This documentation is automatically generated by online-judge-tools/verification-helper
#define PROBLEM "https://judge.yosupo.jp/problem/sum_of_exponential_times_polynomial_limit"
#include "Template/template.hpp"
#include "Utility/fastio.hpp"
#include "Math/modint.hpp"
using Fp=fp<998244353>;
#include "Math/powertable.hpp"
#include "FPS/sumofpolyexp.hpp"
int main(){
Fp r;
int d;
read(r.v,d);
auto pws=powertable<Fp>(d+1,d);
auto ret=LimitSumOfPolyExp(pws,r);
print(ret.v);
return 0;
}
#line 1 "Verify/LC_sum_of_exponential_times_polynomial_limit.test.cpp"
#define PROBLEM "https://judge.yosupo.jp/problem/sum_of_exponential_times_polynomial_limit"
#line 1 "Template/template.hpp"
#include <bits/stdc++.h>
using namespace std;
#define rep(i, a, b) for (int i = (int)(a); i < (int)(b); i++)
#define rrep(i, a, b) for (int i = (int)(b)-1; i >= (int)(a); i--)
#define ALL(v) (v).begin(), (v).end()
#define UNIQUE(v) sort(ALL(v)), (v).erase(unique(ALL(v)), (v).end())
#define SZ(v) (int)v.size()
#define MIN(v) *min_element(ALL(v))
#define MAX(v) *max_element(ALL(v))
#define LB(v, x) int(lower_bound(ALL(v), (x)) - (v).begin())
#define UB(v, x) int(upper_bound(ALL(v), (x)) - (v).begin())
using uint = unsigned int;
using ll = long long int;
using ull = unsigned long long;
using i128 = __int128_t;
using u128 = __uint128_t;
const int inf = 0x3fffffff;
const ll INF = 0x1fffffffffffffff;
template <typename T, typename S = T> S SUM(const vector<T> &a) {
return accumulate(ALL(a), S(0));
}
template <typename S, typename T = S> S POW(S a, T b) {
S ret = 1, base = a;
for (;;) {
if (b & 1)
ret *= base;
b >>= 1;
if (b == 0)
break;
base *= base;
}
return ret;
}
template <typename T> inline bool chmax(T &a, T b) {
if (a < b) {
a = b;
return 1;
}
return 0;
}
template <typename T> inline bool chmin(T &a, T b) {
if (a > b) {
a = b;
return 1;
}
return 0;
}
template <typename T, typename U> T ceil(T x, U y) {
assert(y != 0);
if (y < 0)
x = -x, y = -y;
return (x > 0 ? (x + y - 1) / y : x / y);
}
template <typename T, typename U> T floor(T x, U y) {
assert(y != 0);
if (y < 0)
x = -x, y = -y;
return (x > 0 ? x / y : (x - y + 1) / y);
}
template <typename T> int popcnt(T x) {
return __builtin_popcountll(x);
}
template <typename T> int topbit(T x) {
return (x == 0 ? -1 : 63 - __builtin_clzll(x));
}
template <typename T> int lowbit(T x) {
return (x == 0 ? -1 : __builtin_ctzll(x));
}
template <class T, class U>
ostream &operator<<(ostream &os, const pair<T, U> &p) {
os << "P(" << p.first << ", " << p.second << ")";
return os;
}
template <typename T> ostream &operator<<(ostream &os, const vector<T> &vec) {
os << "{";
for (int i = 0; i < vec.size(); i++) {
os << vec[i] << (i + 1 == vec.size() ? "" : ", ");
}
os << "}";
return os;
}
template <typename T, typename U>
ostream &operator<<(ostream &os, const map<T, U> &map_var) {
os << "{";
for (auto itr = map_var.begin(); itr != map_var.end(); itr++) {
os << "(" << itr->first << ", " << itr->second << ")";
itr++;
if (itr != map_var.end())
os << ", ";
itr--;
