CombinationGenerator< T > Class Template Reference

#include <CombinationGenerator.h>

Public Types
typedef std::vector< T >	Collection
typedef std::vector< Collection >	Combination
typedef std::vector< Combination >	VectorOfCombinations

Public Member Functions
std::vector< Combination >	combinations (const Collection &data, int numberOfCollections) const
std::vector< Combination >	combinations (const Collection &data, const PartitionGenerator::Partition &partition) const
std::vector< Combination >	combinations (const Collection &data) const

Private Member Functions
std::vector< Combination >	separateOneElement (const Collection &data) const
VectorOfCombinations	splitInTwoCollections (const Collection &data, int sizeFirst) const

Detailed Description

template<class T>
class CombinationGenerator< T >

Class to compute all distinct Combinations of a collection 'data' of objects of type 'T'. A Combination is a set of collections, each collection containing one or more objects, with any object in 'data' assigned to exactly one collection.

Definition at line 18 of file CombinationGenerator.h.

Member Typedef Documentation

Collection

template<class T >

typedef std::vector<T> CombinationGenerator< T >::Collection

Definition at line 20 of file CombinationGenerator.h.

Combination

template<class T >

typedef std::vector<Collection> CombinationGenerator< T >::Combination

Definition at line 22 of file CombinationGenerator.h.

VectorOfCombinations

template<class T >

typedef std::vector<Combination> CombinationGenerator< T >::VectorOfCombinations

Definition at line 23 of file CombinationGenerator.h.

Member Function Documentation

combinations() [1/3]

template<class T >

std::vector<Combination> CombinationGenerator< T >::combinations ( const Collection &  data, int  numberOfCollections  ) const

inline

Create combinations obtained by dividing 'data' according to all partitions with 'numberOfCollections' collections.

Definition at line 28 of file CombinationGenerator.h.

References SplitLinear::begin, data, l1ctLayer1_patternWriters_cff::partition, jetUpdater_cfi::sort, and PartitionGenerator::sortedPartitions().

Referenced by CombinationGenerator< T >::combinations(), CombinationGenerator< T >::separateOneElement(), and CombinationGenerator< T >::splitInTwoCollections().

                                                                                             {
    std::vector<Combination> combinations;
    Collection sorted = data;

    // Sort if needed
    if (prev_permutation(sorted.begin(), sorted.end())) {
      sort(sorted.begin(), sorted.end());
    }

    // Create sorted partitions
    PartitionGenerator aPartitionGenerator;
    std::vector<std::vector<PartitionGenerator::Partition> > partitions =
        aPartitionGenerator.sortedPartitions(data.size());

    // Use partitions of size 'numberOfCollections' only
    for (std::vector<PartitionGenerator::Partition>::const_iterator idiv = partitions[numberOfCollections - 1].begin();
         idiv != partitions[numberOfCollections - 1].end();
         idiv++) {
      const PartitionGenerator::Partition& partition = *idiv;
      std::vector<Combination> subCombinations = this->combinations(data, partition);
      for (typename std::vector<Combination>::const_iterator iComb = subCombinations.begin();
           iComb != subCombinations.end();
           iComb++) {
        combinations.push_back(*iComb);
      }
    }
    return combinations;
  }

combinations() [2/3]

template<class T >

std::vector<Combination> CombinationGenerator< T >::combinations ( const Collection &  data, const PartitionGenerator::Partition &  partition  ) const

inline

Create all combinations obtained by dividing 'data' according to Partition 'partition'.

Definition at line 60 of file CombinationGenerator.h.

References bookConverter::comb, CombinationGenerator< T >::combinations(), data, ALPAKA_ACCELERATOR_NAMESPACE::vertexFinder::it, isotrackApplyRegressor::k, isotrackNtupler::merged, l1ctLayer1_patternWriters_cff::partition, findQualityFiles::size, jetUpdater_cfi::sort, and CombinationGenerator< T >::splitInTwoCollections().

                                                                                                                  {
    std::vector<Combination> combinations;

    // Check that sum of collection sizes in 'partition'
    // amounts to number of elements in 'data'
    int nElts = 0;
    for (PartitionGenerator::Partition::const_iterator iSize = partition.begin(); iSize != partition.end(); iSize++) {
      nElts += *iSize;
    }
    if (nElts != data.size())
      return combinations;

    Collection sortedData = data;
    // Sort if needed
    if (prev_permutation(sortedData.begin(), sortedData.end())) {
      sort(sortedData.begin(), sortedData.end());
    }

    // Create initial combination and put it on the stack
    Combination comb;
    comb.push_back(sortedData);
    if (partition.size() == 1) {
      // Return only one combination with only one collection
      combinations.push_back(comb);
      return combinations;
    }
    std::stack<Combination> cStack;
    cStack.push(comb);

    // Sort partitions by size
    // Let 'sortedPartition' = ( n0, n1,... nk, nk+1,... nm )
    // Ordering is >= to speed up partitioning: n0 >= n1 >=... >= nm
    PartitionGenerator::Partition sortedPartition = partition;
    sort(sortedPartition.begin(), sortedPartition.end(), std::greater_equal<int>());

    while (!cStack.empty()) {
      // 'combination' popped out of the stack
      Combination combination = cStack.top();
      cStack.pop();

      // At this stage 'combination' is made of collections
      // of sizes ( n0+n1+...+nk, nk+1,... nm )
      // Now generate all combinations obtained by splitting
      // the first collection of 'combination' in two,
      // according to Partition 'biPartition' = (n0+n1+...+nk-1, nk)
      int k = sortedPartition.size() - combination.size();
      int size = 0;
      for (int iColl = 0; iColl != k; iColl++) {
        size += sortedPartition[iColl];
      }
      PartitionGenerator::Partition biPartition(2);
      biPartition[0] = size;
      biPartition[1] = sortedPartition[k];

