Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from http://www.ncbi.nlm.nih.gov/IEB/ToolBox/CPP_DOC/doxyhtml/makeblastdb_8cpp_source.html below:

NCBI C++ ToolKit: src/app/blastdb/makeblastdb.cpp Source File

();

vector<ESupportedInputFormats>

vector<string>& blastdbs);

<

TObj>

obj.

(

TObj);

<

TObj>

.GetFormatHints().DisableAllNonpreferred();

arg_desc->SetUsageContext(

().GetProgramBasename(),

dflt(

);

dflt +=

+

;

arg_desc->SetCurrentGroup(

);

arg_desc->AddDefaultKey(

,

,

,

arg_desc->AddDefaultKey(

,

,

,

,

,

arg_desc->AddKey(

,

,

arg_desc->SetCurrentGroup(

);

arg_desc->AddOptionalKey(

,

,

+ dflt,

arg_desc->AddFlag(

,

,

);

arg_desc->AddFlag(

,

,

arg_desc->AddFlag(

,

arg_desc->SetCurrentGroup(

);

arg_desc->AddOptionalKey(

,

,

arg_desc->AddOptionalKey(

,

,

,

arg_desc->AddOptionalKey(

,

,

,

arg_desc->AddFlag(

,

,

);

arg_desc->AddOptionalKey(

,

,

,

arg_desc->SetCurrentGroup(

);

arg_desc->AddOptionalKey(

,

,

+ dflt +

,

arg_desc->AddDefaultKey(

,

,

,

arg_desc->SetConstraint(

,

arg_desc->AddDefaultKey(

,

,

,

arg_desc->AddOptionalKey(

,

,

arg_desc->AddOptionalKey(

,

,

,

arg_desc->AddFlag(

,

,

);

arg_desc->AddFlag(

,

,

);

arg_desc->SetCurrentGroup(

);

arg_desc->AddOptionalKey(

,

,

,

arg_desc->AddOptionalKey(

,

,

,

arg_desc->AddOptionalKey(

,

,

kMods =

;

(

= 0;

< kMods.size();

++) {

(v <

(minprec)*1024) {

rv.append(kMods,

, 1);

(p_ptr !=

)

(desc.

()) {

unv(

);

(ch ==

|| ch ==

|| ch ==

|| ch ==

)) {

(ch ==

|| ch ==

) {

(it_bio_set =

(*

); it_bio_set; ++it_bio_set) {

(it_bio_set->CanGetDescr() &&

(it_bio_set->GetDescr())) {

(it_bio =

(*it_bio_set); it_bio; ++it_bio) {

(!

)

rv;

sequence::CDeflineGenerator::TUserFlags

= sequence::CDeflineGenerator::fUseAutoDef;

sequence::CDeflineGenerator

;

(

->IsTitle()) {

old_title =

->GetTitle();

->SetTitle(title);

desr.

().push_back(

);

->SetOrg().SetTaxId(taxid);

(!

.eof())

vector<SBlastDbMaskData> & mask_range,

vector<int> & column_ids,

vector<CTempString> & column_blobs);

: m_Sequence(

), m_Oid(0)

algo_opts, algo_name;

std::mt19937 rng(std::time(

));

vector<SBlastDbMaskData> & mask_range,

vector<int> & column_ids,

vector<CTempString> & column_blobs)

slength(0), alength(0);

mask_data.

.push_back(pair<TSeqPos, TSeqPos>(range->first, range->second));

mask_range.push_back(mask_data);

column_blobs.resize(column_ids.size());

.resize(column_ids.size());

(

= 0;

< (

)column_ids.size();

++) {

(args[

].

