tzrec/models/dat.py

# Copyright (c) 2024-2025, Alibaba Group;
# 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.

from typing import Any, Dict, List, Optional, Tuple, Union

import torch
import torch.nn.functional as F
from torch import nn
from torch._tensor import Tensor

from tzrec.datasets.utils import Batch
from tzrec.features.feature import BaseFeature
from tzrec.models.match_model import MatchModel, MatchTower
from tzrec.modules.mlp import MLP
from tzrec.modules.utils import div_no_nan
from tzrec.protos import model_pb2, simi_pb2, tower_pb2
from tzrec.utils.config_util import config_to_kwargs


@torch.fx.wrap
def _update_dict_tensor(
    tensor_dict: Dict[str, torch.Tensor],
    new_tensor_dict: Optional[Dict[str, Optional[torch.Tensor]]],
) -> None:
    if new_tensor_dict:
        for k, v in new_tensor_dict.items():
            if v is not None:
                tensor_dict[k] = v


class DATTower(MatchTower):
    """DAT user/item tower.

    Args:
        tower_config (Tower): user/item tower config.
        output_dim (int): user/item output embedding dimension.
        similarity (Similarity): when use COSINE similarity,
            will norm the output embedding.
        feature_group (FeatureGroupConfig): feature group config.
        features (list): list of features.
    """

    def __init__(
        self,
        tower_config: tower_pb2.DATTower,
        output_dim: int,
        similarity: simi_pb2.Similarity,
        feature_groups: List[model_pb2.FeatureGroupConfig],
        features: List[BaseFeature],
        model_config: model_pb2.ModelConfig,
    ) -> None:
        super().__init__(
            tower_config, output_dim, similarity, feature_groups, features, model_config
        )
        self.init_input()
        self._augment_group_name = tower_config.augment_input

        # self.augment_embedding_group = EmbeddingGroup(
        #     augment_features, [augment_feature_group]
        # )

        tower_feature_in = self.embedding_group.group_total_dim(self._group_name)
        tower_augment_feature_in = self.embedding_group.group_total_dim(
            self._augment_group_name
        )

        self.mlp = MLP(
            tower_feature_in + tower_augment_feature_in,
            **config_to_kwargs(tower_config.mlp),
        )
        if self._output_dim > 0:
            self.output = nn.Linear(self.mlp.output_dim(), output_dim)

    def forward(
        self, batch: Batch
    ) -> Union[torch.Tensor, Tuple[torch.Tensor, torch.Tensor]]:
        """Forward the tower.

        Args:
            batch (Batch): input batch data.

        Return:
            embedding (dict): tower output embedding.
        """
        grouped_features = self.build_input(batch)
        input_features = grouped_features[self._group_name]
        augmented_feature = grouped_features[self._augment_group_name]

        output = self.mlp(torch.concat([input_features, augmented_feature], dim=1))
        if self._output_dim > 0:
            output = self.output(output)
        if self._similarity == simi_pb2.Similarity.COSINE:
            output = F.normalize(output, p=2.0, dim=1)

        if self.training:
            return output, augmented_feature
        else:
            return output


class DAT(MatchModel):
    """Dual Augmented Two-Towers model.

    Args:
        model_config (ModelConfig): an instance of ModelConfig.
        features (list): list of features.
        labels (list): list of label names.
    """

    def __init__(
        self,
        model_config: model_pb2.ModelConfig,
        features: List[BaseFeature],
        labels: List[str],
        sample_weights: Optional[List[str]] = None,
        **kwargs: Any,
    ) -> None:
        super().__init__(model_config, features, labels, sample_weights, **kwargs)
        name_to_feature_group = {x.group_name: x for x in model_config.feature_groups}

        user_group = name_to_feature_group[self._model_config.user_tower.input]
        user_augment_group = name_to_feature_group[
            self._model_config.user_tower.augment_input
        ]
        item_group = name_to_feature_group[self._model_config.item_tower.input]
        item_augment_group = name_to_feature_group[
            self._model_config.item_tower.augment_input
        ]

