This documentation is automatically generated by online-judge-tools/verification-helper
#include "geometry/convex-hull.hpp"
#include <bits/stdc++.h>
#include "./point.hpp"
using namespace std;
/*
https://oi-wiki.org/geometry/convex-hull/
https://judge.yosupo.jp/problem/line_add_get_min
*/
namespace geometry {
template <typename T>
T floor_div(T n, T d) {
return n / d - ((n ^ d) < 0 & n % d != 0);
}
template <typename T>
bool is_right_of_vector(const TPoint<T> &a, const TPoint<T> &b, const TPoint<T> &c, bool boundary_contain) {
/*
boundary_contain = true 表示包含边界上的点
*/
auto cb = b - c;
auto ba = a - b;
// cb * ba <= 0 判断 ba 是不是在 cb 的右侧
// 直接叉积
T tmp = vmul(cb, ba);
return tmp < 0 || (!boundary_contain && tmp == 0);
}
template <typename T>
using Polygon = vector<TPoint<T>>;
template <typename T>
pair<Polygon<T>, Polygon<T>> convex_hull_by_andrew(vector<TPoint<T>> &points, bool boundary_contain = false) {
sort(points.begin(), points.end(), [&](TPoint<T> &a, TPoint<T> &b) {
if (a.x != b.x) return a.x < b.x;
return a.y < b.y;
});
vector<int> stk;
stk.push_back(0);
int n = points.size();
vector<int> used(n);
for (int i = 1; i < n; i++) {
while (stk.size() >= 2 && is_right_of_vector(points[i], points[stk.back()], points[stk[stk.size() - 2]], boundary_contain)) {
used[stk.back()] = 0;
stk.pop_back();
}
used[i] = 1;
stk.push_back(i);
}
Polygon<T> lower_convex_hull, upper_convex_hull;
int tmp = stk.size(); // 下凸包的大小
for (int i = n - 1; i >= 0; i--) {
if (!used[i]) {
while (stk.size() > tmp && is_right_of_vector(points[i], points[stk.back()], points[stk[stk.size() - 2]], boundary_contain)) {
used[stk.back()] = 0;
stk.pop_back();
}
used[i] = 1;
stk.push_back(i);
}
}
for (int i = 0; i < tmp; i++) {
lower_convex_hull.emplace_back(points[stk[i]]);
}
for (int i = tmp - 1; i < stk.size(); i++) {
upper_convex_hull.emplace_back(points[stk[i]]);
}
return {lower_convex_hull, upper_convex_hull};
}
template <typename T>
Polygon<T> convex_hull_by_graham(vector<TPoint<T>> &points, bool boundary_contain = false) {
int n = points.size();
if (n == 1) {
return points;
}
// 找到 y 最小的,x 最小的点
sort(points.begin(), points.end());
auto p0 = points[0];
for (int i = 0; i < n; i++) {
points[i] = points[i] - p0;
}
// 极角排序
sort(points.begin(), points.end(), compare_by_polar_angle<T>);
if (boundary_contain) {
int sz = n - 1;
while (sz >= 0 && vmul(points[n - 1] - points[0], points[sz] - points[0]) == 0) sz--;
for (int l = sz + 1, r = n - 1; l < r; l++, r--) {
swap(points[l], points[r]);
}
}
vector<int> stk(n);
int k = 0;
stk[k++] = 0;
for (int i = 1; i < n; i++) {
while (k >= 2 && is_right_of_vector(points[i], points[stk[k - 1]], points[stk[k - 2]], boundary_contain)) {
k--;
}
stk[k++] = i;
}
Polygon<T> convex;
for (int i = 0; i < k; i++) {
convex.push_back(points[stk[i]] + p0);
}
return convex;
}
template <typename T>
T getMinDotProduct(const Polygon<T> convex, const TPoint<T> &p) {
/*
求凸包 convex 上的点的向量 x 与向量 p 的点积的最小值, 点积的几何意义是 x 在 p 上的投影的积,方向相同为正,方向不同为负.
