First pass at DRC RTree functionality

This implements a copper-layer RTree with functions for iterating over
the elements in a copper layer and providing Nearest Neighbor returns
for BOARD_CONNECTED_ITEMS
This commit is contained in:
Seth Hillbrand 2020-08-11 16:50:56 -07:00
parent fedc6519cd
commit 7c455f2357
6 changed files with 438 additions and 94 deletions

View File

@ -124,7 +124,7 @@ public:
*/ */
virtual bool Collide( const VECTOR2I& aP, int aClearance = 0, int* aActual = nullptr ) const virtual bool Collide( const VECTOR2I& aP, int aClearance = 0, int* aActual = nullptr ) const
{ {
return Collide( SEG( aP, aP ), aClearance ); return Collide( SEG( aP, aP ), aClearance, aActual );
} }
/** /**

View File

@ -115,11 +115,21 @@ bool SHAPE_COMPOUND::IsSolid() const
bool SHAPE_COMPOUND::Collide( const SEG& aSeg, int aClearance, int* aActual ) const bool SHAPE_COMPOUND::Collide( const SEG& aSeg, int aClearance, int* aActual ) const
{ {
int dist = std::numeric_limits<int>::max();
for( auto& item : m_shapes ) for( auto& item : m_shapes )
{ {
if( item->Collide( aSeg, aClearance, aActual ) ) if( item->Collide( aSeg, aClearance, aActual ) )
return true; {
if( !aActual || *aActual == 0 )
return true;
dist = std::min( dist, *aActual );
}
} }
return false; if( aActual )
*aActual = dist;
return dist != std::numeric_limits<int>::max();
} }

View File

@ -66,6 +66,7 @@ include_directories(
${CMAKE_SOURCE_DIR}/pcbnew/dialogs ${CMAKE_SOURCE_DIR}/pcbnew/dialogs
${CMAKE_SOURCE_DIR}/polygon ${CMAKE_SOURCE_DIR}/polygon
${CMAKE_SOURCE_DIR}/common/geometry ${CMAKE_SOURCE_DIR}/common/geometry
${CMAKE_SOURCE_DIR}/libs/kimath/include/math
${CMAKE_SOURCE_DIR}/qa/common ${CMAKE_SOURCE_DIR}/qa/common
${CMAKE_SOURCE_DIR}/qa ${CMAKE_SOURCE_DIR}/qa
${CMAKE_SOURCE_DIR}/qa/qa_utils ${CMAKE_SOURCE_DIR}/qa/qa_utils

