/*
 * Copyright (c) 2016, NVIDIA CORPORATION.  All rights reserved.
 *
 * Please refer to the NVIDIA end user license agreement (EULA) associated
 * with this source code for terms and conditions that govern your use of
 * this software. Any use, reproduction, disclosure, or distribution of
 * this software and related documentation outside the terms of the EULA
 * is strictly prohibited.
 *
 */

#include <stdio.h>
#include <stdlib.h>
#include <cuda_runtime.h>
#include "helper_cuda.h"
#include "nvgraph.h"

#include<iostream>
#include<string>
#include<iomanip>
#include<fstream>
#include<cstdlib>
#include<bits/stdc++.h>

using namespace std;


/* Single Source Shortest Path (SSSP)
 *  Calculate the shortest path distance from a single vertex in the graph
 *  to all other vertices.
 */

struct tmp_edge{
    int src_id;
    int dst_id;
    float weight;
};

int compare_edges(const void * a, const void * b)
{
    return ( (*(struct tmp_edge *)a).dst_id - (*(struct tmp_edge *)b).dst_id );
}

void check_status(nvgraphStatus_t status)
{
    if ((int)status != 0)
    {
        printf("ERROR : %d\n",status);
        exit(0);
    }
}

int main(int argc, char **argv)
{

    const char * file_name = "./graph.bin";

        // open file
    fstream file(file_name, ios::in | ios::binary);

    int vertices_count = 0;
    long long edges_count = 0;

    // read header
    file.read((char*)(&vertices_count), sizeof(int));
    file.read((char*)(&edges_count), sizeof(long long));

    // print graph
    cout << "Graph has " << vertices_count << " vertices" << endl;
    cout << "Graph has " << edges_count << " edges" << endl;

    int n = vertices_count;
    int nnz = edges_count;//because nvgraphCSCTopology32I_st structure can take only int as 3 parameter, and there is no analog for 64-bit

    const size_t  vertex_numsets = 1, edge_numsets = 1;
    int *destination_offsets_h;
    int *source_indices_h;
    float *weights_h, *sssp_h;
    void** vertex_dim;

    // nvgraph variables
    nvgraphStatus_t status;
    nvgraphHandle_t handle;
    nvgraphGraphDescr_t graph;
    nvgraphCSCTopology32I_t CSC_input;
    cudaDataType_t edge_dimT = CUDA_R_32F;
    cudaDataType_t vertex_dimT = CUDA_R_32F;

    // use command-line specified CUDA device, otherwise use device with highest Gflops/s
    int cuda_device = 0;
    cuda_device = findCudaDevice(argc, (const char **)argv);

    cudaDeviceProp deviceProp;
    checkCudaErrors(cudaGetDevice(&cuda_device));

    checkCudaErrors(cudaGetDeviceProperties(&deviceProp, cuda_device));

    printf("> Detected Compute SM %d.%d hardware with %d multi-processors\n",
           deviceProp.major, deviceProp.minor, deviceProp.multiProcessorCount);

    if (deviceProp.major < 3)
    {
        printf("> nvGraph requires device SM 3.0+\n");
        printf("> Waiving.\n");
        exit(EXIT_WAIVED);
    }

    // Init host data
    destination_offsets_h = (int*) malloc((n+1)*sizeof(int));
    source_indices_h = (int*) malloc(nnz*sizeof(int));
    weights_h = (float*)malloc(nnz*sizeof(float));
    sssp_h = (float*)malloc(n*sizeof(float));
    CSC_input = (nvgraphCSCTopology32I_t) malloc(sizeof(struct nvgraphCSCTopology32I_st));

    for (int i = 0; i < n; i++){
        destination_offsets_h [i] = INT_MAX;
    }



    struct tmp_edge * Edges = (tmp_edge*)malloc(nnz*sizeof(tmp_edge));

    // get & print graph data for WEIGHTED graph
    for(int i = 0; i < nnz; i++){
        int src_id = 0, dst_id = 0;
        float weight = 0;

        // read i-th edge data
        file.read((char*)(&src_id), sizeof(int));
        file.read((char*)(&dst_id), sizeof(int));
        file.read((char*)(&weight), sizeof(float)); // remove it for unweighed graph

        //print edge data
        //cout << i << "  " << src_id << " " << dst_id << " | " << weight << endl;

        Edges[i] = {src_id, dst_id, weight};

        //cout << i << " " <<  Edges[i].src_id << " " << Edges[i].dst_id << " " << Edges[i].weight << endl;
    }

    qsort(Edges, nnz, sizeof(tmp_edge), compare_edges);

    int prev_dst_id = -1;

    for(int i = 0; i < nnz; i++){
        weights_h [i] = Edges[i].weight;

        int tmp_dst_id = Edges[i].dst_id;

        if (tmp_dst_id > prev_dst_id){


            for (int j = prev_dst_id + 1; j <= tmp_dst_id; j++){

                destination_offsets_h [j] = i;

            }

            prev_dst_id = tmp_dst_id;

        }

        source_indices_h [i] = Edges[i].src_id;
        //cout << i << " " <<  Edges[i].src_id << " " << Edges[i].dst_id << " " << Edges[i].weight << endl;
    }

    for (int j = prev_dst_id + 1; j <= n; j++){

        destination_offsets_h [j] = nnz;

    }

    check_status(nvgraphCreate(&handle));
    check_status(nvgraphCreateGraphDescr (handle, &graph));

    CSC_input->nvertices = n;
    CSC_input->nedges = nnz;
    CSC_input->destination_offsets = destination_offsets_h;
    CSC_input->source_indices = source_indices_h;


    // Set graph connectivity and properties (tranfers)
    check_status(nvgraphSetGraphStructure(handle, graph, (void*)CSC_input, NVGRAPH_CSC_32));
    check_status(nvgraphAllocateVertexData(handle, graph, vertex_numsets, &vertex_dimT));
    check_status(nvgraphAllocateEdgeData  (handle, graph, edge_numsets, &edge_dimT));
    check_status(nvgraphSetEdgeData(handle, graph, (void*)weights_h, 0));

    // Solve
    int source_vert = 1;
    check_status(nvgraphSssp(handle, graph, 0,  &source_vert, 0));

    check_status(nvgraphGetVertexData(handle, graph, (void*)sssp_h, 0));

    //for (int i = 0; i<n; i++)  printf("%f\n",sssp_h[i]); printf("\n");


    source_vert = 0;
    check_status(nvgraphSssp(handle, graph, 0,  &source_vert, 0));

    check_status(nvgraphGetVertexData(handle, graph, (void*)sssp_h, 0));

    for (int i = 0; i<n; i++)  printf("%f\n",sssp_h[i]); printf("\n");


    free(destination_offsets_h);
    free(source_indices_h);
    free(weights_h);
    free(CSC_input);

    //Clean
    check_status(nvgraphDestroyGraphDescr (handle, graph));
    check_status(nvgraphDestroy (handle));

    file.close();

    return EXIT_SUCCESS;
}