest=true)
|
|
{
|
|
|
|
start = new iterator(tree,q,tr,Eps,search_nearest);
|
|
past_the_end = new iterator();
|
|
|
|
};
|
|
|
|
// destructor
|
|
~Orthogonal_priority_search() {
|
|
delete start;
|
|
delete past_the_end;
|
|
};
|
|
|
|
iterator begin() {
|
|
return *start;
|
|
}
|
|
|
|
iterator end() {
|
|
return *past_the_end;
|
|
}
|
|
|
|
void statistics() {
|
|
start->statistics();
|
|
}
|
|
|
|
class iterator {
|
|
|
|
public:
|
|
|
|
typedef std::input_iterator_tag iterator_category;
|
|
typedef Item_with_distance value_type;
|
|
typedef int distance_type;
|
|
|
|
class Iterator_implementation;
|
|
Iterator_implementation *Ptr_implementation;
|
|
|
|
|
|
public:
|
|
|
|
// default constructor
|
|
iterator() {Ptr_implementation=0;}
|
|
|
|
int the_number_of_items_visited() {
|
|
return Ptr_implementation->number_of_items_visited;
|
|
}
|
|
|
|
// constructor
|
|
iterator(Tree& tree, Query_item& q, Distance& tr, NT eps=NT(0.0),
|
|
bool search_nearest=true){
|
|
Ptr_implementation =
|
|
new Iterator_implementation(tree, q, tr, eps, search_nearest);
|
|
}
|
|
|
|
// copy constructor
|
|
iterator(const iterator& Iter) {
|
|
Ptr_implementation = Iter.Ptr_implementation;
|
|
if (Ptr_implementation != 0) Ptr_implementation->reference_count++;
|
|
}
|
|
|
|
Item_with_distance& operator* () {
|
|
return *(*Ptr_implementation);
|
|
}
|
|
|
|
// prefix operator
|
|
iterator& operator++() {
|
|
++(*Ptr_implementation);
|
|
return *this;
|
|
}
|
|
|
|
// postfix operator
|
|
std::auto_ptr<Item_with_distance> operator++(int) {
|
|
std::auto_ptr<Item_with_distance> result = (*Ptr_implementation)++;
|
|
return result;
|
|
}
|
|
|
|
bool operator==(const iterator& It) const {
|
|
|
|
if (
|
|
((Ptr_implementation == 0) ||
|
|
Ptr_implementation->Item_PriorityQueue.empty()) &&
|
|
((It.Ptr_implementation == 0) ||
|
|
It.Ptr_implementation->Item_PriorityQueue.empty())
|
|
)
|
|
return true;
|
|
// else
|
|
return (Ptr_implementation == It.Ptr_implementation);
|
|
}
|
|
|
|
bool operator!=(const iterator& It) const {
|
|
return !(*this == It);
|
|
}
|
|
|
|
void statistics () {
|
|
Ptr_implementation->statistics();
|
|
}
|
|
|
|
~iterator() {
|
|
if (Ptr_implementation != 0) {
|
|
Ptr_implementation->reference_count--;
|
|
if (Ptr_implementation->reference_count==0) {
|
|
delete Ptr_implementation;
|
|
Ptr_implementation = 0;
|
|
}
|
|
}
|
|
}
|
|
|
|
|
|
class Iterator_implementation {
|
|
|
|
public:
|
|
|
|
int number_of_neighbours_computed;
|
|
int number_of_internal_nodes_visited;
|
|
int number_of_leaf_nodes_visited;
|
|
int number_of_items_visited;
|
|
|
|
private:
|
|
|
|
|
|
NT multiplication_factor;
|
|
|
|
Item* query_point;
|
|
|
|
int total_item_number;
|
|
|
|
NT distance_to_root;
|
|
|
|
bool search_nearest_neighbour;
|
|
|
|
NT rd;
|
|
|
|
class Priority_higher
|
|
{
|
|
public:
|
|
|
|
bool search_nearest;
|
|
|
|
Priority_higher(bool search_the_nearest_neighbour) {
|
|
search_nearest = search_the_nearest_neighbour;
|
|
}
|
|
|
|
//highest priority is smallest distance
|
|
bool operator() (Node_with_distance* n1, Node_with_distance* n2) const
|
|
{
|
|
if (search_nearest) { return (n1->second > n2->second);}
|
|
else {return (n2->second > n1->second);}
|
|
}
|
|
};
|
|
|
|
class Distance_smaller
|
|
{
|
|
|
|
public:
|
|
|
|
bool search_nearest;
|
|
|
|
Distance_smaller(bool search_the_nearest_neighbour) {
|
|
// Search_traits s;
|
|
search_nearest = search_the_nearest_neighbour;
|
|
}
|
|
|
|
//highest priority is smallest distance
|
|
bool operator() (Item_with_distance* p1, Item_with_distance* p2) const
|
|
{
|
|
