Program Listing for File add_scores_stage_base.cpp

            

            
/*
* SPDX-FileCopyrightText: Copyright (c) 2021-2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
* SPDX-License-Identifier: Apache-2.0
*
* 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 "morpheus/stages/add_scores_stage_base.hpp"

#include "morpheus/messages/memory/tensor_memory.hpp" // for TensorMemory
#include "morpheus/messages/meta.hpp"
#include "morpheus/messages/multi_response.hpp" // for MultiResponseMessage
#include "morpheus/objects/dtype.hpp" // for DType
#include "morpheus/objects/tensor.hpp" // for Tensor
#include "morpheus/objects/tensor_object.hpp" // for TensorObject
#include "morpheus/types.hpp" // for TensorIndex
#include "morpheus/utilities/matx_util.hpp" // for MatxUtil
#include "morpheus/utilities/string_util.hpp" // for StringUtil
#include "morpheus/utilities/tensor_util.hpp" // for TensorUtils

#include <glog/logging.h> // for CHECK, COMPACT_GOOGLE_LOG_FATAL, LogMessageFatal, COMP...
#include <rxcpp/rx.hpp> // for observable_member, trace_activity, decay_t, operator|

#include <cstddef> // for size_t
#include <iterator> // for reverse_iterator
#include <memory> // for shared_ptr, allocator, __shared_ptr_access
#include <ostream> // for basic_ostream, operator<<, basic_ostream::operator<<
#include <stdexcept> // for runtime_error
#include <type_traits> // for is_same_v
#include <typeinfo> // for type_info
#include <utility> // for move, pair
#include <vector> // for vector
// IWYU thinks we need __alloc_traits<>::value_type for vector assignments
// IWYU pragma: no_include <ext/alloc_traits.h>
// IWYU pragma: no_include <operators/rx-map.hpp>

namespace morpheus {

// Component public implementations
// ************ AddClassificationStage **************************** //
template <typename InputT, typename OutputT>
AddScoresStageBase<InputT, OutputT>::AddScoresStageBase(std::map<std::size_t, std::string> idx2label,
                                                        std::optional<float> threshold) :
  base_t(),
  m_idx2label(std::move(idx2label)),
  m_threshold(threshold),
  m_min_col_count(m_idx2label.rbegin()->first)  // Ordered map's largest key will be the last entry
{
    this->pipe(rxcpp::operators::map([this](sink_type_t x) {
        return this->on_data(std::move(x));
    }));
}

template <typename InputT, typename OutputT>
AddScoresStageBase<InputT, OutputT>::source_type_t AddScoresStageBase<InputT, OutputT>::on_data(sink_type_t x)
{
    if constexpr (std::is_same_v<sink_type_t, std::shared_ptr<MultiResponseMessage>>)
    {
        this->on_multi_response_message(x);
    }
    else if constexpr (std::is_same_v<sink_type_t, std::shared_ptr<ControlMessage>>)
    {
        this->on_control_message(x);
    }
    // sink_type_t not supported
    else
    {
        std::string error_msg{"AddScoresStageBase receives unsupported input type: " + std::string(typeid(x).name())};
        LOG(ERROR) << error_msg;
        throw std::runtime_error(error_msg);
    }
    return x;
}

template <>
void AddScoresStageBase<MultiResponseMessage, MultiResponseMessage>::on_multi_response_message(
    std::shared_ptr<MultiResponseMessage> x)
{
    auto probs = x->get_probs_tensor();

    const auto& shape = probs.get_shape();

    // Depending on the input the stride is given in bytes or elements, convert to elements
    auto stride = TensorUtils::get_element_stride(probs.get_stride());

    CHECK(shape.size() == 2 && shape[1] > m_min_col_count)
        << "Model output did not contain enough columns to fufill the requested labels. Label "
           "indexes: "
        << StringUtil::map_to_str(m_idx2label.begin(), m_idx2label.end()) << ", Model output columns: " << shape[1];

    const auto num_rows    = shape[0];
    const auto num_columns = shape[1];

    TensorObject output_tensor;

    if (m_threshold.has_value())
    {
        auto thresh_bool_buffer = MatxUtil::threshold(
            {probs.data(), probs.dtype(), probs.get_memory(), probs.get_shape(), probs.get_stride()},
            *m_threshold,
            false);

        output_tensor.swap(Tensor::create(thresh_bool_buffer, DType::create<bool>(), shape, stride));
    }
    else
    {
        output_tensor.swap(std::move(probs));
    }

    std::vector<std::string> columns;
    std::vector<TensorObject> tensors;

    std::size_t i = 0;
    for (const auto& [column_num, column_name] : m_idx2label)
    {
        columns.push_back(column_name);
        tensors.emplace_back(output_tensor.slice({0, static_cast<TensorIndex>(column_num)},
                                                 {num_rows, static_cast<TensorIndex>(column_num + 1)}));

        ++i;
    }

    x->set_meta(columns, tensors);
}

template <>
void AddScoresStageBase<ControlMessage, ControlMessage>::on_control_message(std::shared_ptr<ControlMessage> x)
{
    // The default of probs_tensor_name is "probs"
    auto probs        = x->tensors()->get_tensor("probs");
    const auto& shape = probs.get_shape();

    // Depending on the input the stride is given in bytes or elements, convert to elements
    auto stride = TensorUtils::get_element_stride(probs.get_stride());

    CHECK(shape.size() == 2 && shape[1] > m_min_col_count)
        << "Model output did not contain enough columns to fufill the requested labels. Label "
           "indexes: "
        << StringUtil::map_to_str(m_idx2label.begin(), m_idx2label.end()) << ", Model output columns: " << shape[1];

    const auto num_rows    = shape[0];
    const auto num_columns = shape[1];

    TensorObject output_tensor;

    if (m_threshold.has_value())
    {
        auto thresh_bool_buffer = MatxUtil::threshold(
            {probs.data(), probs.dtype(), probs.get_memory(), probs.get_shape(), probs.get_stride()},
            *m_threshold,
            false);

        output_tensor.swap(Tensor::create(thresh_bool_buffer, DType::create<bool>(), shape, stride));
    }
    else
    {
        output_tensor.swap(std::move(probs));
    }

    std::vector<std::string> columns;
    std::vector<TensorObject> tensors;

    std::size_t i = 0;
    for (const auto& [column_num, column_name] : m_idx2label)
    {
        columns.push_back(column_name);
        tensors.emplace_back(output_tensor.slice({0, static_cast<TensorIndex>(column_num)},
                                                 {num_rows, static_cast<TensorIndex>(column_num + 1)}));

        ++i;
    }

    x->payload()->set_data(columns, tensors);
}

template class AddScoresStageBase<MultiResponseMessage, MultiResponseMessage>;
template class AddScoresStageBase<ControlMessage, ControlMessage>;

}  // namespace morpheus