#include <iostream>

#include "hipdnn_frontend/attributes/MatmulAttributes.hpp"
#include "utils.hpp"

#include <hipdnn_data_sdk/utilities/Tensor.hpp>
#include <hipdnn_data_sdk/utilities/Workspace.hpp>
#include <hipdnn_frontend.hpp>

int main()
{
    using InputType = hipdnn_data_sdk::types::half;

    const int64_t b = 2; // Batch size
    // Input
    const int64_t n = 16; // Height
    const int64_t m = 32; // Width

    auto buildMatmulsGraph = [=](hipdnnHandle_t handle) {
        auto graph = std::make_shared<hipdnn_frontend::graph::Graph>();

        graph->set_name("matmul_bias_graph")
            .set_io_data_type(hipdnn_frontend::getDataTypeEnumFromType<InputType>())
            .set_intermediate_data_type(hipdnn_frontend::getDataTypeEnumFromType<InputType>())
            .set_compute_data_type(hipdnn_frontend::DataType::FLOAT);

        // create matmul
        auto inputA = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes()
                .set_name("input_a")
                .set_dim({b, n, m})
                .set_stride({n * m, m, 1}));
        auto inputB = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes()
                .set_name("input_b")
                .set_dim({b, m, n})
                .set_stride({n * m, n, 1}));
        auto matmulAttributes = hipdnn_frontend::graph::MatmulAttributes().set_name("matmul_node");
        auto output = graph->matmul(inputA, inputB, matmulAttributes);
        output->set_output(true);

        // build graph
        HIPDNN_FE_CHECK(graph->build(handle));

        return std::make_tuple(graph, inputA, inputB, output);
    };

    auto backend = hipdnn_frontend::detail::hipdnnBackend();
    if(!backend)
    {
        std::cout << "Creat backend failed. \n";
        return 1;
    }

    hipdnnHandle_t handle;
    HIPDNN_CHECK(backend->create(&handle));

    auto [graph, inputA, inputB, output] = buildMatmulsGraph(handle);

    // Allocate DCU memory
    hipdnn_data_sdk::utilities::Tensor<InputType> inputATensor(inputA->get_dim(),
                                                               inputA->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> inputBTensor(inputB->get_dim(),
                                                               inputB->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> outTensor(output->get_dim(),
                                                            output->get_stride());

    std::unordered_map<int64_t, void*> variantPack;
    variantPack[inputA->get_uid()] = inputATensor.memory().deviceData();
    variantPack[inputB->get_uid()] = inputBTensor.memory().deviceData();
    variantPack[output->get_uid()] = outTensor.memory().deviceData();

    int64_t workspaceSize = 0;
    HIPDNN_FE_CHECK(graph->get_workspace_size(workspaceSize));
    const hipdnn_data_sdk::utilities::Workspace workspace(static_cast<size_t>(workspaceSize));

    HIPDNN_FE_CHECK(graph->execute(handle, variantPack, workspace.get()));

    std::cout << "Matmul graph execution complete. \n";

    HIPDNN_CHECK(backend->destroy(handle));
    return 0;
}