#include <iostream>

#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 = float;

    const int64_t n = 1; // Batch size
    const int64_t c = 2; // Number of channels
    const int64_t h = 3; // Height
    const int64_t w = 4; // Width
    const auto permutation
        = std::vector<int64_t>{0, 2, 3, 1}; // nhwc->nchw[0, 1, 2, 3] or nchw->nhwc[0, 2, 3, 1]

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

        graph->set_name("transpose_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);

        auto input = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes()
                .set_name("input")
                .set_dim({n, c, h, w})
                .set_stride({c * h * w, h * w, w, 1})); // nhwc->nchw下，修改为nhwc的stride步长

        auto transposeAttributes = hipdnn_frontend::graph::TransposeAttributes()
                                       .set_name("transpose_node")
                                       .set_permutation(permutation);

        auto output = graph->transpose(input, transposeAttributes);
        output->set_output(true);

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

        return std::make_tuple(graph, input, 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, input, output] = buildTransposeGraph(handle);

    hipdnn_data_sdk::utilities::Tensor<InputType> inputTensor(input->get_dim(),
                                                              input->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> outputTensor(output->get_dim(),
                                                               output->get_stride());

    std::unordered_map<int64_t, void*> variantPack;
    variantPack[input->get_uid()] = inputTensor.memory().deviceData();
    variantPack[output->get_uid()] = outputTensor.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 << "Transpose graph execution complete. \n";

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