}
os << "}";
return os;
}
template <typename T> ostream &operator<<(ostream &os, const set<T> &set_var) {
os << "{";
for (auto itr = set_var.begin(); itr != set_var.end(); itr++) {
os << *itr;
++itr;
if (itr != set_var.end())
os << ", ";
itr--;
}
os << "}";
return os;
}
#ifdef LOCAL
#define debug 1
#define show(...) _show(0, #__VA_ARGS__, __VA_ARGS__)
#else
#define debug 0
#define show(...) true
#endif
template <typename T> void _show(int i, T name) {
cerr << '\n';
}
template <typename T1, typename T2, typename... T3>
void _show(int i, const T1 &a, const T2 &b, const T3 &...c) {
for (; a[i] != ',' && a[i] != '\0'; i++)
cerr << a[i];
cerr << ":" << b << " ";
_show(i + 1, a, c...);
}
#line 2 "Utility/fastio.hpp"
#include <unistd.h>
namespace fastio {
static constexpr uint32_t SZ = 1 << 17;
char ibuf[SZ];
char obuf[SZ];
char out[100];
// pointer of ibuf, obuf
uint32_t pil = 0, pir = 0, por = 0;
struct Pre {
char num[10000][4];
constexpr Pre() : num() {
for (int i = 0; i < 10000; i++) {
int n = i;
for (int j = 3; j >= 0; j--) {
num[i][j] = n % 10 | '0';
n /= 10;
}
}
}
} constexpr pre;
inline void load() {
memmove(ibuf, ibuf + pil, pir - pil);
pir = pir - pil + fread(ibuf + pir - pil, 1, SZ - pir + pil, stdin);
pil = 0;
if (pir < SZ)
ibuf[pir++] = '\n';
}
inline void flush() {
fwrite(obuf, 1, por, stdout);
por = 0;
}
void rd(char &c) {
do {
if (pil + 1 > pir)
load();
c = ibuf[pil++];
} while (isspace(c));
}
void rd(string &x) {
x.clear();
char c;
do {
if (pil + 1 > pir)
load();
c = ibuf[pil++];
} while (isspace(c));
do {
x += c;
if (pil == pir)
load();
c = ibuf[pil++];
} while (!isspace(c));
}
template <typename T> void rd_real(T &x) {
string s;
rd(s);
x = stod(s);
}
template <typename T> void rd_integer(T &x) {
if (pil + 100 > pir)
load();
char c;
do
c = ibuf[pil++];
while (c < '-');
bool minus = 0;
if constexpr (is_signed<T>::value || is_same_v<T, i128>) {
if (c == '-') {
minus = 1, c = ibuf[pil++];
}
}
x = 0;
while ('0' <= c) {
x = x * 10 + (c & 15), c = ibuf[pil++];
}
if constexpr (is_signed<T>::value || is_same_v<T, i128>) {
if (minus)
x = -x;
}
}
void rd(int &x) {
rd_integer(x);
}
void rd(ll &x) {
rd_integer(x);
}
void rd(i128 &x) {
rd_integer(x);
}
void rd(uint &x) {
rd_integer(x);
}
void rd(ull &x) {
rd_integer(x);
}
void rd(u128 &x) {
rd_integer(x);
}
void rd(double &x) {
rd_real(x);
}
void rd(long double &x) {
rd_real(x);
}
template <class T, class U> void rd(pair<T, U> &p) {
return rd(p.first), rd(p.second);
}
template <size_t N = 0, typename T> void rd_tuple(T &t) {
if constexpr (N < std::tuple_size<T>::value) {
auto &x = std::get<N>(t);
rd(x);
rd_tuple<N + 1>(t);
}
}
template <class... T> void rd(tuple<T...> &tpl) {
rd_tuple(tpl);
}
template <size_t N = 0, typename T> void rd(array<T, N> &x) {
for (auto &d : x)
rd(d);
}
template <class T> void rd(vector<T> &x) {
for (auto &d : x)
rd(d);
}
void read() {}
template <class H, class... T> void read(H &h, T &...t) {
rd(h), read(t...);
}
void wt(const char c) {
if (por == SZ)
flush();
obuf[por++] = c;
}
