/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.

   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 "paddle/operators/margin_rank_loss_op.h"

namespace paddle {
namespace operators {

class MarginRankLossOp : public framework::OperatorWithKernel {
 public:
  using framework::OperatorWithKernel::OperatorWithKernel;

 protected:
  void InferShape(framework::InferShapeContext *ctx) const override {
    // input check
    PADDLE_ENFORCE(ctx->HasInput("Label"), "Input(Label) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("X1"), "Input(X1) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("X2"), "Input(X2) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasOutput("Out"), "Output(Out) shouldn't be null.");
    auto label_dims = ctx->GetInputDim("Label");
    auto x1_dims = ctx->GetInputDim("X1");
    auto x2_dims = ctx->GetInputDim("X2");
    PADDLE_ENFORCE(
        (label_dims == x1_dims) && (x1_dims == x2_dims) &&
            (label_dims.size() == 2) && (label_dims[1] == 1),
        "All inputs must be 2-D tensor with shape [batch_size x 1].");
    ctx->SetOutputDim("Activated", label_dims);
    ctx->SetOutputDim("Out", label_dims);
  }
};

template <typename T>
class MarginRankLossOpMaker : public framework::OpProtoAndCheckerMaker {
 public:
  MarginRankLossOpMaker(framework::OpProto *proto,
                        framework::OpAttrChecker *op_checker)
      : OpProtoAndCheckerMaker(proto, op_checker) {
    AddInput("X1",
             "(2-D tensor with shape [batch_size x 1]) The score for "
             "one item X1 to be ranked, from pairwise ranking model.");
    AddInput("X2",
             "(2-D tensor with shape [batch_size x 1]) The score for "
             "another item X2 to be ranked, from pairwise ranking model.");
    AddInput("Label",
             "(2-D tensor with shape [batch_size x 1]) "
             "The label indicating X1 ranked higher than X2 or not, "
             "can only be +1 or -1.");
    AddAttr<T>("margin", "(scalar, default 0) Margin for MarginRankLossOp.")
        .SetDefault(static_cast<T>(0));
    AddOutput("Activated",
              "(2-D tensor with shape [batch_size x 1]) Intermediate tensor "
              "to indicate whether each element of Output(Out) is activated.")
        .AsIntermediate();
    AddOutput("Out",
              "(2-D tensor with shape [batch_size x 1]) "
              "The output loss of MarginRankLoss operator.");
    AddComment(R"DOC(

MarginRankLoss operator measures the loss given a pair of training sample
{`X1`, `X2`} and the `Label` with attribute `margin`, where `Label = +1` 
indicating X1 is ranked higher than `X2`, otherwise `Label = -1`. The loss 
turns out

loss(X1, X2, Label) = max(0, -Label * (X1 - X2) + margin).

The attribute `margin` involved here helps make the predictions more robust.
Denote the item ranked higher as the positive sample, otherwise negative 
sample. If the score of the two samples statisfies 

positive sample - negative sample < margin,

the pair of samples will contribute to the loss, which will backpropogate and 
train the ranking model to enlarge the difference of the two score.

For batch input with size `batch_size`, `X1`, `X2` and `Label`
all have the same shape [batch_size x 1].

)DOC");
  }
};

class MarginRankLossGradOp : public framework::OperatorWithKernel {
 public:
  using framework::OperatorWithKernel::OperatorWithKernel;

 protected:
  void InferShape(framework::InferShapeContext *ctx) const override {
    PADDLE_ENFORCE(ctx->HasInput("Label"), "Input(Label) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("X1"), "Input(X1) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("X2"), "Input(X2) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput(framework::GradVarName("Out")),
                   "Input(Out@GRAD) shouldn't be null.");
    PADDLE_ENFORCE(ctx->HasInput("Activated"),
                   "Intermediate(Activated) shouldn't be null.");
    auto dims = ctx->GetInputDim("Label");
    ctx->SetOutputDim(framework::GradVarName("X1"), dims);
    ctx->SetOutputDim(framework::GradVarName("X2"), dims);
  }
};

}  // namespace operators
}  // namespace paddle
namespace ops = paddle::operators;

REGISTER_OP(margin_rank_loss, ops::MarginRankLossOp,
            ops::MarginRankLossOpMaker<float>, margin_rank_loss_grad,
            ops::MarginRankLossGradOp);
REGISTER_OP_CPU_KERNEL(
    margin_rank_loss,
    ops::MarginRankLossKernel<paddle::platform::CPUPlace, float>);
REGISTER_OP_CPU_KERNEL(
    margin_rank_loss_grad,
    ops::MarginRankLossGradKernel<paddle::platform::CPUPlace, float>);