Showing content from http://www.ncbi.nlm.nih.gov/IEB/ToolBox/CPP_DOC/doxyhtml/blastkmerutils_8cpp.html below:
NCBI C++ ToolKit: src/algo/blast/proteinkmer/blastkmerutils.cpp File Reference
Go to the source code of this file.
Go to the SVN repository for this file.
USING_SCOPE (objects) USING_SCOPE (blast) DEFINE_STATIC_MUTEX (randMutex) uint32_t uhash (uint64_t x, uint64_t a, uint64_t b) static uint32_t FNV_hash (uint32_t num) FNV hash, see http://www.isthe.com/chongo/tech/comp/fnv/index.html. More...
double estimate_jaccard (vector< uint32_t > &query_hash, vector< uint32_t > &subject, int num_hashes) double estimate_jaccard2 (vector< uint32_t > &query_hash, vector< uint32_t > &subject, int num_hashes) set< uint32_t > BlastKmerGetKmerSetStats (const string &query_sequence, int kmerNum, map< string, int > &kmerCount, map< string, int > &kmerCountPlus, int alphabetChoice, bool perQuery) Simplified version of BlastKmerGetKmerSet. More...
set< uint32_t > BlastKmerGetKmerSet (const string &query_sequence, bool do_seg, TSeqRange &range, int kmerNum, int alphabetChoice) Get KMERs for a given sequence using a compressed alphabet. More...
set< uint32_t > BlastKmerGetKmerSet2 (const string &query_sequence, TSeqRange &range, int kmerNum, int alphabetChoice, vector< int > badMers) Get KMERs for a given sequence using a compressed alphabet. More...
void BlastKmerGetCompressedTranslationTable (vector< Uint1 > &trans_table, int alphabetChoice) Creates translation table for compressed alphabets. More...
int BlastKmerBreakUpSequence (int length, vector< TSeqRange > &range_v, int ChunkSize) Breaks a sequences up into chunks if the sequences is above a certain length. More...
int BlastKmerGetDistance (const vector< uint32_t > &minhash1, const vector< uint32_t > &minhash2) Calculates the number of differences between two minhash arrays. More...
bool minhash_query (const string &query, vector< vector< uint32_t > > &seq_hash, int num_hashes, uint32_t *a, uint32_t *b, int do_seg, int kmerNum, int alphabetChoice, int chunkSize) bool minhash_query2 (const string &query, vector< vector< uint32_t > > &seq_hash, int kmerNum, int numHashes, int alphabetChoice, vector< int > badMers, int chunkSize) Hash the query for the minimum values;. More...
void get_LSH_hashes (vector< vector< uint32_t > > &query_hash, vector< vector< uint32_t > > &lsh_hash_vec, int num_bands, int rows_per_band) void get_LSH_hashes5 (vector< vector< uint32_t > > &query_hash, vector< vector< uint32_t > > &lsh_hash_vec, int numHashes, int numRows) Gets the LSH hash for one hash function. More...
void GetRandomNumbers (uint32_t *a, uint32_t *b, int numHashes) Get the random numbers for the hash function. More...
void GetKValues (vector< vector< int > > &kvector, int k_value, int l_value, int array_size) Function to get the k sites to compare for Buhler LSH. More...
void get_LSH_hashes2 (vector< vector< uint32_t > > &query_hash, vector< vector< uint32_t > > &lsh_hash_vec, int num_k, int num_l, vector< vector< int > > &kvector) void get_LSH_match_from_hash (const vector< vector< uint32_t > > &query_LSH_hash, const uint64_t *lsh_array, vector< set< uint32_t > > &candidates) void s_HashHashQuery (const vector< vector< uint32_t > > &query_hash, vector< vector< uint32_t > > &query_hash_hash, int compress, int version) void neighbor_query (const vector< vector< uint32_t > > &query_hash, const uint64_t *lsh, vector< set< uint32_t > > &candidates, CMinHashFile &mhfile, int num_hashes, int min_hits, double thresh, TBlastKmerPrelimScoreVector &score_vector, BlastKmerStats &kmer_stats, int kmerVer) static int s_BlastKmerVerifyVolume (CMinHashFile &mhfile, string &error_msg, int volume) int BlastKmerVerifyIndex (CRef< CSeqDB > seqdb, string &error_msg) ◆ BlastKmerBreakUpSequence() ◆ BlastKmerGetCompressedTranslationTable() void BlastKmerGetCompressedTranslationTable ( vector< Uint1 > & trans_table, int alphabetChoice ) ◆ BlastKmerGetDistance()
Calculates the number of differences between two minhash arrays.
Used to decide whether two arrays are similar enough. The assumption is made that both arrays are of the same size.
-
Parameters
-
minhash1 First array [in] minhash2 Second array [in]
-
Returns
-
distance.
Definition at line 399 of file blastkmerutils.cpp.
Referenced by s_MinhashSequences(), and s_MinhashSequences2().
◆ BlastKmerGetKmerSet()
Get KMERs for a given sequence using a compressed alphabet.