      VectorOfCombinations subCombinations = splitInTwoCollections(combination[0], biPartition[0]);
      for (typename std::vector<Combination>::const_iterator iComb = subCombinations.begin();
           iComb != subCombinations.end();
           iComb++) {
        // Check ordering of successive bins of equal size
        if (combination.size() > 1) {
          if ((*iComb)[1].size() == combination[1].size()) {
            Collection adjBins;
            adjBins.push_back((*iComb)[1][0]);
            adjBins.push_back(combination[1][0]);
            // Drop 'combination' if successive bins of equal size not ordered
            // i.e. if previous permutation exists
            if (prev_permutation(adjBins.begin(), adjBins.end()))
              continue;
          }
        }

        // Append remaining collections of 'combination'
        Combination merged = *iComb;
        typename Combination::const_iterator it = combination.begin();
        it++;
        for (; it != combination.end(); it++) {
          merged.push_back(*it);
        }

        // Store combination 'merged' if partitioning is complete,
        // otherwise put it on the stack for further partitioning.
        if (merged.size() == sortedPartition.size()) {
          combinations.push_back(merged);
        } else {
          cStack.push(merged);
        }
      }
    }
    return combinations;
  }

combinations() [3/3]

template<class T >

std::vector<Combination> CombinationGenerator< T >::combinations ( const Collection &  data ) const

inline

Create all combinations of elements from 'data'.

Definition at line 153 of file CombinationGenerator.h.

References CombinationGenerator< T >::combinations(), data, l1ctLayer1_patternWriters_cff::partition, PartitionGenerator::partitions(), and jetUpdater_cfi::sort.

                                                                    {
    std::vector<Combination> combinations;
    Collection sorted = data;
    // Sort if needed
    if (prev_permutation(sorted.begin(), sorted.end())) {
      sort(sorted.begin(), sorted.end());
    }

    PartitionGenerator aPartitionGenerator;
    std::vector<PartitionGenerator::Partition> partitions = aPartitionGenerator.partitions(data.size());

    for (std::vector<PartitionGenerator::Partition>::const_iterator idiv = partitions.begin(); idiv != partitions.end();
         idiv++) {
      const PartitionGenerator::Partition& partition = *idiv;

      std::vector<Combination> subCombinations = this->combinations(data, partition);
      for (typename std::vector<Combination>::const_iterator iComb = subCombinations.begin();
           iComb != subCombinations.end();
           iComb++) {
        combinations.push_back(*iComb);
      }
    }
    return combinations;
  }

separateOneElement()

template<class T >

std::vector<Combination> CombinationGenerator< T >::separateOneElement ( const Collection &  data ) const

inlineprivate

Create all combinations obtained by dividing 'data' in two collections, the second one having only one element.

Definition at line 239 of file CombinationGenerator.h.

References bookConverter::comb, CombinationGenerator< T >::combinations(), data, mps_fire::i, dqmiolumiharvest::j, and single().

Referenced by CombinationGenerator< T >::splitInTwoCollections().

                                                                          {
    std::vector<Combination> combinations;
    for (typename Collection::const_iterator i = data.begin(); i != data.end(); i++) {
      Combination comb;
      Collection single;
      single.push_back(*i);
      Collection coll;
      typename Collection::const_iterator j = data.begin();
      for (; j != i; j++) {
        coll.push_back(*j);
      }
      j++;
      for (; j != data.end(); j++) {
        coll.push_back(*j);
      }
      comb.push_back(coll);
      comb.push_back(single);
      combinations.push_back(comb);
    }
    return combinations;
  }

splitInTwoCollections()

template<class T >

VectorOfCombinations CombinationGenerator< T >::splitInTwoCollections ( const Collection &  data, int  sizeFirst  ) const

inlineprivate

Create all combinations obtained by dividing 'data' in two collections, the first one being of size 'sizeFirst'

Definition at line 182 of file CombinationGenerator.h.

References universalConfigTemplate::collection, bookConverter::comb, CombinationGenerator< T >::combinations(), data, edm::second(), CombinationGenerator< T >::separateOneElement(), and findQualityFiles::size.

Referenced by CombinationGenerator< T >::combinations().

                                                                                          {
    std::vector<Combination> combinations;
    std::stack<Combination> cStack;

    // Create first combination with 2 partitions
    Combination comb;
    comb.push_back(data);
    comb.push_back(Collection());
    cStack.push(comb);

    while (!cStack.empty()) {
      Combination combination = cStack.top();
      cStack.pop();

      Collection collection = combination[0];
      std::vector<Combination> subCombinations = separateOneElement(collection);

      for (typename std::vector<Combination>::const_iterator iComb = subCombinations.begin();
           iComb != subCombinations.end();
           iComb++) {
        Collection second = combination[1];
        second.push_back((*iComb)[1][0]);

        // Abandon current combination if not ordered,
        // i.e. if previous permutation exists
        bool ordered = !prev_permutation(second.begin(), second.end());
        if (!ordered)
          continue;
        next_permutation(second.begin(), second.end());

        if ((*iComb)[0].size() == second.size()) {
          Collection adjBins;
          adjBins.push_back((*iComb)[0][0]);
          adjBins.push_back(second[0]);
          // Abandon current combination if successive bins of equal size
          // not ordered, i.e. if previous permutation exists
          if (prev_permutation(adjBins.begin(), adjBins.end()))
            continue;
        }

        Combination stored;
        stored.push_back((*iComb)[0]);
        stored.push_back(second);

        if (stored[0].size() == sizeFirst) {
          combinations.push_back(stored);
        } else {
          cStack.push(stored);
        }
      }
    }
    return combinations;
  }