308
qa/drc_proto/drc_rtree.h Normal file
View File

@ -0,0 +1,308 @@
/*
* This program source code file is part of KiCad, a free EDA CAD application.
*
* Copyright (C) 2019 KiCad Developers, see AUTHORS.txt for contributors.
* Copyright (C) 2020 CERN
*
* 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 3
* 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, you may find one here:
* http://www.gnu.org/licenses/old-licenses/gpl-3.0.html
* or you may search the http://www.gnu.org website for the version 3 license,
* or you may write to the Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
*/
#ifndef DRC_RTREE_H_
#define DRC_RTREE_H_
#include <eda_rect.h>
#include <board_connected_item.h>
#include <set>
#include <vector>
#include <geometry/rtree.h>
#include <vector2d.h>
/**
* DRC_RTREE -
* Implements an R-tree for fast spatial and layer indexing of connectable items.
* Non-owning.
*/
class DRC_RTREE
{
private:
using drc_rtree = RTree<BOARD_CONNECTED_ITEM*, int, 2, double>;
public:
DRC_RTREE()
{
for( int layer : LSET::AllCuMask().Seq() )
m_tree[layer] = new drc_rtree();
m_count = 0;
}
~DRC_RTREE()
{
for( auto tree : m_tree )
delete tree;
}
/**
* Function Insert()
* Inserts an item into the tree. Item's bounding box is taken via its GetBoundingBox() method.
*/
void insert( BOARD_CONNECTED_ITEM* aItem )
{
const EDA_RECT& bbox = aItem->GetBoundingBox();
const int mmin[2] = { bbox.GetX(), bbox.GetY() };
const int mmax[2] = { bbox.GetRight(), bbox.GetBottom() };
for( int layer : aItem->GetLayerSet().Seq() )
m_tree[layer]->Insert( mmin, mmax, aItem );
m_count++;
}
/**
* Function Remove()
* Removes an item from the tree. Removal is done by comparing pointers, attempting
* to remove a copy of the item will fail.
*/
bool remove( BOARD_CONNECTED_ITEM* aItem )
{
// First, attempt to remove the item using its given BBox
const EDA_RECT& bbox = aItem->GetBoundingBox();
const int mmin[2] = { bbox.GetX(), bbox.GetY() };
const int mmax[2] = { bbox.GetRight(), bbox.GetBottom() };
bool removed = false;
for( auto layer : aItem->GetLayerSet().Seq() )
{
// If we are not successful ( true == not found ), then we expand
// the search to the full tree
if( m_tree[int( layer )]->Remove( mmin, mmax, aItem ) )
{
// N.B. We must search the whole tree for the pointer to remove
// because the item may have been moved before we have the chance to
// delete it from the tree
const int mmin2[2] = { INT_MIN, INT_MIN };
const int mmax2[2] = { INT_MAX, INT_MAX };
if( m_tree[int( layer )]->Remove( mmin2, mmax2, aItem ) )
continue;
}
removed = true;
}
m_count -= int( removed );
return removed;
}
/**
* Function RemoveAll()
* Removes all items from the RTree
*/
void clear()
{
for( auto tree : m_tree )
tree->RemoveAll();
m_count = 0;
}
/**
* Determine if a given item exists in the tree. Note that this does not search the full tree
* so if the item has been moved, this will return false when it should be true.
*
* @param aItem Item that may potentially exist in the tree
* @param aRobust If true, search the whole tree, not just the bounding box
* @return true if the item definitely exists, false if it does not exist within bbox
*/
bool contains( BOARD_CONNECTED_ITEM* aItem, bool aRobust = false )
{
const EDA_RECT& bbox = aItem->GetBoundingBox();
const int mmin[2] = { bbox.GetX(), bbox.GetY() };
const int mmax[2] = { bbox.GetRight(), bbox.GetBottom() };
bool found = false;
