#include "op_plugin/AclOpsInterface.h"
#include "op_plugin/OpInterface.h"
#include "op_plugin/utils/op_api_common.h"
namespace op_api {
using npu_preparation = at_npu::native::OpPreparation;
std::tuple<at::Tensor, at::Tensor, at::Tensor> unique_dim(const at::Tensor &self, const int64_t dim, const bool sorted,
const bool return_inverse, const bool return_counts)
{
DO_COMPATIBILITY(aclnnUniqueDim, acl_op::unique_dim(self, dim, sorted, return_inverse, return_counts));
TORCH_CHECK(dim < self.dim(), "Dim's value must be smaller than self's dim.", OPS_ERROR(ErrCode::VALUE));
auto sizes = self.sizes().vec();
auto num_zero_dims = std::count(sizes.begin(), sizes.end(), 0);
if (self.size(dim) == 0) {
TORCH_CHECK(num_zero_dims == 1, "Number of zero sized dimensions is more than one, so unique cannot be applied.",
OPS_ERROR(ErrCode::PARAM));
at::Tensor y = at::empty(sizes, self.options());
at::Tensor y_inverse = at::empty({0}, self.options().dtype(at::kLong));
at::Tensor y_counts = at::empty({0}, self.options().dtype(at::kLong));
return std::tuple<at::Tensor, at::Tensor, at::Tensor>(y, y_inverse, y_counts);
}
TORCH_CHECK(num_zero_dims == 0, "There are 0 sized dimensions, and they aren't selected, so unique cannot be applied.",
OPS_ERROR(ErrCode::PARAM));
at::Tensor y = npu_preparation::apply_tensor_without_format(self);
at::Tensor y_inverse = npu_preparation::apply_tensor_without_format(self.size(dim), self.options().dtype(at::kLong));
at::Tensor y_counts = npu_preparation::apply_tensor_without_format(self.size(dim), self.options().dtype(at::kLong));
static auto opApiFuncAddr = []() {
auto ret = GetOpApiFuncAddr("aclGetViewShape");
TORCH_CHECK(ret != nullptr, "GetOpApiFuncAddr failed.", OPS_ERROR(ErrCode::INTERNAL));
return ret;
}();
using aclGetViewShapeFunc = int (*)(const aclTensor *tensor, int64_t **view_dims, uint64_t *view_dims_num);
auto aclGetViewShape = reinterpret_cast<aclGetViewShapeFunc>(opApiFuncAddr);
OP_EXEC_LOG(aclnnUniqueDim, "EXEC_NPU_CMD_SYNC", self, sorted, return_inverse, dim, y, y_inverse, y_counts);
auto npuAclParams = EXEC_NPU_CMD_SYNC(aclnnUniqueDim, self, sorted, return_inverse, dim, y,
y_inverse, y_counts);
int64_t *view_dims = nullptr;
uint64_t view_dim_num = 0;
constexpr int64_t Y_IDX = 4;
auto ret1 = aclGetViewShape(npuAclParams.Get<Y_IDX>(), &view_dims, &view_dim_num);
TORCH_CHECK(ret1 == 0, "aclGetViewShape for y failed.", OPS_ERROR(ErrCode::ACL));
c10::SmallVector<int64_t, SIZE> output_size_y(view_dims, view_dims + view_dim_num);
y.resize_(output_size_y);
constexpr int64_t Y_COUNTS_IDX = 6;
auto ret2 = aclGetViewShape(npuAclParams.Get<Y_COUNTS_IDX>(), &view_dims, &view_dim_num);
TORCH_CHECK(ret2 == 0, "aclGetViewShape for y_counts failed.", OPS_ERROR(ErrCode::ACL));
c10::SmallVector<int64_t, SIZE> output_size_y_counts(view_dims, view_dims + view_dim_num);
y_counts.resize_(output_size_y_counts);
delete[] view_dims;
view_dims = nullptr;
return std::tuple<at::Tensor, at::Tensor, at::Tensor>(y, y_inverse, y_counts);
}
}