Commit adba6823 authored by Swift Genomics's avatar Swift Genomics Committed by swiftgenomics
Browse files

Move funcs to index utils.

parent 212145ea

src/chromap.h

+1 −1
Original line number Diff line number Diff line
@@ -29,7 +29,7 @@ 
#include "temp_mapping.h"

#include "utils.h"



#define CHROMAP_VERSION "0.2.3-r446"

#define CHROMAP_VERSION "0.2.3-r447"



namespace chromap {

src/index.cc

+18 −57
Original line number Diff line number Diff line
@@ -6,7 +6,6 @@ 
#include <iostream>



#include "minimizer_generator.h"

#include "minimizer_utils.h"



namespace chromap {
@@ -246,62 +245,6 @@ void Index::CheckIndex(uint32_t num_sequences, 
  }

}



void Index::HeapMergeCandidatePositionLists(

    const std::vector<std::vector<uint64_t>> sorted_candidate_position_lists,

    std::vector<uint64_t> &candidate_positions) const {

  std::priority_queue<CandidatePositionWithListIndex> heap;

  std::vector<uint32_t> candidate_position_list_indices(

      sorted_candidate_position_lists.size(), 0);



  for (uint32_t li = 0; li < sorted_candidate_position_lists.size(); ++li) {

    if (sorted_candidate_position_lists[li].size() == 0) {

      continue;

    }

    heap.emplace(li, sorted_candidate_position_lists[li][0]);

  }



  while (!heap.empty()) {

    const CandidatePositionWithListIndex min_candidate_position = heap.top();

    heap.pop();

    candidate_positions.push_back(min_candidate_position.position);

    ++candidate_position_list_indices[min_candidate_position.list_index];



    const uint32_t min_candidate_position_list_index =

        candidate_position_list_indices[min_candidate_position.list_index];

    const std::vector<uint64_t> &min_sorted_candidate_position_list =

        sorted_candidate_position_lists[min_candidate_position.list_index];

    if (min_candidate_position_list_index <

        min_sorted_candidate_position_list.size()) {

      heap.emplace(min_candidate_position.list_index,

                   min_sorted_candidate_position_list

                       [min_candidate_position_list_index]);

    }

  }

}



uint64_t Index::GenerateCandidatePositionForSingleSeedHit(

    uint64_t reference_seed_hit, uint64_t read_seed_hit) const {

  const uint32_t reference_position =

      GenerateSequencePosition(reference_seed_hit);



  const uint32_t read_position = GenerateSequencePosition(read_seed_hit);



  // For now we can't see the reference here. So let us don't validate

  // this seed hit. Instead, we do it later some time when we check the

  // candidates.

  const uint32_t mapping_start_position =

      AreTwoHitsOnTheSameStrand(reference_seed_hit, read_seed_hit)

          ? reference_position - read_position

          : reference_position + read_position - kmer_size_ + 1;



  const uint64_t reference_id = GenerateSequenceIndex(reference_seed_hit);



  const uint64_t candidate_position =

      GenerateCandidatePosition(reference_id, mapping_start_position);



  return candidate_position;

}



int Index::GenerateCandidatePositions(

    const CandidatePositionGeneratingConfig &generating_config,

    MappingMetadata &mapping_metadata) const {
@@ -668,4 +611,22 @@ int Index::GenerateCandidatePositionsFromRepetitiveReadWithMateInfoOnOneStrand( 
  return max_minimizer_count;

}



uint64_t Index::GenerateCandidatePositionForSingleSeedHit(

    uint64_t reference_seed_hit, uint64_t read_seed_hit) const {

  const uint32_t reference_position =

      GenerateSequencePosition(reference_seed_hit);

  const uint32_t read_position = GenerateSequencePosition(read_seed_hit);

  // For now we can't see the reference here. So let us don't validate this

  // candidate position. Instead, we do it later some time when we check the

  // candidates.

  const uint32_t mapping_start_position =

      AreTwoHitsOnTheSameStrand(reference_seed_hit, read_seed_hit)