auto search = [&found, &aItem]( const BOARD_CONNECTED_ITEM* aSearchItem ) {
if( aSearchItem == aItem )
{
found = true;
return false;
}
return true;
};
for( int layer : aItem->GetLayerSet().Seq() )
{
m_tree[layer]->Search( mmin, mmax, search );
if( found )
break;
}
if( !found && aRobust )
{
for( int layer : LSET::AllCuMask().Seq() )
{
// N.B. We must search the whole tree for the pointer to remove
// because the item may have been moved. We do not expand the item
// layer search as this should not change.
const int mmin2[2] = { INT_MIN, INT_MIN };
const int mmax2[2] = { INT_MAX, INT_MAX };
m_tree[layer]->Search( mmin2, mmax2, search );
if( found )
break;
}
}
return found;
}
std::vector<std::pair<int, BOARD_CONNECTED_ITEM*>> GetNearest( const wxPoint &aPoint,
PCB_LAYER_ID aLayer,
int aLimit )
{
const int point[2] = { aPoint.x, aPoint.y };
auto result = m_tree[int( aLayer )]->NearestNeighbors( point,
[aLimit]( std::size_t a_count, int a_maxDist ) -> bool
{
return a_count >= aLimit;
},
[]( BOARD_CONNECTED_ITEM* aElement) -> bool
{
// Don't remove any elements from the list
return false;
},
[aLayer]( const int* a_point, BOARD_CONNECTED_ITEM* a_data ) -> int
{
switch( a_data->Type() )
{
case PCB_TRACE_T:
{
TRACK* track = static_cast<TRACK*>( a_data );
SEG seg( track->GetStart(), track->GetEnd() );
return seg.Distance( VECTOR2I( a_point[0], a_point[1] ) ) -
( track->GetWidth() + 1 ) / 2;
}
case PCB_VIA_T:
{
VIA* via = static_cast<VIA*>( a_data );
return ( VECTOR2I( via->GetPosition() ) -
VECTOR2I( a_point[0], a_point[1] ) ).EuclideanNorm() -
( via->GetWidth() + 1 ) / 2;
}
default:
{
VECTOR2I point( a_point[0], a_point[1] );
int dist = 0;
auto shape = a_data->GetEffectiveShape( aLayer );
// Here we use a hack to get the distance by colliding with a large area
// However, we can't use just MAX_INT because we will overflow the collision calculations
shape->Collide( point, std::numeric_limits<int>::max() / 2, &dist);
return dist;
}
}
return 0;
});
return result;
}
/**
* Returns the number of items in the tree
* @return number of elements in the tree;
*/
size_t size()
{
return m_count;
}
bool empty()
{
return m_count == 0;
}
using iterator = typename drc_rtree::Iterator;
/**
* The DRC_LAYER struct provides a layer-specific auto-range iterator to the RTree. Using
* this struct, one can write lines like:
*
* for( auto item : rtree.OnLayer( In1_Cu ) )
*
* and iterate over only the RTree items that are on In1
*/
struct DRC_LAYER
{
DRC_LAYER( drc_rtree* aTree ) : layer_tree( aTree )
{
m_rect = { { INT_MIN, INT_MIN }, { INT_MAX, INT_MAX } };
};
DRC_LAYER( drc_rtree* aTree, const EDA_RECT aRect ) : layer_tree( aTree )
{
m_rect = { { aRect.GetX(), aRect.GetY() },
{ aRect.GetRight(), aRect.GetBottom() } };
};
drc_rtree::Rect m_rect;
drc_rtree* layer_tree;
iterator begin()
{
return layer_tree->begin( m_rect );
}
iterator end()
{
return layer_tree->end( m_rect );
}
};
DRC_LAYER OnLayer( PCB_LAYER_ID aLayer )
{
return DRC_LAYER( m_tree[int( aLayer )] );
}
DRC_LAYER Overlapping( PCB_LAYER_ID aLayer, const wxPoint& aPoint, int aAccuracy = 0 )
{
EDA_RECT rect( aPoint, wxSize( 0, 0 ) );
rect.Inflate( aAccuracy );
return DRC_LAYER( m_tree[int( aLayer )], rect );
}
DRC_LAYER Overlapping( PCB_LAYER_ID aLayer, const EDA_RECT& aRect )
{
return DRC_LAYER( m_tree[int( aLayer )], aRect );
}
private:
drc_rtree* m_tree[MAX_CU_LAYERS];
size_t m_count;
};
#endif /* DRC_RTREE_H_ */