void wt(const string s) {
for (char c : s)
wt(c);
}
void wt(const char *s) {
size_t len = strlen(s);
for (size_t i = 0; i < len; i++)
wt(s[i]);
}
template <typename T> void wt_integer(T x) {
if (por > SZ - 100)
flush();
if (x < 0) {
obuf[por++] = '-', x = -x;
}
int outi;
for (outi = 96; x >= 10000; outi -= 4) {
memcpy(out + outi, pre.num[x % 10000], 4);
x /= 10000;
}
if (x >= 1000) {
memcpy(obuf + por, pre.num[x], 4);
por += 4;
} else if (x >= 100) {
memcpy(obuf + por, pre.num[x] + 1, 3);
por += 3;
} else if (x >= 10) {
int q = (x * 103) >> 10;
obuf[por] = q | '0';
obuf[por + 1] = (x - q * 10) | '0';
por += 2;
} else
obuf[por++] = x | '0';
memcpy(obuf + por, out + outi + 4, 96 - outi);
por += 96 - outi;
}
template <typename T> void wt_real(T x) {
ostringstream oss;
oss << fixed << setprecision(15) << double(x);
string s = oss.str();
wt(s);
}
void wt(int x) {
wt_integer(x);
}
void wt(ll x) {
wt_integer(x);
}
void wt(i128 x) {
wt_integer(x);
}
void wt(uint x) {
wt_integer(x);
}
void wt(ull x) {
wt_integer(x);
}
void wt(u128 x) {
wt_integer(x);
}
void wt(double x) {
wt_real(x);
}
void wt(long double x) {
wt_real(x);
}
template <class T, class U> void wt(const pair<T, U> val) {
wt(val.first);
wt(' ');
wt(val.second);
}
template <size_t N = 0, typename T> void wt_tuple(const T t) {
if constexpr (N < std::tuple_size<T>::value) {
if constexpr (N > 0) {
wt(' ');
}
const auto x = std::get<N>(t);
wt(x);
wt_tuple<N + 1>(t);
}
}
template <class... T> void wt(tuple<T...> tpl) {
wt_tuple(tpl);
}
template <class T, size_t S> void wt(const array<T, S> val) {
auto n = val.size();
for (size_t i = 0; i < n; i++) {
if (i)
wt(' ');
wt(val[i]);
}
}
template <class T> void wt(const vector<T> val) {
auto n = val.size();
for (size_t i = 0; i < n; i++) {
if (i)
wt(' ');
wt(val[i]);
}
}
void print() {
wt('\n');
}
template <class Head, class... Tail> void print(Head &&head, Tail &&...tail) {
wt(head);
if (sizeof...(Tail))
wt(' ');
print(forward<Tail>(tail)...);
}
void __attribute__((destructor)) _d() {
flush();
}
} // namespace fastio
using fastio::flush;
using fastio::print;
using fastio::read;
inline void first(bool i = true) {
print(i ? "first" : "second");
}
inline void Alice(bool i = true) {
print(i ? "Alice" : "Bob");
}
inline void Takahashi(bool i = true) {
print(i ? "Takahashi" : "Aoki");
}
inline void yes(bool i = true) {
print(i ? "yes" : "no");
}
inline void Yes(bool i = true) {
print(i ? "Yes" : "No");
}
inline void No() {
print("No");
}
inline void YES(bool i = true) {
print(i ? "YES" : "NO");
}
inline void NO() {
print("NO");
}
inline void Yay(bool i = true) {
print(i ? "Yay!" : ":(");
}
inline void Possible(bool i = true) {
print(i ? "Possible" : "Impossible");
}
inline void POSSIBLE(bool i = true) {
print(i ? "POSSIBLE" : "IMPOSSIBLE");
}
/**
* @brief Fast IO
*/
#line 5 "Verify/LC_sum_of_exponential_times_polynomial_limit.test.cpp"
#line 2 "Math/comb.hpp"
template <typename T> T Inv(ll n) {
static int md;
static vector<T> buf({0, 1});
if (md != T::get_mod()) {
md = T::get_mod();
