vector<SBlastDbMaskData> & mask_ranges,
98vector<int> & column_ids,
99vector<CTempString> & column_blobs) = 0;
101 #if ((!defined(NCBI_COMPILER_WORKSHOP) || (NCBI_COMPILER_VERSION > 550)) && \ 102 (!defined(NCBI_COMPILER_MIPSPRO)) ) 160 const string& title,
165 boollong_seqids =
false,
167 boollimit_defline =
false,
169 boolscan_bioseq_4_cfastareader_usrobj =
true);
196 const string& title,
202 boollong_seqids =
false,
204 boollimit_defline =
false,
206 boolscan_bioseq_4_cfastareader_usrobj =
true);
228 voidSetMaskLetters(
const string& mask_letters);
237 voidSetSourceDb(
const string& src_db_name);
285 boolBuild(
constvector<string> & ids,
312 boolAddIds(
constvector<string> & ids);
335 boolAddSequences(
IBioseqSource& src,
booladd_pig =
false);
356 boolEndBuild(
boolerase =
false);
387m_UseRemote = use_remote;
398m_SkipCopyingGis = v;
409 voidSetMaxFileSize(
Uint8max_file_size);
427 const string& options,
428 const string& name =
"");
444RegisterMaskingAlgorithm(
const string& program,
445 const string& description,
446 const string& options);
463 static voidCreateDirectories(
const string&
dbname);
468objects::CScope & x_GetScope();
526objects::CSeqVector * sv,
527 booladd_pig =
false);
549 boolx_ReportUnresolvedIds(
const CInputGiList& gi_list)
const;
574 voidx_AddOneRemoteSequence(
constobjects::CSeq_id & seqid,
594 boolx_EndBuild(
boolerase,
const CException* close_exception);
Build BlastDB format databases from various data sources.
bool m_IsProtein
True for a protein database, false for nucleotide.
int m_DeflineCount
Define count.
void x_SetLeafTaxids(CRef< objects::CBlast_def_line_set > headers)
Store leaf taxids in provided headers.
bool m_KeepLinks
True to keep linkout bits from source dbs, false to discard.
bool m_SkipLargeGis
If set to true, skip GIs with value > 0x7FFFFFFF.
TIdToBits m_Id2Links
Table of linkout bits to apply to sequences.
bool m_FoundMatchingMasks
If true, there were sequences whose IDs matched those in the provided masking locations (via SetMaskD...
void SetVerbosity(bool v)
Specify level of output verbosity.
bool m_UseRemote
Whether to use remote resolution and sequence fetching.
bool m_ParseIDs
If true, string IDs found in FASTA input will be parsed as Seq-ids.
int m_OIDCount
Number of OIDs stored in this database.
void SetUseRemote(bool use_remote)
Specify whether to use remote fetching for locally absent IDs.
bool m_ScanBioseq4CFastaReaderUsrObjct
bool m_KeepMbits
True to keep membership bits from source dbs, false to discard.
CRef< CTaxIdSet > m_Taxids
Set of TaxIDs configured to apply to sequences.
CRef< objects::CScope > m_Scope
Sequence scope, used for remote fetching.
CRef< objects::CObjectManager > m_ObjMgr
Object manager, used for remote fetching.
ostream & m_LogFile
Logfile.
bool m_Verbose
If true, more detailed log messages will be produced.
bool m_KeepLeafs
True to keep leaf taxids from source dbs, false to discard.
bool m_SkipCopyingGis
If set to true, when copying BLASTDBs, skip the GIs.
CRef< CWriteDB > m_OutputDb
Database being produced here.
CRef< CSeqDBExpert > m_SourceDb
Database for duplicating sequences locally (-sourcedb option.)
CRef< IMaskDataSource > m_MaskData
Subject masking data.
bool m_LongIDs
If true, use long sequence ids (database|accession)
TIdToLeafs m_Id2Leafs
Table of leaf taxids to apply to sequences.
string GetOutputDbName() const
TIdToBits m_Id2Mbits
Table of membership bits to apply to sequences.
void SetSkipCopyingGis(bool v)
Gi List for database construction.
This represents a set of masks for a given sequence.
CTempString implements a light-weight string on top of a storage buffer whose lifetime management is ...
int TIndexType
Bitwise OR of "EIndexType".
Interface to a source of Bioseq objects.
virtual CConstRef< objects::CBioseq > GetNext()=0
Get a Bioseq object if there are any more to get.
An interface providing lookups of mask-data by Seq-id.
virtual CMaskedRangesVector & GetRanges(const list< CRef< CSeq_id > > &id)=0
Get ranges of masking data for the given Seq-ids.
Interface to a source of raw sequence data.
virtual const map< string, string > & GetColumnMetaData(int id)=0
Get metadata for the column with the specified Column ID.
virtual bool GetNext(CTempString &sequence, CTempString &ambiguities, CRef< objects::CBlast_def_line_set > &deflines, vector< SBlastDbMaskData > &mask_ranges, vector< int > &column_ids, vector< CTempString > &column_blobs)=0
Get a raw sequence.
virtual void GetColumnNames(vector< string > &names)=0
Get the names of all columns defined by this sequence source.
virtual int GetColumnId(const string &name)=0
Get the column ID for a column mentioned by name.
Include a standard set of the NCBI C++ Toolkit most basic headers.
static const struct name_t names[]
uint64_t Uint8
8-byte (64-bit) unsigned integer
#define END_NCBI_SCOPE
End previously defined NCBI scope.
#define BEGIN_NCBI_SCOPE
Define ncbi namespace.
IO_PREFIX::istream CNcbiIstream
Portable alias for istream.
#define NCBI_XOBJWRITE_EXPORT
EBlast_filter_program
This defines the possible sequence filtering algorithms to be used in a BLAST database.
char * dbname(DBPROCESS *dbproc)
Get name of current database.
EBlastDbVersion
BLAST database version.
Defines `expert' version of CSeqDB interfaces.
Class which defines sequence id to taxid mapping.
Defines BLAST database construction classes.
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