comparison CSP2/CSP2_env/env-d9b9114564458d9d-741b3de822f2aaca6c6caa4325c4afce/include/mash/MinHashHeap.h @ 69:33d812a61356

planemo upload commit 2e9511a184a1ca667c7be0c6321a36dc4e3d116d
author jpayne
date Tue, 18 Mar 2025 17:55:14 -0400
parents
children
comparison
equal deleted inserted replaced
67:0e9998148a16 69:33d812a61356
1 #ifndef HashHeapCounted_h
2 #define HashHeapCounted_h
3
4 #include "HashList.h"
5 #include "HashPriorityQueue.h"
6 #include "HashSet.h"
7 #include <math.h>
8 #include "bloom_filter.hpp"
9
10 class MinHashHeap
11 {
12 public:
13
14 MinHashHeap(bool use64New, uint64_t cardinalityMaximumNew, uint64_t multiplicityMinimumNew = 1, uint64_t memoryBoundBytes = 0);
15 ~MinHashHeap();
16 void computeStats();
17 void clear();
18 double estimateMultiplicity() const;
19 double estimateSetSize() const;
20 void toCounts(std::vector<uint32_t> & counts) const;
21 void toHashList(HashList & hashList) const;
22 void tryInsert(hash_u hash);
23
24 private:
25
26 bool use64;
27
28 HashSet hashes;
29 HashPriorityQueue hashesQueue;
30
31 HashSet hashesPending;
32 HashPriorityQueue hashesQueuePending;
33
34 uint64_t cardinalityMaximum;
35 uint64_t multiplicityMinimum;
36
37 uint64_t multiplicitySum;
38
39 bloom_filter * bloomFilter;
40
41 uint64_t kmersTotal;
42 uint64_t kmersUsed;
43 };
44
45 inline double MinHashHeap::estimateMultiplicity() const {return hashes.size() ? (double)multiplicitySum / hashes.size() : 0;}
46 inline double MinHashHeap::estimateSetSize() const {return hashes.size() ? pow(2.0, use64 ? 64.0 : 32.0) * (double)hashes.size() / (use64 ? (double)hashesQueue.top().hash64 : (double)hashesQueue.top().hash32) : 0;}
47 inline void MinHashHeap::toHashList(HashList & hashList) const {hashes.toHashList(hashList);}
48 inline void MinHashHeap::toCounts(std::vector<uint32_t> & counts) const {hashes.toCounts(counts);}
49
50 #endif