View File

@ -36,6 +36,7 @@
#include <drc_proto/drc_engine.h> #include <drc_proto/drc_engine.h>
#include <drc_proto/drc_item.h> #include <drc_proto/drc_item.h>
#include <drc_proto/drc_rtree.h>
#include <drc_proto/drc_rule.h> #include <drc_proto/drc_rule.h>
#include <drc_proto/drc_test_provider_clearance_base.h> #include <drc_proto/drc_test_provider_clearance_base.h>

View File

@ -15,10 +15,32 @@
// * 2004 Templated C++ port by Greg Douglas // * 2004 Templated C++ port by Greg Douglas
// * 2013 CERN (www.cern.ch) // * 2013 CERN (www.cern.ch)
// * 2020 KiCad Developers - Add std::iterator support for searching // * 2020 KiCad Developers - Add std::iterator support for searching
// * 2020 KiCad Developers - Add container nearest neighbor based on Hjaltason & Samet
// //
//LICENSE:
// /*
// Entirely free for all uses. Enjoy! * This program source code file is part of KiCad, a free EDA CAD application.
*
* Copyright (C) 2020 KiCad Developers, see AUTHORS.txt for contributors.
* Copyright (C) 2013 CERN
*
* 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 3
* 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, you may find one here:
* http://www.gnu.org/licenses/old-licenses/gpl-3.0.html
* or you may search the http://www.gnu.org website for the version 3 license,
* or you may write to the Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
*/
#ifndef RTREE_H #ifndef RTREE_H
#define RTREE_H #define RTREE_H
@ -35,6 +57,8 @@
#include <array> #include <array>
#include <functional> #include <functional>
#include <iterator> #include <iterator>
#include <queue>
#include <vector>
#ifdef DEBUG #ifdef DEBUG
#define ASSERT assert // RTree uses ASSERT( condition ) #define ASSERT assert // RTree uses ASSERT( condition )
@ -202,23 +226,19 @@ public:
/// Save tree contents to stream /// Save tree contents to stream
bool Save( RTFileStream& a_stream ); bool Save( RTFileStream& a_stream );
/// Find the nearest neighbor of a specific point. /**
/// It uses the MINDIST method, simplifying the one from "R-Trees: Theory and Applications" by Yannis Manolopoulos et al. * Gets an ordered vector of the nearest data elements to a specified point
/// The bounding rectangle is used to calculate the distance to the DATATYPE. * @param aPoint coordinate to measure against
/// \param a_point point to start the search * @param aTerminate Callback routine to check when we have gathered sufficient elements
/// \return Returns the DATATYPE located closest to a_point, 0 if the tree is empty. * @param aFilter Callback routine to remove specific elements from the query results
DATATYPE NearestNeighbor( const ELEMTYPE a_point[NUMDIMS] ); * @param aSquaredDist Callback routine to measure the distance from the point to the data element
* @return vector of matching elements and their distance to the point
/// Find the nearest neighbor of a specific point. */
/// It uses the MINDIST method, simplifying the one from "R-Trees: Theory and Applications" by Yannis Manolopoulos et al. std::vector<std::pair<ELEMTYPE, DATATYPE>> NearestNeighbors(
/// It receives a callback function to calculate the distance to a DATATYPE object, instead of using the bounding rectangle. const ELEMTYPE aPoint[NUMDIMS],
/// \param a_point point to start the search std::function<bool( const std::size_t aNumResults, const ELEMTYPE aMinDist )> aTerminate,
/// \param a_squareDistanceCallback function that performs the square distance calculation for the selected DATATYPE. std::function<bool( const DATATYPE aElement )> aFilter,