if (search_nearest) {return (p1->second > p2->second);}
|
|
else {return (p2->second > p1->second);}
|
|
}
|
|
};
|
|
|
|
std::priority_queue<Node_with_distance*, Node_with_distance_vector,
|
|
Priority_higher>* PriorityQueue;
|
|
|
|
std::priority_queue<Item_with_distance*, Item_with_distance_vector,
|
|
Distance_smaller>* Item_PriorityQueue;
|
|
|
|
Distance* Orthogonal_Distance_instance;
|
|
|
|
public:
|
|
|
|
int reference_count;
|
|
|
|
|
|
|
|
// constructor
|
|
Iterator_implementation(Tree& tree, Query_item& q, Distance& tr,
|
|
NT Eps=NT(0.0), bool search_nearest=true)
|
|
{
|
|
PriorityQueue= new std::priority_queue<Node_with_distance*,
|
|
Node_with_distance_vector,
|
|
Priority_higher>
|
|
(Priority_higher(search_nearest));
|
|
|
|
Item_PriorityQueue = new std::priority_queue<Item_with_distance*,
|
|
Item_with_distance_vector,
|
|
Distance_smaller>
|
|
(Distance_smaller(search_nearest));
|
|
|
|
search_nearest_neighbour=search_nearest;
|
|
reference_count=1;
|
|
Orthogonal_Distance_instance=&tr;
|
|
multiplication_factor=
|
|
Orthogonal_Distance_instance->transformed_distance(NT(1.0)+Eps);
|
|
|
|
if (search_nearest) distance_to_root=
|
|
Orthogonal_Distance_instance->min_distance_to_queryitem(q,
|
|
*(tree.bounding_box()));
|
|
else distance_to_root=
|
|
Orthogonal_Distance_instance->max_distance_to_queryitem(q,
|
|
*(tree.bounding_box()));
|
|
|
|
|
|
query_point = &q;
|
|
|
|
total_item_number=tree.item_number();
|
|
|
|
number_of_leaf_nodes_visited=0;
|
|
number_of_internal_nodes_visited=0;
|
|
number_of_items_visited=0;
|
|
number_of_neighbours_computed=0;
|
|
|
|
|
|
|
|
Node_with_distance *The_Root = new Node_with_distance(tree.root(),
|
|
distance_to_root);
|
|
PriorityQueue->push(The_Root);
|
|
|
|
// rd is the distance of the top of the priority queue to q
|
|
rd=The_Root->second;
|
|
Compute_the_next_nearest_neighbour();
|
|
}
|
|
|
|
// * operator
|
|
Item_with_distance& operator* () {
|
|
return *(Item_PriorityQueue->top());
|
|
}
|
|
|
|
// prefix operator
|
|
Iterator_implementation& operator++() {
|
|
// std::cout << "called ++" << std::endl;
|
|
Delete_the_current_item_top();
|
|
Compute_the_next_nearest_neighbour();
|
|
return *this;
|
|
}
|
|
|
|
// postfix operator
|
|
std::auto_ptr<Item_with_distance> operator++(int) {
|
|
Item_with_distance Value = *(Item_PriorityQueue->top());
|
|
std::auto_ptr<Item_with_distance>
|
|
result(new Item_with_distance(Value));
|
|
++*this;
|
|
return result;
|
|
}
|
|
|
|
// Print statistics of the general priority search process.
|
|
void statistics () {
|
|
std::cout << "Orthogonal priority search statistics:"
|
|
<< std::endl;
|
|
std::cout << "Number of internal nodes visited:"
|
|
<< number_of_internal_nodes_visited << std::endl;
|
|
std::cout << "Number of leaf nodes visited:"
|
|
<< number_of_leaf_nodes_visited << std::endl;
|
|
std::cout << "Number of items visited:"
|
|
<< number_of_items_visited << std::endl;
|
|
std::cout << "Number of neighbours computed:"
|
|
<< number_of_neighbours_computed << std::endl;
|
|
}
|
|
|
|
|
|
//destructor
|
|
~Iterator_implementation() {
|
|
|
|
|
|
while (PriorityQueue->size()>0) {
|
|
Node_with_distance* The_top=PriorityQueue->top();
|
|
PriorityQueue->pop();
|
|
delete The_top;
|
|
};
|
|
while (Item_PriorityQueue->size()>0) {
|
|
Item_with_distance* The_top=Item_PriorityQueue->top();
|
|
Item_PriorityQueue->pop();
|
|
delete The_top;
|
|
};
|
|
delete PriorityQueue;
|
|
delete Item_PriorityQueue;
|
|
|
|
}
|
|
|
|
private:
|
|
|
|
void Delete_the_current_item_top() {
|
|
Item_with_distance* The_item_top=Item_PriorityQueue->top();
