QTClusterFinder.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: Hendrik Weisser $
// $Authors: Steffen Sass, Hendrik Weisser $
// --------------------------------------------------------------------------
 
#pragma once
 
#include <OpenMS/ANALYSIS/MAPMATCHING/BaseGroupFinder.h>
#include <OpenMS/CONCEPT/ProgressLogger.h>
#include <OpenMS/DATASTRUCTURES/GridFeature.h>
#include <OpenMS/DATASTRUCTURES/QTCluster.h>
#include <OpenMS/ANALYSIS/MAPMATCHING/FeatureDistance.h>
 
#include <unordered_map>
 
#include <list>
#include <vector>
#include <unordered_set>
#include <utility> // for pair<>
 
namespace OpenMS
{
 
  class OPENMS_DLLAPI QTClusterFinder :
    public BaseGroupFinder
  {
  public:
 
    typedef std::unordered_map< 
              std::pair<OpenMS::GridFeature*, OpenMS::GridFeature*>,
              double> PairDistances;
 
    typedef std::unordered_map<
              const OpenMS::GridFeature*, std::unordered_set<Size> > ElementMapping;
 
    struct Heap;
 
    struct HeapHandles;
 
    struct Grid;
 
  private:
    Size num_maps_;
 
    bool use_IDs_;
 
    //TODO we could also bin by equal sized RT bins, or by a fixed RT size
    //TODO could be made dependent on nr. of maps (e.g. with 5 maps you get [around] 4 diffs per ID already)
    Size min_nr_diffs_per_bin_;
 
    double min_score_;
 
    double noID_penalty_;
 
    double max_diff_rt_;
 
    double max_diff_mz_;
 
    int nr_partitions_;
 
    FeatureDistance feature_distance_;
 
    std::unordered_set<const OpenMS::GridFeature*> already_used_;
 
    std::map<double, double> bin_tolerances_;
 
    double getDistance_(const OpenMS::GridFeature* left, const
        OpenMS::GridFeature* right);
 
    void setParameters_(double max_intensity, double max_mz);
 
    bool makeConsensusFeature_(Heap& cluster_heads,
                               ConsensusFeature& feature,
                               ElementMapping& element_mapping,
                               const Grid& grid,
                               const HeapHandles& handles);
 
    void computeClustering_(const Grid& grid,
                            Heap& cluster_heads,
                            std::vector<QTCluster::BulkData>& cluster_data,
                            HeapHandles& handles,
                            ElementMapping& element_mapping);
 
    void removeFromElementMapping_(const QTCluster& cluster,
                                   ElementMapping& element_mapping);
    
    void createConsensusFeature_(ConsensusFeature& feature, const double quality, 
                                 const QTCluster::Elements& elements);
 
    void updateClustering_(ElementMapping& element_mapping,
                           const Grid& grid,
                           const QTCluster::Elements& elements,
                           Heap& cluster_heads,
                           const HeapHandles& handles,
                           Size best_id);
 
    template <typename MapType>
    void run_(const std::vector<MapType>& input_maps, ConsensusMap& result_map);
 
    template <typename MapType>
    void run_internal_(const std::vector<MapType>& input_maps,
                       ConsensusMap& result_map, bool do_progress);
 
    void addClusterElements_(const Grid& grid, QTCluster& cluster);
 
    bool distIsOutlier_(double dist, double rt);
 
protected:
 
    enum
    {
      RT = Peak2D::RT,
      MZ = Peak2D::MZ
    };
 
public:
 
    QTClusterFinder();
 
    ~QTClusterFinder() override;
 
    void run(const std::vector<ConsensusMap>& input_maps,
             ConsensusMap& result_map) override;
 
    void run(const std::vector<FeatureMap>& input_maps,
             ConsensusMap& result_map);
 
  };
} // namespace OpenMS