buf = vector<T>({0, 1});
}
assert(n > 0);
n %= md;
while (SZ(buf) <= n) {
int k = SZ(buf), q = (md + k - 1) / k;
buf.push_back(buf[k * q - md] * q);
}
return buf[n];
}
template <typename T> T Fact(ll n, bool inv = 0) {
static int md;
static vector<T> buf({1, 1}), ibuf({1, 1});
if (md != T::get_mod()) {
md = T::get_mod();
buf = ibuf = vector<T>({1, 1});
}
assert(n >= 0 and n < md);
while (SZ(buf) <= n) {
buf.push_back(buf.back() * SZ(buf));
ibuf.push_back(ibuf.back() * Inv<T>(SZ(ibuf)));
}
return inv ? ibuf[n] : buf[n];
}
template <typename T> T nPr(int n, int r, bool inv = 0) {
if (n < 0 || n < r || r < 0)
return 0;
return Fact<T>(n, inv) * Fact<T>(n - r, inv ^ 1);
}
template <typename T> T nCr(int n, int r, bool inv = 0) {
if (n < 0 || n < r || r < 0)
return 0;
return Fact<T>(n, inv) * Fact<T>(r, inv ^ 1) * Fact<T>(n - r, inv ^ 1);
}
// sum = n, r tuples
template <typename T> T nHr(int n, int r, bool inv = 0) {
return nCr<T>(n + r - 1, r - 1, inv);
}
// sum = n, a nonzero tuples and b tuples
template <typename T> T choose(int n, int a, int b) {
if (n == 0)
return !a;
return nCr<T>(n + b - 1, a + b - 1);
}
/**
* @brief Combination
*/
#line 3 "Math/modint.hpp"
template <unsigned mod = 1000000007> struct fp {
static_assert(mod < uint(1) << 31);
unsigned v;
static constexpr int get_mod() {
return mod;
}
constexpr unsigned inv() const {
assert(v != 0);
int x = v, y = mod, p = 1, q = 0, t = 0, tmp = 0;
while (y > 0) {
t = x / y;
x -= t * y, p -= t * q;
tmp = x, x = y, y = tmp;
tmp = p, p = q, q = tmp;
}
if (p < 0)
p += mod;
return p;
}
constexpr fp(ll x = 0) : v(x >= 0 ? x % mod : (mod - (-x) % mod) % mod) {}
fp operator-() const {
return fp() - *this;
}
fp pow(ull t) {
fp res = 1, b = *this;
while (t) {
if (t & 1)
res *= b;
b *= b;
t >>= 1;
}
return res;
}
fp &operator+=(const fp &x) {
if ((v += x.v) >= mod)
v -= mod;
return *this;
}
fp &operator-=(const fp &x) {
if ((v += mod - x.v) >= mod)
v -= mod;
return *this;
}
fp &operator*=(const fp &x) {
v = ull(v) * x.v % mod;
return *this;
}
fp &operator/=(const fp &x) {
if (x.v < 15000000) {
return *this *= Inv<fp>(x.v);
}
v = ull(v) * x.inv() % mod;
return *this;
}
fp operator+(const fp &x) const {
return fp(*this) += x;
}
fp operator-(const fp &x) const {
return fp(*this) -= x;
}
fp operator*(const fp &x) const {
return fp(*this) *= x;
}
fp operator/(const fp &x) const {
return fp(*this) /= x;
}
bool operator==(const fp &x) const {
return v == x.v;
}
bool operator!=(const fp &x) const {
return v != x.v;
}
friend istream &operator>>(istream &is, fp &x) {
return is >> x.v;
}
friend ostream &operator<<(ostream &os, const fp &x) {
return os << x.v;
}
};
template <unsigned mod> void rd(fp<mod> &x) {
fastio::rd(x.v);
}
template <unsigned mod> void wt(fp<mod> x) {
fastio::wt(x.v);
}
/**
* @brief Modint
*/
#line 7 "Verify/LC_sum_of_exponential_times_polynomial_limit.test.cpp"
using Fp=fp<998244353>;
#line 2 "Math/sieve.hpp"
template <int L = 101010101> vector<int> sieve(int N) {
bitset<L> isp;
int n, sq = ceil(sqrt(N));