()) {

& input_type = args[

].AsString();

(input_type ==

) {

}

(input_type ==

) {

}

(input_type ==

) {

}

(input_type ==

) {

error_msg =

+ seq_file +

;

error_msg =

+ seq_file +

;

unique_ptr<CDecompressIStream> comp_seq_istream;

is_gz =

, is_bz2 =

, is_zst=

;

data_in = comp_seq_istream.get();

=

+ seq_file +

\

+

;

.GetFormatHints().DisableAllNonpreferred();

.GuessFormat();

(

);

+=

;

& files = args[

];

& ids = args[

];

& descs = args[

];

& gi_names = args[

];

vector<string> file_list;

vector<string> id_list;

vector<string> desc_list;

vector<string> gi_mask_names;

(! file_list.size()) {

);

(file_list.size() != id_list.size()) {

);

(

= 0;

< id_list.size(); ++

) {

);

(file_list.size() != desc_list.size()) {

);

vector<string> default_desc(id_list.size(),

);

desc_list.swap(default_desc);

(file_list.size() != gi_mask_names.size()) {

);

(

= 0;

< file_list.size(); ++

) {

( !

(file_list[

]).Exists() ) {

<<

);

.GetFormatHints().DisableAllNonpreferred();

mask_file_format =

.GuessFormat();

(mask_file, mask_file_format, first_obj,

+ file_list[

] +

);

*

<<

<< file_list[

] << endl;

opts = first_obj->GetAlgo_options();

(id_list.size()) {

name = gi_mask_names.size() ? gi_mask_names[

] : file_list[

];

(seqid.

()) {

);

(mask_file, mask_file_format, masks,

);

(

!= args[

].AsString())

(args[

])

vector<CTempString>

;

vector<string> blastdb;

(

[0] ==

) {

(

= 0;

<

.size();

++) {

seq_file =

[

];

unique_ptr<CDecompressIStream> comp_seq_istream;

is_gz =

, is_bz2 =

, is_zst=

;

data_in = comp_seq_istream.get();

vector<string> blastdb;

randomize = args[

].AsBoolean();

vector<string> final_blastdb;

(blastdb.size()==1);

db(blastdb[0], seqtype);

vector<string> paths;

(paths.size() > 1) {

);

final_blastdb.push_back(blastdb[0]);

(vector<string>, iter, blastdb) {

& s = *iter;

<< s <<

);

final_blastdb.push_back(s);

(final_blastdb.size()) {

(!leafs.

()) {

bdl->GetSeqid().front()->AsFastaString();

leafTaxIds[id] = ids;

);

env_skip =

dont_scan_bioseq =

().

(

);

scan_bioseq_4_cfastareader_usrobj =

(dont_scan_bioseq.empty());

is_protein = (args[

].AsString() ==

);

: args[

]).AsString();

(input_fmt !=

&& args[

].AsBoolean()) {

);

vector<string> input_files;

(

==

|| input_files.size() > 1) {

+

);

: args[

]).AsString();

vector<CTempString>

;

(

.size() > 1) {

);

(title ==

) {

);

= & (args[

].HasValue()

? args[

].AsOutputFile()

hash_index = args[

];

use_gi_mask = args[

];

long_seqids =

;

long_seqids = (

.

(

,

) ==

);

limit_defline =

;

(args[

]) {

limit_defline =

;

oid_masks = 0;

(args[

]) {

oid_masks = args[

].AsInteger();

scan_bioseq_4_cfastareader_usrobj));

(args[

]) {

MAX_VOL_FILE_SIZE = 0x100000000;

(bytes >= MAX_VOL_FILE_SIZE) {

);

(args[

].

()) {

}

(args[

].

()) {

output_prefix = args[

]

? args[

].AsString()

metadata_filename = new_db +

+ extn;

ofstream

(metadata_filename.c_str());

json_out->PreserveKeyNames();

json_out->WriteObject(obj_info);

(args[

].

()) {

(args[

].

()) {

(args[

].

()) {

(args[

].

()) {

(

argc,

* argv[]

)

User-defined methods of the data storage class.

User-defined methods of the data storage class.

Declares singleton objects to store the version and reference for the BLAST engine.

#define CATCH_ALL(exit_code)

Standard catch statement for all BLAST command line programs.

Interface for converting sources of sequence data into blast sequence input.

Auxiliary classes/functions for BLAST input library.

Code to build a database given various sources of sequence data.

Class to constrain the values of an argument to those greater than or equal to the value specified in...

CSeq_entry * GetParentEntry(void) const

TTaxId GetTaxId() const

Determine the tax-id for this bioseq.

void AddParam(EUsageParams p, int val)

Keeps track of the version of the BLAST engine in the NCBI C++ toolkit.