/// \param a_squareDistance Pointer in which the square distance to the nearest neighbour will be returned. std::function<ELEMTYPE( const ELEMTYPE a_point[NUMDIMS], const DATATYPE a_data )> aSquaredDist );
/// \return Returns the DATATYPE located closest to a_point, 0 if the tree is empty.
DATATYPE NearestNeighbor( const ELEMTYPE a_point[NUMDIMS],
ELEMTYPE a_squareDistanceCallback( const ELEMTYPE a_point[NUMDIMS], DATATYPE a_data ),
ELEMTYPE* a_squareDistance );
public: public:
/// Iterator is not remove safe. /// Iterator is not remove safe.
@ -495,6 +515,12 @@ protected:
Branch m_branch; Branch m_branch;
ELEMTYPE minDist; ELEMTYPE minDist;
bool isLeaf; bool isLeaf;
inline bool operator<(const NNNode &other) const
{
/// This is reversed on purpose to use std::priority_queue
return other.minDist < minDist;
}
}; };
Node* AllocNode(); Node* AllocNode();
@ -531,8 +557,7 @@ protected:
void FreeListNode( ListNode* a_listNode ); void FreeListNode( ListNode* a_listNode );
static bool Overlap( Rect* a_rectA, Rect* a_rectB ); static bool Overlap( Rect* a_rectA, Rect* a_rectB );
void ReInsert( Node* a_node, ListNode** a_listNode ); void ReInsert( Node* a_node, ListNode** a_listNode );
ELEMTYPE MinDist( const ELEMTYPE a_point[NUMDIMS], Rect* a_rect ); ELEMTYPE MinDist( const ELEMTYPE a_point[NUMDIMS], const Rect& a_rect );
void InsertNNListSorted( std::vector<NNNode*>* nodeList, NNNode* newNode );
bool Search( Node * a_node, Rect * a_rect, int& a_foundCount, bool Search( Node * a_node, Rect * a_rect, int& a_foundCount,
std::function<bool (const DATATYPE&)> a_callback ) const; std::function<bool (const DATATYPE&)> a_callback ) const;
@ -815,58 +840,67 @@ int RTREE_QUAL::Search( const ELEMTYPE a_min[NUMDIMS], const ELEMTYPE a_max[NUMD
RTREE_TEMPLATE RTREE_TEMPLATE
DATATYPE RTREE_QUAL::NearestNeighbor( const ELEMTYPE a_point[NUMDIMS] ) std::vector<std::pair<ELEMTYPE, DATATYPE>> RTREE_QUAL::NearestNeighbors(
const ELEMTYPE a_point[NUMDIMS],
std::function<bool( const std::size_t aNumResults, const ELEMTYPE aMinDist )> aTerminate,
std::function<bool( const DATATYPE aElement )> aFilter,
std::function<ELEMTYPE( const ELEMTYPE a_point[NUMDIMS], const DATATYPE a_data )> aSquaredDist )
{ {
return this->NearestNeighbor( a_point, 0, 0 ); std::vector<std::pair<ELEMTYPE, DATATYPE>> result;
} std::priority_queue<NNNode> search_q;
for( int i = 0; i < m_root->m_count; ++i )
RTREE_TEMPLATE
DATATYPE RTREE_QUAL::NearestNeighbor( const ELEMTYPE a_point[NUMDIMS],
ELEMTYPE a_squareDistanceCallback( const ELEMTYPE a_point[NUMDIMS], DATATYPE a_data ),
ELEMTYPE* a_squareDistance )
{
std::vector<NNNode*> nodeList;
Node* node = m_root;
NNNode* closestNode = 0;
while( !closestNode || !closestNode->isLeaf )
{ {
//check every node on this level if( m_root->IsLeaf() )
for( int index = 0; index < node->m_count; ++index )
{ {
NNNode* newNode = new NNNode; search_q.push( NNNode{ m_root->m_branch[i],
newNode->isLeaf = node->IsLeaf(); aSquaredDist( a_point, m_root->m_branch[i].m_data ),
newNode->m_branch = node->m_branch[index]; m_root->IsLeaf() });
if( newNode->isLeaf && a_squareDistanceCallback )
newNode->minDist = a_squareDistanceCallback( a_point, newNode->m_branch.m_data );
else
newNode->minDist = this->MinDist( a_point, &(node->m_branch[index].m_rect) );
//TODO: a custom list could be more efficient than a vector
this->InsertNNListSorted( &nodeList, newNode );
