#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
    // Input
    const int64_t c = 16; // Number of channels
    const int64_t h = 16; // Height
    const int64_t w = 16; // Width

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

        graph->set_name("groupnorm_backward_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); //

        const int numGroups = 2;
        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}));
        auto dy = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes()
                .set_name("dy")
                .set_dim({n, c, h, w})
                .set_stride({c * h * w, h * w, w, 1}));
        auto scale = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes().set_name("scale").set_dim({c}).set_stride(
                {1}));
        auto epsilon = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes()
                .set_name("epsilon")
                .set_data_type(hipdnn_frontend::DataType::FLOAT)
                .set_value(1e-5f));
        auto groups = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes()
                .set_name("groups")
                .set_data_type(hipdnn_frontend::DataType::INT32)
                .set_value(numGroups));

        auto mean = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes()
                .set_name("mean")
                .set_dim({numGroups * n})
                .set_stride({1}));
        auto invVariance = std::make_shared<hipdnn_frontend::graph::TensorAttributes>(
            hipdnn_frontend::graph::Tensor_attributes()
                .set_name("inv_variance")
                .set_dim({numGroups * n})
                .set_stride({1}));

        auto groupnormBwdAttributes = hipdnn_frontend::graph::GroupnormBwdAttributes()
                                          .set_name("groupnorm_backward_node")
                                          .set_epsilon(epsilon)
                                          .set_groups(groups);

        auto [dx, dbias, dscale] = graph->groupnorm_backward(
            input, dy, scale, mean, invVariance, groupnormBwdAttributes);
        dx->set_output(true);
        dbias->set_output(true);
        dscale->set_output(true);

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

        return std::make_tuple(graph, input, dy, scale, mean, invVariance, dx, dbias, dscale);
    };

    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, dy, scale, mean, invVariance, dx, dbias, dscale]
        = buildGNBackwardGraph(handle);

    // Allocate DCU memory
    hipdnn_data_sdk::utilities::Tensor<InputType> inputTensor(input->get_dim(),
                                                              input->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> dyTensor(dy->get_dim(), dy->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> scaleTensor(scale->get_dim(),
                                                              scale->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> meanTensor(mean->get_dim(), mean->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> invVarianceTensor(invVariance->get_dim(),
                                                                    invVariance->get_stride());

    hipdnn_data_sdk::utilities::Tensor<InputType> dxTensor(dx->get_dim(), dx->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> dbiasTensor(dbias->get_dim(),
                                                              dbias->get_stride());
    hipdnn_data_sdk::utilities::Tensor<InputType> dscaleTensor(dscale->get_dim(),
                                                               dscale->get_stride());

    std::unordered_map<int64_t, void*> variantPack;
    variantPack[input->get_uid()] = inputTensor.memory().deviceData();
    variantPack[dy->get_uid()] = dyTensor.memory().deviceData();
    variantPack[scale->get_uid()] = scaleTensor.memory().deviceData();
    variantPack[mean->get_uid()] = meanTensor.memory().deviceData();
    variantPack[invVariance->get_uid()] = invVarianceTensor.memory().deviceData();
    variantPack[dx->get_uid()] = dxTensor.memory().deviceData();
    variantPack[dbias->get_uid()] = dbiasTensor.memory().deviceData();
    variantPack[dscale->get_uid()] = dscaleTensor.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 << "GNBackward graph execution complete. \n";

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