tensorflow/core/kernels/data/concatenate_dataset_op.cc

/* Copyright 2017 The TensorFlow Authors. All Rights Reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
==============================================================================*/
#include "tensorflow/core/kernels/data/concatenate_dataset_op.h"

#include "tensorflow/core/framework/partial_tensor_shape.h"
#include "tensorflow/core/framework/tensor.h"
#include "tensorflow/core/kernels/data/name_utils.h"

namespace tensorflow {
namespace data {

// See documentation in ../../ops/dataset_ops.cc for a high-level
// description of the following op.

/* static */ constexpr const char* const ConcatenateDatasetOp::kDatasetType;
/* static */ constexpr const char* const ConcatenateDatasetOp::kInputDataset;
/* static */ constexpr const char* const ConcatenateDatasetOp::kAnotherDataset;
/* static */ constexpr const char* const ConcatenateDatasetOp::kOutputTypes;
/* static */ constexpr const char* const ConcatenateDatasetOp::kOutputShapes;

constexpr char kIndex[] = "i";
constexpr char kInputImplUninitialized[] = "input_impl_uninitialized";

class ConcatenateDatasetOp::Dataset : public DatasetBase {
 public:
  explicit Dataset(OpKernelContext* ctx, const DatasetBase* input,
                   const DatasetBase* to_concatenate)
      : DatasetBase(DatasetContext(ctx)),
        input_(input),
        to_concatenate_(to_concatenate) {
    input_->Ref();
    to_concatenate_->Ref();

    auto os_input = input->output_shapes();
    auto os_concatenate = to_concatenate->output_shapes();
    for (int i = 0; i < os_input.size(); i++) {
      output_shapes_.push_back(
          MostSpecificCompatibleShape(os_input[i], os_concatenate[i]));
    }
  }
  ~Dataset() override {
    input_->Unref();
    to_concatenate_->Unref();
  }

  std::unique_ptr<IteratorBase> MakeIteratorInternal(
      const string& prefix) const override {
    return absl::make_unique<Iterator>(Iterator::Params{
        this, name_utils::IteratorPrefix(kDatasetType, prefix)});
  }

  const DataTypeVector& output_dtypes() const override {
    return input_->output_dtypes();
  }

  const std::vector<PartialTensorShape>& output_shapes() const override {
    return output_shapes_;
  }

  string DebugString() const override {
    return name_utils::DatasetDebugString(kDatasetType);
  }

  int64 Cardinality() const override {
    int64 n1 = input_->Cardinality();
    int64 n2 = to_concatenate_->Cardinality();
    if (n1 == kInfiniteCardinality || n2 == kInfiniteCardinality) {
      return kInfiniteCardinality;
    }
    if (n1 == kUnknownCardinality || n2 == kUnknownCardinality) {
      return kUnknownCardinality;
    }
    return n1 + n2;
  }

  Status CheckExternalState() const override {
    TF_RETURN_IF_ERROR(input_->CheckExternalState());
    return to_concatenate_->CheckExternalState();
  }

 protected:
  Status AsGraphDefInternal(SerializationContext* ctx,
                            DatasetGraphDefBuilder* b,
                            Node** output) const override {
    Node* input_graph = nullptr;
    TF_RETURN_IF_ERROR(b->AddInputDataset(ctx, input_, &input_graph));
    Node* to_concatenate_graph = nullptr;
    TF_RETURN_IF_ERROR(
        b->AddInputDataset(ctx, to_concatenate_, &to_concatenate_graph));
    TF_RETURN_IF_ERROR(
        b->AddDataset(this, {input_graph, to_concatenate_graph}, output));
    return Status::OK();
  }

 private:
  class Iterator : public DatasetIterator<Dataset> {
   public:
    explicit Iterator(const Params& params)
        : DatasetIterator<Dataset>(params), i_(0) {}

    Status Initialize(IteratorContext* ctx) override {
      return dataset()->input_->MakeIterator(
          ctx, strings::StrCat(prefix(), "[0]"), &input_impl_);
    }