|
|
Item_PriorityQueue->pop();
|
|
delete The_item_top;
|
|
}
|
|
|
|
void Compute_the_next_nearest_neighbour() {
|
|
|
|
// compute the next item
|
|
bool next_neighbour_found=false;
|
|
if (!(Item_PriorityQueue->empty())) {
|
|
if (search_nearest_neighbour)
|
|
next_neighbour_found=
|
|
(multiplication_factor*rd > Item_PriorityQueue->top()->second);
|
|
else
|
|
next_neighbour_found=
|
|
(multiplication_factor*rd < Item_PriorityQueue->top()->second);
|
|
}
|
|
// otherwise browse the tree further
|
|
while ((!next_neighbour_found) && (!PriorityQueue->empty())) {
|
|
Node_with_distance* The_node_top=PriorityQueue->top();
|
|
Node* N= The_node_top->first;
|
|
PriorityQueue->pop();
|
|
delete The_node_top;
|
|
|
|
while (!(N->is_leaf())) { // compute new distance
|
|
number_of_internal_nodes_visited++;
|
|
int new_cut_dim=N->cutting_dimension();
|
|
NT old_off, new_rd;
|
|
NT new_off =
|
|
(*query_point)[new_cut_dim] -
|
|
N->cutting_value();
|
|
if ( ((new_off < NT(0.0)) &&
|
|
(search_nearest_neighbour)) ||
|
|
(( new_off >= NT(0.0)) && (!search_nearest_neighbour))
|
|
) {
|
|
old_off=
|
|
(*query_point)[new_cut_dim]-N->low_value();
|
|
if (old_off>NT(0.0)) old_off=NT(0.0);
|
|
new_rd=
|
|
Orthogonal_Distance_instance->
|
|
new_distance(rd,old_off,new_off,new_cut_dim);
|
|
Node_with_distance *Upper_Child =
|
|
new Node_with_distance(N->upper(),new_rd);
|
|
PriorityQueue->push(Upper_Child);
|
|
N=N->lower();
|
|
|
|
}
|
|
else { // compute new distance
|
|
old_off= N->high_value() -
|
|
(*query_point)[new_cut_dim];
|
|
if (old_off>NT(0.0)) old_off=NT(0.0);
|
|
new_rd=Orthogonal_Distance_instance->
|
|
new_distance(rd,old_off,new_off,new_cut_dim);
|
|
Node_with_distance *Lower_Child =
|
|
new Node_with_distance(N->lower(),new_rd);
|
|
PriorityQueue->push(Lower_Child);
|
|
N=N->upper();
|
|
}
|
|
}
|
|
// n is a leaf
|
|
number_of_leaf_nodes_visited++;
|
|
if (N->size() > 0) {
|
|
for (Item_iterator it=N->begin(); it != N->end(); it++) {
|
|
number_of_items_visited++;
|
|
NT distance_to_query_point=
|
|
Orthogonal_Distance_instance->
|
|
distance(*query_point,**it);
|
|
Item_with_distance *NN_Candidate=
|
|
new Item_with_distance(*it,distance_to_query_point);
|
|
Item_PriorityQueue->push(NN_Candidate);
|
|
};
|
|
// old top of PriorityQueue has been processed,
|
|
// hence update rd
|
|
|
|
if (!(PriorityQueue->empty())) {
|
|
rd = PriorityQueue->top()->second;
|
|
if (search_nearest_neighbour)
|
|
next_neighbour_found =
|
|
(multiplication_factor*rd >
|
|
Item_PriorityQueue->top()->second);
|
|
else
|
|
next_neighbour_found =
|
|
(multiplication_factor*rd <
|
|
Item_PriorityQueue->top()->second);
|
|
}
|
|
else // priority queue empty => last neighbour found
|
|
{
|
|
next_neighbour_found=true;
|
|
};
|
|
|
|
number_of_neighbours_computed++;
|
|
}
|
|
} // next_neighbour_found or priority queue is empty
|
|
// in the latter case also the item priority quee is empty
|
|
}
|
|
}; // class Iterator_implementaion
|
|
}; // class iterator
|
|
}; // class
|
|
|
|
template <class Traits, class Query_item, class Distance>
|
|
void swap (typename Orthogonal_priority_search<Traits,
|
|
Query_item, Distance>::iterator& x,
|
|
typename Orthogonal_priority_search<Traits,
|
|
Query_item, Distance>::iterator& y) {
|
|
typename Orthogonal_priority_search<Traits,
|
|
Query_item, Distance>::iterator::Iterator_implementation
|
|
*tmp = x.Ptr_implementation;
|
|
x.Ptr_implementation = y.Ptr_implementation;
|
|
y.Ptr_implementation = tmp;
|
|
}
|
|
|
|
|
|
|
|
} // namespace CGAL
|
|
|
|
|
|
#endif // ORTHOGONAL_PRIORITY_SEARCH_H
|