XICParquetFile.h

Go to the documentation of this file.

// Copyright (c) 2002-present, OpenMS Inc. -- EKU Tuebingen, ETH Zurich, and FU Berlin
// SPDX-License-Identifier: BSD-3-Clause
//
// --------------------------------------------------------------------------
// $Maintainer: Justin Sing $
// $Authors: Justin Sing $
// --------------------------------------------------------------------------
 
#pragma once
 
#include <OpenMS/CONCEPT/Types.h>
#include <OpenMS/DATASTRUCTURES/String.h>
#include <OpenMS/FORMAT/ParquetFilter.h>
 
#include <vector>
 
namespace OpenMS
{
  class OPENMS_DLLAPI XICParquetFile
  {
  public:
    struct OPENMS_DLLAPI XICChromatogram
    {
      Int64 run_id{0};
      String source_file;
      Int64 ms_level{0};
 
      bool has_precursor_id{false};
      Int64 precursor_id{0};
      bool has_transition_id{false};
      Int64 transition_id{0};
      String modified_sequence;
      bool has_precursor_charge{false};
      Int64 precursor_charge{0};
      bool has_product_charge{false};
      Int64 product_charge{0};
      bool has_detecting_transition{false};
      Int64 detecting_transition{0};
      bool has_precursor_decoy{false};
      Int64 precursor_decoy{0};
      bool has_product_decoy{false};
      Int64 product_decoy{0};
      bool has_transition_ordinal{false};
      Int64 transition_ordinal{0};
      String transition_type;
      String annotation;
 
      std::vector<double> rt;
      std::vector<double> intensity;
    };
 
    struct OPENMS_DLLAPI XICRunInfo
    {
      Int64 run_id{0};
      String source_file;
    };
 
    struct OPENMS_DLLAPI XICAnalyte
    {
      bool has_precursor_id{false};
      Int64 precursor_id{0};
      String modified_sequence;
      bool has_precursor_charge{false};
      Int64 precursor_charge{0};
      bool has_precursor_decoy{false};
      Int64 precursor_decoy{0};
 
      bool has_transition_id{false};
      Int64 transition_id{0};
      bool has_product_charge{false};
      Int64 product_charge{0};
      bool has_transition_ordinal{false};
      Int64 transition_ordinal{0};
      bool has_detecting_transition{false};
      Int64 detecting_transition{0};
      bool has_product_decoy{false};
      Int64 product_decoy{0};
      String transition_type;
      String annotation;
 
      std::vector<Int64> transition_ids;
      std::vector<Int64> product_charges;
      std::vector<Int64> transition_ordinals;
      std::vector<Int64> detecting_transitions;
      std::vector<Int64> product_decoys;
      std::vector<String> transition_types;
      std::vector<String> annotations;
    };
 
    explicit XICParquetFile(const String& filename);
 
    explicit XICParquetFile(const std::vector<String>& filenames);
    XICParquetFile(const XICParquetFile& rhs) = default;
    XICParquetFile& operator=(const XICParquetFile& rhs) = default;
 
    const String& getFilename() const;
 
    const std::vector<String>& getFilenames() const;
 
    void load(std::vector<XICChromatogram>& output) const;
 
    void getChromatograms(std::vector<XICChromatogram>& output,
                          Int64 precursor_id = -1,
                          Int64 transition_id = -1,
                          const String& modified_sequence = "",
                          Int64 precursor_charge = -1,
                          Int64 product_charge = -1,
                          Int64 ms_level = -1,
                          Int64 run_id = -1,
                          const String& filter = "") const;
 
    void getChromatograms(std::vector<XICChromatogram>& output,
                          const ParquetFilter& filter) const;
 
    void getChromatograms(std::vector<XICChromatogram>& output,
                          const ParquetFilterBuilder& filter) const;
 
    void getRuns(std::vector<XICRunInfo>& output) const;
 
    void getAnalytes(std::vector<XICAnalyte>& output,
                     const std::vector<String>& columns = {},
                     bool nest_transitions = true) const;
 
    void getColumns(std::vector<String>& output) const;
 
  private:
    void getChromatograms_(std::vector<XICChromatogram>& output,
                           const FilterExpression& extra_filter,
                           Int64 precursor_id,
                           Int64 transition_id,
                           const String& modified_sequence,
                           Int64 precursor_charge,
                           Int64 product_charge,
                           Int64 ms_level,
                           Int64 run_id,
                           const String& filter) const;
 
    String filename_;
    std::vector<String> filenames_;
  };
 
  typedef XICParquetFile::XICChromatogram XICChromatogram;
  typedef XICParquetFile::XICRunInfo XICRunInfo;
  typedef XICParquetFile::XICAnalyte XICAnalyte;
} // namespace OpenMS