#include "testlib.h"

#include <iomanip>
#include <thread>
#include <tuple>
#include <unordered_map>
#include <vector>

using namespace std;

using ll = unsigned long long;
using pll = pair<ll, ll>;

const double eps = 1e-3;

void generate_edges_parallel(ll s_start, ll s_end, ll n, const vector<ll> &cnt,
                             unsigned int thread_seed,
                             vector<tuple<ll, ll, double>> &buffer) {
    random_t thread_rnd;
    thread_rnd.setSeed(thread_seed);

    unordered_map<ll, ll> selected;

    for (ll s = s_start; s < s_end; ++s) {
        ll k = cnt[s];
        if (k == 0) continue;

        selected.clear();

        for (ll i = 0; i < k; ++i) {
            ll range_size = (n - 1) - i;
            ll j = thread_rnd.next(1uLL, range_size);

            ll val_j = selected.count(j) ? selected[j] : j;
            ll val_last =
                selected.count(range_size) ? selected[range_size] : range_size;

            ll e = (val_j < s) ? val_j : val_j + 1;
            double len = thread_rnd.next(eps, 1.0 / eps);

            buffer.emplace_back(s, e, len);

            selected[j] = val_last;
        }
    }
}

int main(int argc, char *argv[]) {
    // 메인 RNG 초기화
    registerGen(argc, argv, 1);

    ll n = opt<ll>("N");
    ll m = opt<ll>("M");

    println(n, m);

    vector<ll> cnt(n + 1, 0);
    for (ll i = 0; i < m; ++i) {
        ll s = rnd.next(1uLL, n);
        cnt[s]++;
    }

    const int num_threads = thread::hardware_concurrency();
    vector<thread> threads;

    vector<unsigned int> thread_seeds;
    vector<vector<tuple<ll, ll, double>>> thread_buffers(num_threads);

    for (int i = 0; i < num_threads; ++i) {
        thread_seeds.push_back(rnd.next(1u, 1000000u));
    }

    ll s_batch_size = (n + num_threads) / num_threads;
    ll s_current = 1;

    for (int i = 0; i < num_threads; ++i) {
        ll s_start = s_current;
        ll s_end = min(s_start + s_batch_size, n + 1);
        if (s_start >= s_end) break;

        threads.emplace_back(generate_edges_parallel, s_start, s_end, n,
                             ref(cnt), thread_seeds[i], ref(thread_buffers[i]));

        s_current = s_end;
    }

    for (auto &t : threads) {
        t.join();
    }

    cout << fixed << setprecision(15);
    for (int i = 0; i < num_threads; ++i) {
        for (const auto &edge : thread_buffers[i]) {
            println(get<0>(edge), get<1>(edge), get<2>(edge));
        }
    }
}