for (int z = 1; z <= 5; z += 4) {
for (int y = z; y <= sq; y += 6) {
for (int x = 1; x <= sq and (n = 4 * x * x + y * y) <= N; ++x) {
isp[n].flip();
}
for (int x = y + 1; x <= sq and (n = 3 * x * x - y * y) <= N;
x += 2) {
isp[n].flip();
}
}
}
for (int z = 2; z <= 4; z += 2) {
for (int y = z; y <= sq; y += 6) {
for (int x = 1; x <= sq and (n = 3 * x * x + y * y) <= N; x += 2) {
isp[n].flip();
}
for (int x = y + 1; x <= sq and (n = 3 * x * x - y * y) <= N;
x += 2) {
isp[n].flip();
}
}
}
for (int y = 3; y <= sq; y += 6) {
for (int z = 1; z <= 2; ++z) {
for (int x = z; x <= sq and (n = 4 * x * x + y * y) <= N; x += 3) {
isp[n].flip();
}
}
}
for (int n = 5; n <= sq; ++n)
if (isp[n]) {
for (int k = n * n; k <= N; k += n * n) {
isp[k] = false;
}
}
isp[2] = isp[3] = true;
vector<int> ret;
for (int i = 2; i <= N; i++)
if (isp[i]) {
ret.push_back(i);
}
return ret;
}
/**
* @brief Prime Sieve
*/
#line 3 "Math/powertable.hpp"
template<typename T>vector<T> powertable(int n,ll k){ //0^k,1^k,..,n^k
assert(k>=0);
auto ps=sieve(n+1);
vector<T> f(n+1,1);
if(k)f[0]=0;
for(auto& p:ps){
T pk=T(p).pow(k);
for(ll q=p;q<=n;q*=p){
for(ll i=q;i<=n;i+=q)f[i]*=pk;
}
}
return f;
}
/**
* @brief Enumrate $n^k$
*/
#line 3 "FPS/interpolate.hpp"
// f(0),..,f(d) -> f(t)
template <typename T> T Interpolate(vector<T> &ys, ll t) {
int d = ys.size() - 1;
if (t <= d)
return ys[t];
vector<T> L(d + 1, 1), R(d + 1, 1);
rep(i, 0, d) L[i + 1] = L[i] * (t - i);
for (int i = d; i; i--)
R[i - 1] = R[i] * (t - i);
T ret;
rep(i, 0, d + 1) {
T add = ys[i] * L[i] * R[i] * Fact<T>(i, 1) * Fact<T>(d - i, 1);
if ((d - i) & 1)
ret -= add;
else
ret += add;
}
return ret;
}
/**
* @brief interpolate (one point)
*/
#line 4 "FPS/sumofpolyexp.hpp"
template <typename T>
T LimitSumOfPolyExp(vector<T> &f, T r) { // sum_{k=0}^inf r^k*f(k)
assert(r != 1);
int d = f.size() - 1;
vector<T> rs(d + 1);
rs[0] = 1;
rep(i, 0, d) rs[i + 1] = rs[i] * r;
T c, add;
rep(i, 0, d + 1) {
add += rs[i] * f[i];
if ((d - i) & 1)
c -= nCr<T>(d + 1, i + 1) * rs[d - i] * add;
else
c += nCr<T>(d + 1, i + 1) * rs[d - i] * add;
}
c /= (-r + 1).pow(d + 1);
return c;
}
template <typename T>
T SumOfPolyExp(vector<T> &f, T r, ll n) { // sum_{k=0}^{n-1} r^k*f(k)
n--;
if (n < 0)
return 0;
int d = f.size() - 1;
vector<T> rs(d + 1), rui(d + 1);
rs[0] = 1;
rep(i, 0, d) rs[i + 1] = rs[i] * r;
rep(i, 0, d + 1) rui[i] = rs[i] * f[i];
rep(i, 0, d) rui[i + 1] += rui[i];
if (r == 0)
return f[0];
else if (r == 1)
return Interpolate(rui, n);
else {
T c;
rep(i, 0, d + 1) c +=
nCr<T>(d + 1, i + 1) * rs[d - i] * rui[i] * ((d - i) & 1 ? -1 : 1);
c /= T(-r + 1).pow(d + 1);
vector<T> ys(d + 1);
T pwr = 1, invr = T(r).inv();
rep(i, 0, d + 1) ys[i] = (rui[i] - c) * pwr, pwr *= invr;
return T(r).pow(n) * Interpolate(ys, n) + c;
}
}
/**
* @brief $\sum_{k} r^k\cdot poly(k)$
*/
#line 11 "Verify/LC_sum_of_exponential_times_polynomial_limit.test.cpp"
int main(){
Fp r;
int d;
read(r.v,d);
auto pws=powertable<Fp>(d+1,d);
auto ret=LimitSumOfPolyExp(pws,r);
print(ret.v);
return 0;
}