Build BlastDB format databases from various data sources.

bool AddSequences(IBioseqSource &src, bool add_pig=false)

Add sequences from an IBioseqSource object.

bool AddFasta(CNcbiIstream &fasta_file)

Add sequences from a file containing FASTA data.

void SetVerbosity(bool v)

Specify level of output verbosity.

void SetMaskDataSource(IMaskDataSource &ranges)

Specify an object mapping Seq-id to subject masking data.

void SetMembBits(const TLinkoutMap &membbits, bool keep_mbits)

Specify a membership bit lookup object.

int RegisterMaskingAlgorithm(EBlast_filter_program program, const string &options, const string &name="")

Define a masking algorithm.

void SetTaxids(CTaxIdSet &taxids)

Specify a mapping of sequence ids to taxonomic ids.

string GetOutputDbName() const

bool EndBuild(bool erase=false)

Finish building a new database.

void SetMaxFileSize(Uint8 max_file_size)

Set the maximum size of database component files.

void SetLeafTaxIds(const TIdToLeafs &taxids, bool keep_taxids)

Specify a leaf-taxids object.

Class implements different ad-hoc unreliable file format identifications.

EFormat

The formats are checked in the same order as declared here.

@ eBinaryASN

Binary ASN.1.

@ eFasta

FASTA format sequence record, CFastaReader.

@ eUnknown

unknown format

static const char * GetFormatName(EFormat format)

Defines invalid user input exceptions.

The main application class.

CRef< CMaskedRangeSet > m_Ranges

void x_ProcessInputData(const string &paths, bool is_protein)

void x_AddSeqEntries(CNcbiIstream &data, TFormat fmt)

CFormatGuess::EFormat TFormat

Convenience typedef.

void x_AddFasta(CNcbiIstream &data)

void x_AddSequenceData(CNcbiIstream &input, TFormat fmt)

virtual void Init()

@inheritDoc

CRef< CBuildDatabase > m_DB

bool x_ShouldParseSeqIds(void)

CMakeBlastDBApp()

@inheritDoc

virtual int Run()

@inheritDoc

ESupportedInputFormats x_GetUserInputTypeHint(void)

CBlastUsageReport m_UsageReport

TFormat x_ConvertToCFormatGuessType(ESupportedInputFormats fmt)

ESupportedInputFormats x_ConvertToSupportedType(TFormat fmt)

vector< ESupportedInputFormats > x_GuessInputType(const vector< CTempString > &filenames, vector< string > &blastdbs)

void x_VerifyInputFilesType(const vector< CTempString > &filenames, CMakeBlastDBApp::ESupportedInputFormats input_type)

TFormat x_GuessFileType(CNcbiIstream &input)

void Insert(int algo_id, const CSeq_id &id, const CSeq_loc &v)

static CNcbiApplication * Instance(void)

Singleton method.

virtual ~CRawSeqDBSource()

virtual int GetColumnId(const string &name)

Get the column ID for a column mentioned by name.

virtual const map< string, string > & GetColumnMetaData(int id)

Get metadata for the column with the specified Column ID.

CRef< CSeqDBExpert > m_Source

vector< CBlastDbBlob > m_Blobs

vector< int > m_RandomOids

virtual bool GetNext(CTempString &sequence, CTempString &ambiguities, CRef< CBlast_def_line_set > &deflines, vector< SBlastDbMaskData > &mask_range, vector< int > &column_ids, vector< CTempString > &column_blobs)

vector< string > m_ColumnNames

CRawSeqDBSource(const string &name, bool protein, CBuildDatabase *outdb, bool randomize)

vector< int > m_ColumnIds

virtual void GetColumnNames(vector< string > &names)

Get the names of all columns defined by this sequence source.

map< int, int > m_MaskIdMap

void GetRawSeqAndAmbig(int oid, const char **buffer, int *seq_length, int *ambig_length) const

Raw Sequence and Ambiguity Data.

void GetColumnBlob(int col_id, int oid, CBlastDbBlob &blob)

Fetch the data blob for the given column and oid.

static void FindVolumePaths(const string &dbname, ESeqType seqtype, vector< string > &paths, vector< string > *alias_paths=NULL, bool recursive=true, bool expand_links=true)

Find volume paths.

int GetNumOIDs() const

Returns the size of the (possibly sparse) OID range.

void GetAvailableMaskAlgorithms(vector< int > &algorithms)

Get a list of algorithm IDs for which mask data exists.