        user_features = self.get_features_in_feature_groups([user_group])
        user_augment_features = self.get_features_in_feature_groups(
            [user_augment_group]
        )
        item_features = self.get_features_in_feature_groups([item_group])
        item_augment_features = self.get_features_in_feature_groups(
            [item_augment_group]
        )

        self.user_tower = DATTower(
            self._model_config.user_tower,
            self._model_config.output_dim,
            self._model_config.similarity,
            [user_group, user_augment_group],
            user_features + user_augment_features,
            model_config,
        )

        self.item_tower = DATTower(
            self._model_config.item_tower,
            self._model_config.output_dim,
            self._model_config.similarity,
            [item_group, item_augment_group],
            item_features + item_augment_features,
            model_config,
        )

        self.amm_u_weight = self._model_config.amm_u_weight
        self.amm_i_weight = self._model_config.amm_i_weight

    def predict(self, batch: Batch) -> Dict[str, Tensor]:
        """Forward the model.

        Args:
            batch (Batch): input batch data.

        Return:
            predictions (dict): a dict of predicted result.
        """
        if self.training:
            user_tower_emb, user_augment = self.user_tower(batch)
            item_tower_emb, item_augment = self.item_tower(batch)
        else:
            user_tower_emb = self.user_tower(batch)
            item_tower_emb = self.item_tower(batch)
            user_augment, item_augment = None, None

        _update_dict_tensor(
            self._loss_collection, self.user_tower.group_variational_dropout_loss
        )
        _update_dict_tensor(
            self._loss_collection, self.item_tower.group_variational_dropout_loss
        )

        ui_sim = (
            self.sim(user_tower_emb, item_tower_emb) / self._model_config.temperature
        )

        if self.training:
            return {
                "similarity": ui_sim,
                "user_augment": user_augment,
                "item_augment": item_augment,
                "user_tower_emb": user_tower_emb.detach(),
                "item_tower_emb": item_tower_emb.detach(),
            }
        else:
            return {"similarity": ui_sim}

    def loss(
        self, predictions: Dict[str, torch.Tensor], batch: Batch
    ) -> Dict[str, torch.Tensor]:
        """Compute the Adaptive-Mimic Mechanism loss."""
        losses = super().loss(predictions, batch)

        sample_weight = (
            batch.sample_weights[self._sample_weight]
            if self._sample_weight
            else torch.Tensor([1.0])
        )

        amm_loss = {}
        batch_size = predictions["similarity"].size(0)
        if "user_augment" in predictions and "item_tower_emb" in predictions:
            user_augment = predictions["user_augment"]
            item_tower_emb = predictions["item_tower_emb"]
            amm_loss_u = self.amm_u_weight * torch.sum(
                torch.square(
                    F.normalize(user_augment, p=2.0, dim=1)
                    - item_tower_emb[:batch_size]
                ),
                dim=1 if self._sample_weight else (0, 1),
            )
            amm_loss["amm_loss_u"] = amm_loss_u

        if "item_augment" in predictions and "user_tower_emb" in predictions:
            item_augment = predictions["item_augment"]
            user_tower_emb = predictions["user_tower_emb"]
            amm_loss_i = self.amm_i_weight * torch.sum(
                torch.square(
                    F.normalize(item_augment[:batch_size], p=2.0, dim=1)
                    - user_tower_emb
                ),
                dim=1 if self._sample_weight else (0, 1),
            )
            amm_loss["amm_loss_i"] = amm_loss_i

        if self._sample_weight:
            for loss_name in amm_loss.keys():
                amm_loss[loss_name] = div_no_nan(
                    torch.mean(amm_loss[loss_name] * sample_weight),
                    torch.mean(sample_weight),
                )

        losses.update(amm_loss)

        return losses