// ----------------------------------------------------------------------------
// -                        Open3D: www.open3d.org                            -
// ----------------------------------------------------------------------------
// Copyright (c) 2018-2023 www.open3d.org
// SPDX-License-Identifier: MIT
// ----------------------------------------------------------------------------
//

#include <vector>

#include "open3d/ml/impl/continuous_conv/ContinuousConvTranspose.h"
#include "torch/script.h"

using namespace open3d::ml::impl;

template <class TFeat, class TOut, class TReal, class TIndex>
void ContinuousConvTransposeCPU(
        const torch::Tensor& filters,
        const torch::Tensor& out_positions,
        const torch::Tensor& out_importance,
        const torch::Tensor& extents,
        const torch::Tensor& offset,
        const torch::Tensor& inp_positions,
        const torch::Tensor& inp_features,
        const torch::Tensor& inp_neighbors_index,
        const torch::Tensor& inp_neighbors_importance_sum,
        const torch::Tensor& inp_neighbors_row_splits,
        const torch::Tensor& neighbors_index,
        const torch::Tensor& neighbors_importance,
        const torch::Tensor& neighbors_row_splits,
        const bool align_corners,
        const CoordinateMapping coordinate_mapping,
        const bool normalize,
        const InterpolationMode interpolation,
        const int64_t max_temp_mem_MB,
        torch::Tensor& out_features) {
    const bool individual_extents = extents.size(0) > 1;
    const bool isotropic_extents = extents.size(1) == 1;
    std::vector<int> filter_dims;
    for (auto d : filters.sizes()) {
        filter_dims.push_back(d);
    }

    CConvTransposeComputeFeaturesCPU<TFeat, TOut, TReal, TIndex>(
            out_features.data_ptr<TOut>(), filter_dims,
            filters.data_ptr<TFeat>(), out_positions.size(0),
            out_positions.data_ptr<TReal>(),
            out_importance.size(0) ? out_importance.data_ptr<TFeat>() : nullptr,
            inp_positions.size(0), inp_positions.data_ptr<TReal>(),
            inp_features.data_ptr<TFeat>(),
            inp_neighbors_importance_sum.size(0)
                    ? inp_neighbors_importance_sum.data_ptr<TFeat>()
                    : nullptr,
            inp_neighbors_row_splits.data_ptr<int64_t>(),
            neighbors_index.size(0),
            (TIndex*)neighbors_index.data_ptr<TIndex>(),
            neighbors_importance.size(0)
                    ? neighbors_importance.data_ptr<TFeat>()
                    : nullptr,
            neighbors_row_splits.data_ptr<int64_t>(), extents.data_ptr<TReal>(),
            offset.data_ptr<TReal>(), interpolation, coordinate_mapping,
            align_corners, individual_extents, isotropic_extents, normalize);
}
#define INSTANTIATE(TFeat, TOut, TReal, TIndex)                                \
    template void ContinuousConvTransposeCPU<TFeat, TOut, TReal, TIndex>(      \
            const torch::Tensor& filters, const torch::Tensor& out_positions,  \
            const torch::Tensor& out_importance, const torch::Tensor& extents, \
            const torch::Tensor& offset, const torch::Tensor& inp_positions,   \
            const torch::Tensor& inp_features,                                 \
            const torch::Tensor& inp_neighbors_index,                          \
            const torch::Tensor& inp_neighbors_importance_sum,                 \
            const torch::Tensor& inp_neighbors_row_splits,                     \
            const torch::Tensor& neighbors_index,                              \
            const torch::Tensor& neighbors_importance,                         \
            const torch::Tensor& neighbors_row_splits,                         \
            const bool align_corners,                                          \
            const CoordinateMapping coordinate_mapping, const bool normalize,  \
            const InterpolationMode interpolation,                             \
            const int64_t max_temp_mem_MB, torch::Tensor& out_features);

INSTANTIATE(float, float, float, int32_t)