可以发现当 p.y 是正的,那么最小值一定出现在凸包的下半部分。如果 p.y 是负的,那么最小值出现在凸包的上半部分.(感觉很正确)
*/
if (p.y < 0) {
// 翻转
}
}
} // namespace geometry
#line 1 "geometry/convex-hull.hpp"
#include <bits/stdc++.h>
#line 2 "geometry/point.hpp"
#line 3 "geometry/base.hpp"
using namespace std;
namespace geometry {
template <typename T>
const constexpr T EPS = static_cast<T>(1e-8);
template <typename T>
const constexpr T PI = acos(static_cast<T>(-1));
template <typename T>
inline int sign(const T &r) {
return r < -EPS<T> ? -1 : r > EPS<T> ? 1
: 0;
}
template <typename T>
inline bool equals(const T &a, const T &b) { return sign(a - b) == 0; }
} // namespace geometry
#line 4 "geometry/point.hpp"
using namespace std;
namespace geometry {
template <typename T>
struct TPoint {
T x, y;
int id;
TPoint() : x(0), y(0), id(-1) {}
TPoint(const T& x_, const T& y_) : x(x_), y(y_), id(-1) {}
TPoint(const T& x_, const T& y_, int id_) : x(x_), y(y_), id(id_) {}
static constexpr T eps = static_cast<T>(1e-9);
inline TPoint operator+(const TPoint& rhs) const { return TPoint(x + rhs.x, y + rhs.y); }
inline TPoint operator-(const TPoint& rhs) const { return TPoint(x - rhs.x, y - rhs.y); }
inline TPoint operator*(const T& rhs) const { return TPoint(x * rhs, y * rhs); }
inline TPoint operator/(const T& rhs) const { return TPoint(x / rhs, y / rhs); }
friend T smul(const TPoint& a, const TPoint& b) {
// 点积
return a.x * b.x + a.y * b.y;
}
friend T vmul(const TPoint& a, const TPoint& b) {
// 叉积
return a.x * b.y - a.y * b.x;
}
inline T abs2() const {
return x * x + y * y;
}
inline bool operator<(const TPoint& rhs) const {
return (y < rhs.y || (y == rhs.y && x < rhs.x));
}
};
template <typename T>
string to_string(const TPoint<T>& p) {
return "(" + std::to_string(p.x) + ", " + std::to_string(p.y) + ")";
}
template <typename T>
bool compare_by_polar_angle(const TPoint<T>& a, const TPoint<T>& b) {
T x = vmul(a, b);
return x == 0 ? (a.abs2() < b.abs2()) : x > 0;
}
} // namespace geometry
#line 5 "geometry/convex-hull.hpp"
using namespace std;
/*
https://oi-wiki.org/geometry/convex-hull/
https://judge.yosupo.jp/problem/line_add_get_min
*/
namespace geometry {
template <typename T>
T floor_div(T n, T d) {
return n / d - ((n ^ d) < 0 & n % d != 0);
}
template <typename T>
bool is_right_of_vector(const TPoint<T> &a, const TPoint<T> &b, const TPoint<T> &c, bool boundary_contain) {
/*
boundary_contain = true 表示包含边界上的点
*/
auto cb = b - c;
auto ba = a - b;
// cb * ba <= 0 判断 ba 是不是在 cb 的右侧
// 直接叉积
T tmp = vmul(cb, ba);
return tmp < 0 || (!boundary_contain && tmp == 0);
}
template <typename T>
using Polygon = vector<TPoint<T>>;
template <typename T>
pair<Polygon<T>, Polygon<T>> convex_hull_by_andrew(vector<TPoint<T>> &points, bool boundary_contain = false) {
sort(points.begin(), points.end(), [&](TPoint<T> &a, TPoint<T> &b) {
if (a.x != b.x) return a.x < b.x;
return a.y < b.y;
});
vector<int> stk;
stk.push_back(0);
int n = points.size();
vector<int> used(n);
for (int i = 1; i < n; i++) {
while (stk.size() >= 2 && is_right_of_vector(points[i], points[stk.back()], points[stk[stk.size() - 2]], boundary_contain)) {
used[stk.back()] = 0;
stk.pop_back();
}
used[i] = 1;
stk.push_back(i);
}
Polygon<T> lower_convex_hull, upper_convex_hull;
int tmp = stk.size(); // 下凸包的大小
for (int i = n - 1; i >= 0; i--) {
if (!used[i]) {
while (stk.size() > tmp && is_right_of_vector(points[i], points[stk.back()], points[stk[stk.size() - 2]], boundary_contain)) {
used[stk.back()] = 0;
stk.pop_back();
}
used[i] = 1;
stk.push_back(i);
}
}
for (int i = 0; i < tmp; i++) {
lower_convex_hull.emplace_back(points[stk[i]]);
}
for (int i = tmp - 1; i < stk.size(); i++) {
upper_convex_hull.emplace_back(points[stk[i]]);
}
return {lower_convex_hull, upper_convex_hull};
}
template <typename T>
Polygon<T> convex_hull_by_graham(vector<TPoint<T>> &points, bool boundary_contain = false) {
int n = points.size();
if (n == 1) {
return points;
}
// 找到 y 最小的,x 最小的点
sort(points.begin(), points.end());
auto p0 = points[0];
for (int i = 0; i < n; i++) {
points[i] = points[i] - p0;
}
// 极角排序
sort(points.begin(), points.end(), compare_by_polar_angle<T>);
if (boundary_contain) {
int sz = n - 1;
while (sz >= 0 && vmul(points[n - 1] - points[0], points[sz] - points[0]) == 0) sz--;
for (int l = sz + 1, r = n - 1; l < r; l++, r--) {
swap(points[l], points[r]);
}
}
vector<int> stk(n);
int k = 0;
stk[k++] = 0;
for (int i = 1; i < n; i++) {
while (k >= 2 && is_right_of_vector(points[i], points[stk[k - 1]], points[stk[k - 2]], boundary_contain)) {
k--;
}
stk[k++] = i;
}
Polygon<T> convex;
for (int i = 0; i < k; i++) {
convex.push_back(points[stk[i]] + p0);
}
return convex;
}
template <typename T>
T getMinDotProduct(const Polygon<T> convex, const TPoint<T> &p) {
/*
求凸包 convex 上的点的向量 x 与向量 p 的点积的最小值, 点积的几何意义是 x 在 p 上的投影的积,方向相同为正,方向不同为负.
可以发现当 p.y 是正的,那么最小值一定出现在凸包的下半部分。如果 p.y 是负的,那么最小值出现在凸包的上半部分.(感觉很正确)
*/
if (p.y < 0) {
// 翻转
}
}
} // namespace geometry