3D-printing/OrcaSlicer
1
0
Fork 0
mirror of https://github.com/SoftFever/OrcaSlicer.git synced 2025-07-22 14:13:57 -06:00
Code Issues Projects Releases Packages Wiki Activity Actions 2

Separate support tree routing and meshing, remove Common.hpp/.cpp .

Browse source 
* Remove Common.hpp and Common.cpp, move things into their respective modules in sla.
This commit is contained in:
tamasmeszaros 2020-06-16 13:17:06 +02:00
parent 2ff04e6f68
commit 184f64f828
27 changed files with 897 additions and 912 deletions
Download patch file Download diff file Expand all files Collapse all files

50
src/libslic3r/SLA/SupportTreeBuildsteps.hpp
View file

@ -5,6 +5,7 @@
#include <libslic3r/SLA/SupportTreeBuilder.hpp>
#include <libslic3r/SLA/Clustering.hpp>
#include <libslic3r/SLA/SpatIndex.hpp>
namespace Slic3r {
namespace sla {
@ -108,55 +109,6 @@ public:
EigenMesh3D::hit_result query_hit(const SupportableMesh &msh, const Bridge &br, double safety_d = std::nan(""));
EigenMesh3D::hit_result query_hit(const SupportableMesh &msh, const Head &br, double safety_d = std::nan(""));
// This function returns the position of the centroid in the input 'clust'
// vector of point indices.
template<class DistFn>
long cluster_centroid(const ClusterEl& clust,
const std::function<Vec3d(size_t)> &pointfn,
DistFn df)
{
switch(clust.size()) {
case 0: /* empty cluster */ return ID_UNSET;
case 1: /* only one element */ return 0;
case 2: /* if two elements, there is no center */ return 0;
default: ;
}
// The function works by calculating for each point the average distance
// from all the other points in the cluster. We create a selector bitmask of
// the same size as the cluster. The bitmask will have two true bits and
// false bits for the rest of items and we will loop through all the
// permutations of the bitmask (combinations of two points). Get the
// distance for the two points and add the distance to the averages.
// The point with the smallest average than wins.
// The complexity should be O(n^2) but we will mostly apply this function
// for small clusters only (cca 3 elements)
std::vector<bool> sel(clust.size(), false); // create full zero bitmask
std::fill(sel.end() - 2, sel.end(), true); // insert the two ones
std::vector<double> avgs(clust.size(), 0.0); // store the average distances
do {
std::array<size_t, 2> idx;
for(size_t i = 0, j = 0; i < clust.size(); i++) if(sel[i]) idx[j++] = i;
double d = df(pointfn(clust[idx[0]]),
pointfn(clust[idx[1]]));
// add the distance to the sums for both associated points
for(auto i : idx) avgs[i] += d;
// now continue with the next permutation of the bitmask with two 1s
} while(std::next_permutation(sel.begin(), sel.end()));
// Divide by point size in the cluster to get the average (may be redundant)
for(auto& a : avgs) a /= clust.size();
// get the lowest average distance and return the index
auto minit = std::min_element(avgs.begin(), avgs.end());
return long(minit - avgs.begin());
}
inline Vec3d dirv(const Vec3d& startp, const Vec3d& endp) {
return (endp - startp).normalized();