} }
if( nodeList.size() == 0 ) else
{ {
return 0; search_q.push( NNNode{ m_root->m_branch[i],
MinDist(a_point, m_root->m_branch[i].m_rect),
m_root->IsLeaf() });
} }
closestNode = nodeList.back();
node = closestNode->m_branch.m_child;
nodeList.pop_back();
free(closestNode);
} }
// free memory used for remaining NNNodes in nodeList while( !search_q.empty() )
for( auto node_it : nodeList )
{ {
NNNode* nnode = node_it; const NNNode curNode = search_q.top();
free(nnode);
if( aTerminate( result.size(), curNode.minDist ) )
break;
search_q.pop();
if( curNode.isLeaf )
{
if( aFilter( curNode.m_branch.m_data ) )
result.emplace_back( curNode.minDist, curNode.m_branch.m_data );
}
else
{
Node* node = curNode.m_branch.m_child;
for( int i = 0; i < node->m_count; ++i )
{
NNNode newNode;
newNode.isLeaf = node->IsLeaf();
newNode.m_branch = node->m_branch[i];
if( newNode.isLeaf )
newNode.minDist = aSquaredDist( a_point, newNode.m_branch.m_data );
else
newNode.minDist = this->MinDist( a_point, node->m_branch[i].m_rect );
search_q.push( newNode );
}
}
} }
*a_squareDistance = closestNode->minDist; return result;
return closestNode->m_branch.m_data;
} }
RTREE_TEMPLATE RTREE_TEMPLATE
int RTREE_QUAL::Count() int RTREE_QUAL::Count()
{ {
@ -1895,7 +1929,7 @@ void RTREE_QUAL::ReInsert( Node* a_node, ListNode** a_listNode )
} }
// Search in an index tree or subtree for all data retangles that overlap the argument rectangle. // Search in an index tree or subtree for all data rectangles that overlap the argument rectangle.
RTREE_TEMPLATE RTREE_TEMPLATE
bool RTREE_QUAL::Search( Node* a_node, Rect* a_rect, int& a_foundCount, bool RTREE_QUAL::Search( Node* a_node, Rect* a_rect, int& a_foundCount,
std::function<bool (const DATATYPE&)> a_callback ) const std::function<bool (const DATATYPE&)> a_callback ) const
@ -1939,44 +1973,34 @@ bool RTREE_QUAL::Search( Node* a_node, Rect* a_rect, int& a_foundCount,
//calculate the minimum distance between a point and a rectangle as defined by Manolopoulos et al. //calculate the minimum distance between a point and a rectangle as defined by Manolopoulos et al.
//it uses the square distance to avoid the use of ELEMTYPEREAL values, which are slower. // returns Euclidean norm to ensure value fits in ELEMTYPE
RTREE_TEMPLATE RTREE_TEMPLATE
ELEMTYPE RTREE_QUAL::MinDist( const ELEMTYPE a_point[NUMDIMS], Rect* a_rect ) ELEMTYPE RTREE_QUAL::MinDist( const ELEMTYPE a_point[NUMDIMS], const Rect& a_rect )
{ {
ELEMTYPE *q, *s, *t; const ELEMTYPE *q, *s, *t;
q = (ELEMTYPE*) a_point; q = a_point;
s = a_rect->m_min; s = a_rect.m_min;
t = a_rect->m_max; t = a_rect.m_max;
int minDist = 0; double minDist = 0.0;
for( int index = 0; index < NUMDIMS; index++ ) for( int index = 0; index < NUMDIMS; index++ )
{ {
int r = q[index]; int r = q[index];
if( q[index] < s[index] ) if( q[index] < s[index] )
{ {
r = s[index]; r = s[index];
} }
else if( q[index] >t[index] ) else if( q[index] > t[index] )
{ {
r = t[index]; r = t[index];
} }
int addend = q[index] - r;
double addend = q[index] - r;
minDist += addend * addend; minDist += addend * addend;
} }
return minDist;
}
return std::lround( std::sqrt( minDist ) );
//insert a NNNode in a list sorted by its minDist (desc.)
RTREE_TEMPLATE
void RTREE_QUAL::InsertNNListSorted( std::vector<NNNode*>* nodeList, NNNode* newNode )
{
auto iter = nodeList->begin();
while( iter != nodeList->end() && (*iter)->minDist > newNode->minDist )
{
++iter;
}
nodeList->insert(iter, newNode);
} }