ESeqType

Sequence types (eUnknown tries protein, then nucleotide).

void GetMaskAlgorithmDetails(int algorithm_id, objects::EBlast_filter_program &program, string &program_name, string &algo_opts)

Get information about one type of masking available here.

void RetSequence(const char **buffer) const

Returns any resources associated with the sequence.

string GetTitle() const

Returns the database title.

CRef< CBlast_db_metadata > GetDBMetaData(string user_path=kEmptyStr)

void ListColumns(vector< string > &titles)

List columns titles found in this database.

bool CheckOrFindOID(int &next_oid) const

Find an included OID, incrementing next_oid if necessary.

int GetColumnId(const string &title)

Get an ID number for a given column title.

CRef< CBlast_def_line_set > GetHdr(int oid) const

Get the ASN.1 header for the sequence.

void GetMaskData(int oid, const vector< int > &algo_ids, TSequenceRanges &ranges)

Get masked ranges of a sequence.

const map< string, string > & GetColumnMetaData(int column_id)

Get all metadata for the specified column.

virtual CConstRef< CBioseq > GetNext()

Get a Bioseq object if there are any more to get.

CRef< CObjectManager > m_objmgr

set< CBioseq * > m_bio_skipped

CTypeIterator< CBioseq > m_bio

CSeqEntrySource(CNcbiIstream &is, TFormat fmt, bool skip_unver)

CRef< CSeq_entry > m_entry

CFormatGuess::EFormat TFormat

Convenience typedef.

bool IsUnverified(const CSeq_descr &descr)

@Seq_descr.hpp User-defined methods of the data storage class.

const CSeq_descr & GetDescr(void) const

bool IsSetDescr(void) const

CSeq_entry * GetParentEntry(void) const

void SetMappingFromFile(CNcbiIstream &f)

CTempString implements a light-weight string on top of a storage buffer whose lifetime management is ...

Template class for iteration on objects of class C.

@ eFullIndex

Use several forms of each Seq-id in the string index.

@ eAddHash

Add an index from sequence hash to OID.

@ eNoIndex

Build a database without any indices.

int TIndexType

Bitwise OR of "EIndexType".

Interface to a source of Bioseq objects.

Interface to a source of raw sequence data.

iterator_bool insert(const value_type &val)

const_iterator begin() const

const_iterator find(const key_type &key) const

const_iterator end() const

const string kArgDbTitle

Title for the BLAST database.

const string kArgDbType

BLAST database molecule type.

void Print(const CCompactSAMApplication::AlignInfo &ai)

API (CDeflineGenerator) for computing sequences' titles ("definitions").

std::ofstream out("events_result.xml")

main entry point for tests

static const struct name_t names[]

static const char * str(char *buf, int n)

static void des(const char *src, const char *out)

void SetFullVersion(CRef< CVersionAPI > version)

Set version data for the program.

void HideStdArgs(THideStdArgs hide_mask)

Set the hide mask for the Hide Std Flags.

const CNcbiEnvironment & GetEnvironment(void) const

Get the application's cached environment.

const CNcbiRegistry & GetConfig(void) const

Get the application's cached configuration parameters (read-only).

virtual const CArgs & GetArgs(void) const

Get parsed command line arguments.

int AppMain(int argc, const char *const *argv, const char *const *envp=0, EAppDiagStream diag=eDS_Default, const char *conf=NcbiEmptyCStr, const string &name=NcbiEmptyString)

Main function (entry point) for the NCBI application.

CVersionInfo GetVersion(void) const

Get the program version information.

virtual void SetupArgDescriptions(CArgDescriptions *arg_desc)

Setup the command line argument descriptions.

#define ITERATE(Type, Var, Cont)

ITERATE macro to sequence through container elements.

SStrictId_Tax::TId TTaxId

Taxon id type.

const CNcbiArguments & GetArguments(void) const

Get the application's cached unprocessed command-line arguments.