-
Parameters
-
query_sequence string with one sequence [in] do_seg Should the sequence be segged (not recommended) [in] range portion of sequence to be processed [in] kmerNum size of kmer [in] alphabetChoice 0 is 15 letter, 1 is 10 letter alphabet [in]
-
Returns
-
set of unsigned ints for the kmers.
Definition at line 206 of file blastkmerutils.cpp.
References BlastKmerGetCompressedTranslationTable(), BlastSeqLocFree(), free(), CRange_Base::GetFrom(), COpenRange< Position >::GetLength(), CRange_Base::GetTo(), i, set< Key, Compare >::insert(), malloc(), BlastSeqLoc::next, NULL, query, SegParametersFree(), SegParametersNewAa(), and SeqBufferSeg().
Referenced by minhash_query(), and s_MinhashSequences().
◆ BlastKmerGetKmerSet2()
Get KMERs for a given sequence using a compressed alphabet.
This version can read in overrepresented KMERs and extend them by one.
-
Parameters
-
query_sequence string with one sequence [in] range portion of sequence to be processed [in] kmerNum size of kmer [in] alphabetChoice 0 is 15 letter, 1 is 10 letter alphabet [in] badMers Overrepresented KMERs [in]
-
Returns
-
set of unsigned ints for the kmers.
Definition at line 270 of file blastkmerutils.cpp.
References BlastKmerGetCompressedTranslationTable(), free(), CRange_Base::GetFrom(), COpenRange< Position >::GetLength(), CRange_Base::GetTo(), i, set< Key, Compare >::insert(), malloc(), and query.
Referenced by minhash_query2(), and s_MinhashSequences2().
◆ BlastKmerGetKmerSetStats() ◆ BlastKmerVerifyIndex() ◆ DEFINE_STATIC_MUTEX() DEFINE_STATIC_MUTEX ( randMutex ) ◆ estimate_jaccard() double estimate_jaccard ( vector< uint32_t > & query_hash, vector< uint32_t > & subject, int num_hashes ) inline ◆ estimate_jaccard2() double estimate_jaccard2 ( vector< uint32_t > & query_hash, vector< uint32_t > & subject, int num_hashes ) inline ◆ FNV_hash() ◆ get_LSH_hashes() void get_LSH_hashes ( vector< vector< uint32_t > > & query_hash, vector< vector< uint32_t > > & lsh_hash_vec, int num_bands, int rows_per_band ) ◆ get_LSH_hashes2() void get_LSH_hashes2 ( vector< vector< uint32_t > > & query_hash, vector< vector< uint32_t > > & lsh_hash_vec, int num_k, int num_l, vector< vector< int > > & kvector ) ◆ get_LSH_hashes5() void get_LSH_hashes5 ( vector< vector< uint32_t > > & query_hash, vector< vector< uint32_t > > & lsh_hash_vec, int numHashes, int numRows ) ◆ get_LSH_match_from_hash() ◆ GetKValues() void GetKValues ( vector< vector< int > > & kvector, int k_value, int l_value, int array_size ) ◆ GetRandomNumbers() ◆ minhash_query()
Definition at line 415 of file blastkmerutils.cpp.
References a, b, set< Key, Compare >::begin(), BlastKmerBreakUpSequence(), BlastKmerGetKmerSet(), set< Key, Compare >::empty(), set< Key, Compare >::end(), i, query, and uhash().
Referenced by BOOST_AUTO_TEST_CASE(), and CBlastKmer::x_ProcessQuery().
◆ minhash_query2()
Hash the query for the minimum values;.
-
Parameters
-
query as a ASCII string [in] seq_hash hash values for all kmers [out] kmerNum number of letters in a KMER [in] numHashes number of hashes in a signature [in] alphabetChoice 15 or 10 letters [in] badMers Overrepresented KMERS [in]
Definition at line 479 of file blastkmerutils.cpp.
References set< Key, Compare >::begin(), BlastKmerBreakUpSequence(), BlastKmerGetKmerSet2(), set< Key, Compare >::empty(), set< Key, Compare >::end(), FNV_hash(), i, query, and ct::sort().
Referenced by BOOST_AUTO_TEST_CASE(), and CBlastKmer::x_ProcessQuery().
◆ neighbor_query()
Definition at line 744 of file blastkmerutils.cpp.
References map_checker< Container >::begin(), map_checker< Container >::end(), estimate_jaccard(), estimate_jaccard2(), CMinHashFile::GetDataWidth(), CMinHashFile::GetHits(), CMinHashFile::GetMinHits(), CMinHashFile::GetVersion(), BlastKmerStats::hit_count, i, map_checker< Container >::insert(), int, BlastKmerStats::jd_count, BlastKmerStats::jd_oid_count, n, offset, BlastKmerStats::oids_considered, read_size(), s_HashHashQuery(), ct::sort(), and BlastKmerStats::total_matches.
Referenced by CBlastKmer::x_RunKmerFile().
◆ s_BlastKmerVerifyVolume() ◆ s_HashHashQuery() void s_HashHashQuery ( const vector< vector< uint32_t > > & query_hash, vector< vector< uint32_t > > & query_hash_hash, int compress, int version ) ◆ uhash() ◆ USING_SCOPE() [1/2] ◆ USING_SCOPE() [2/2]
RetroSearch is an open source project built by @garambo
| Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4