    Status GetNextInternal(IteratorContext* ctx,
                           std::vector<Tensor>* out_tensors,
                           bool* end_of_sequence) override {
      mutex_lock l(mu_);
      if (!input_impl_) {
        *end_of_sequence = true;
        return Status::OK();
      }
      while (i_ < 2) {
        TF_RETURN_IF_ERROR(
            input_impl_->GetNext(ctx, out_tensors, end_of_sequence));
        if (!*end_of_sequence) {
          return Status::OK();
        }
        if (++i_ < 2) {
          TF_RETURN_IF_ERROR(dataset()->to_concatenate_->MakeIterator(
              ctx, strings::StrCat(prefix(), "[1]"), &input_impl_));
        }
      }
      *end_of_sequence = true;
      input_impl_.reset();
      return Status::OK();
    }

   protected:
    std::shared_ptr<model::Node> CreateNode(
        IteratorContext* ctx, model::Node::Args args) const override {
      return model::MakeKnownRatioNode(std::move(args),
                                       /*ratio=*/1);
    }

    Status SaveInternal(IteratorStateWriter* writer) override {
      mutex_lock l(mu_);
      TF_RETURN_IF_ERROR(writer->WriteScalar(full_name(kIndex), i_));
      if (input_impl_) {
        TF_RETURN_IF_ERROR(SaveInput(writer, input_impl_));
      } else {
        TF_RETURN_IF_ERROR(
            writer->WriteScalar(full_name(kInputImplUninitialized), ""));
      }
      return Status::OK();
    }

    Status RestoreInternal(IteratorContext* ctx,
                           IteratorStateReader* reader) override {
      mutex_lock l(mu_);
      TF_RETURN_IF_ERROR(reader->ReadScalar(full_name(kIndex), &i_));
      if (reader->Contains(full_name(kInputImplUninitialized))) {
        input_impl_.reset();
        return Status::OK();
      }
      if (!TF_PREDICT_TRUE(i_ >= 0 && i_ <= 2))
        return errors::InvalidArgument("i_ must be in range [0, 2].");
      if (i_ == 1) {
        TF_RETURN_IF_ERROR(dataset()->to_concatenate_->MakeIterator(
            ctx, strings::StrCat(prefix(), "[1]"), &input_impl_));
      } else if (i_ == 2) {
        input_impl_.reset();
      }
      if (input_impl_) {
        TF_RETURN_IF_ERROR(RestoreInput(ctx, reader, input_impl_));
      }
      return Status::OK();
    }

   private:
    mutex mu_;
    int64 i_ GUARDED_BY(mu_);
    std::unique_ptr<IteratorBase> input_impl_ GUARDED_BY(mu_);
  };

  static PartialTensorShape MostSpecificCompatibleShape(
      const PartialTensorShape& ts1, const PartialTensorShape& ts2) {
    if (ts1.dims() != ts2.dims() || ts1.unknown_rank() || ts2.unknown_rank())
      return PartialTensorShape();
    PartialTensorShape output_tensorshape({});
    auto dims1 = ts1.dim_sizes();
    auto dims2 = ts2.dim_sizes();
    for (int d = 0; d < ts1.dims(); d++) {
      if (dims1[d] == dims2[d])
        output_tensorshape.AddDim(dims1[d]);
      else
        output_tensorshape.AddDim(-1);
    }
    return output_tensorshape;
  }

  const DatasetBase* input_;
  const DatasetBase* to_concatenate_;
  std::vector<PartialTensorShape> output_shapes_;
};

ConcatenateDatasetOp::ConcatenateDatasetOp(OpKernelConstruction* ctx)
    : BinaryDatasetOpKernel(ctx) {}

void ConcatenateDatasetOp::MakeDataset(OpKernelContext* ctx, DatasetBase* input,
                                       DatasetBase* to_concatenate,
                                       DatasetBase** output) {
  OP_REQUIRES(ctx, input->output_dtypes() == to_concatenate->output_dtypes(),
              errors::InvalidArgument(
                  "input dataset and dataset to concatenate"
                  " have different output_types %s and %s",
                  (DataTypeVectorString(input->output_dtypes()),
                   DataTypeVectorString(to_concatenate->output_dtypes()))));
  *output = new Dataset(ctx, input, to_concatenate);
}

namespace {
REGISTER_KERNEL_BUILDER(Name("ConcatenateDataset").Device(DEVICE_CPU),
                        ConcatenateDatasetOp);
}  // namespace
}  // namespace data
}  // namespace tensorflow