          ? reference_position - read_position

          : reference_position + read_position - kmer_size_ + 1;

  const uint64_t reference_id = GenerateSequenceIndex(reference_seed_hit);

  const uint64_t candidate_position =

      GenerateCandidatePosition(reference_id, mapping_start_position);

  return candidate_position;

}



}  // namespace chromap

src/index.h

+0 −5
Original line number Diff line number Diff line
@@ -81,14 +81,9 @@ class Index { 
  uint32_t GetLookupTableSize() const { return kh_size(lookup_table_); }



 private:

  void HeapMergeCandidatePositionLists(

      const std::vector<std::vector<uint64_t>> sorted_candidate_position_lists,

      std::vector<uint64_t> &candidate_positions) const;



  uint64_t GenerateCandidatePositionForSingleSeedHit(

      uint64_t reference_seed_hit, uint64_t read_seed_hit) const;



 protected:

  int kmer_size_ = 0;

  int window_size_ = 0;

  // Number of threads to build the index, which is not used right now.

src/index_utils.h

+47 −0
Original line number Diff line number Diff line
@@ -46,6 +46,53 @@ inline static uint64_t GenerateCandidatePositionFromOccurrenceTableEntry( 
  return entry >> 1;

}



// Only used in Index to merge sorted candidate position lists using heap.

struct CandidatePositionWithListIndex {

  uint32_t list_index;

  uint64_t position;



  CandidatePositionWithListIndex(uint32_t list_index, uint64_t position)

      : list_index(list_index), position(position) {}



  bool operator<(const CandidatePositionWithListIndex &h) const {

    // The inversed direction is to make a min-heap.

    return position > h.position;

  }

};



inline static void HeapMergeCandidatePositionLists(

    const std::vector<std::vector<uint64_t>> sorted_candidate_position_lists,

    std::vector<uint64_t> &candidate_positions) {

  std::priority_queue<CandidatePositionWithListIndex> heap;

  std::vector<uint32_t> candidate_position_list_indices(

      sorted_candidate_position_lists.size(), 0);



  for (uint32_t li = 0; li < sorted_candidate_position_lists.size(); ++li) {

    if (sorted_candidate_position_lists[li].size() == 0) {

      continue;

    }

    heap.emplace(li, sorted_candidate_position_lists[li][0]);

  }



  while (!heap.empty()) {

    const CandidatePositionWithListIndex min_candidate_position = heap.top();

    heap.pop();

    candidate_positions.push_back(min_candidate_position.position);

    ++candidate_position_list_indices[min_candidate_position.list_index];



    const uint32_t min_candidate_position_list_index =

        candidate_position_list_indices[min_candidate_position.list_index];

    const std::vector<uint64_t> &min_sorted_candidate_position_list =

        sorted_candidate_position_lists[min_candidate_position.list_index];

    if (min_candidate_position_list_index <

        min_sorted_candidate_position_list.size()) {

      heap.emplace(min_candidate_position.list_index,

                   min_sorted_candidate_position_list

                       [min_candidate_position_list_index]);

    }

  }

}



}  // namespace chromap



#endif  // INDEX_UTILS_H_

src/utils.h

+0 −14
Original line number Diff line number Diff line
@@ -42,20 +42,6 @@ struct _mm_history { 
  uint32_t repetitive_seed_length;

};



// Only used in Index to merge sorted candidate position lists using heap.

struct CandidatePositionWithListIndex {

  uint32_t list_index;

  uint64_t position;



  CandidatePositionWithListIndex(uint32_t list_index, uint64_t position)

      : list_index(list_index), position(position) {}



  bool operator<(const CandidatePositionWithListIndex &h) const {

    // The inversed direction is to make a min-heap.

    return position > h.position;

  }

};



KHASH_MAP_INIT_INT64(k128, uint128_t);

KHASH_MAP_INIT_INT64(k64_seq, uint64_t);

KHASH_SET_INIT_INT(k32_set);

CRA Git | Maintained and supported by SUSTech CRA and CCSE