#define TAX_ID_FROM(T, value)

@ fHideXmlHelp

Hide XML help description.

@ fHideFullVersion

Hide full version description.

@ fHideDryRun

Hide dryrun description.

@ fHideConffile

Hide configuration file description.

@ eNoOwnership

No ownership is assumed.

virtual const string & AsString(void) const =0

Get the argument's string value.

virtual bool HasValue(void) const =0

Check if argument holds a value.

@ fAppend

Append to end-of-file; for eOutputFile or eIOFile.

@ fHidden

Hide it in Usage.

@ eRequires

One argument requires another.

@ eExcludes

One argument excludes another.

@ eInputFile

Name of file (must exist and be readable)

@ eString

An arbitrary string.

@ eOutputFile

Name of file (must be writable)

@ eInteger

Convertible into an integer number (int or Int8)

EMethod

Compression/decompression methods.

@ eZstd

ZStandard (raw zstd data)

@ eGZipFile

.gz file (including concatenated files)

void SetDiagPostPrefix(const char *prefix)

Specify a string to prefix all subsequent error postings with.

EDiagSev SetDiagPostLevel(EDiagSev post_sev=eDiag_Error)

Set the threshold severity for posting the messages.

#define ERR_POST(message)

Error posting with file, line number information but without error codes.

@ eDiag_Warning

Warning message.

const string & Get(const string &name, bool *found=NULL) const

Get environment value by name.

void Error(CExceptionArgs_Base &args)

TErrCode GetErrCode(void) const

Get error code.

#define NCBI_THROW(exception_class, err_code, message)

Generic macro to throw an exception, given the exception class, error code and message string.

void Warning(CExceptionArgs_Base &args)

static char GetPathSeparator(void)

Get path separator symbol specific for the current platform.

#define MSerial_AsnBinary

#define MSerial_AsnText

I/O stream manipulators –.

CBeginInfo Begin(C &obj)

Get starting point of object hierarchy.

CSeq_entry_Handle AddTopLevelSeqEntry(CSeq_entry &top_entry, TPriority pri=kPriority_Default, EExist action=eExist_Default)

Add seq_entry, default priority is higher than for defaults or loaders Add object to the score with p...

bool Empty(void) const THROWS_NONE

Check if CConstRef is empty – not pointing to any object which means having a null value.

void Reset(void)

Reset reference object.

void Reset(void)

Reset reference object.

bool NotEmpty(void) const THROWS_NONE

Check if CRef is not empty – pointing to an object and has a non-null value.

bool Empty(void) const THROWS_NONE

Check if CRef is empty – not pointing to any object, which means having a null value.

int32_t Int4

4-byte (32-bit) signed integer

uint64_t Uint8

8-byte (64-bit) unsigned integer

virtual const string & Get(const string &section, const string &name, TFlags flags=0) const

Get the parameter value.

bool IsEnabled(void)

Indicates whether application usage statistics collection is enabled for a current reporter instance.

IO_PREFIX::ostream CNcbiOstream

Portable alias for ostream.

IO_PREFIX::istream CNcbiIstream

Portable alias for istream.

IO_PREFIX::ifstream CNcbiIfstream

Portable alias for ifstream.

static int CompareNocase(const CTempString s1, SIZE_TYPE pos, SIZE_TYPE n, const char *s2)

Case-insensitive compare of a substring with another string.

static list< string > & Split(const CTempString str, const CTempString delim, list< string > &arr, TSplitFlags flags=0, vector< SIZE_TYPE > *token_pos=NULL)

Split a string using specified delimiters.

static bool EndsWith(const CTempString str, const CTempString end, ECase use_case=eCase)

Check if a string ends with a specified suffix value.

static Uint8 StringToUInt8_DataSize(const CTempString str, TStringToNumFlags flags=0)

Convert string that can contain "software" qualifiers to Uint8.

static TNumeric StringToNumeric(const CTempString str, TStringToNumFlags flags=0, int base=10)

Convert string to a numeric value.

static string IntToString(int value, TNumToStringFlags flags=0, int base=10)

Convert int to string.

static enable_if< is_arithmetic< TNumeric >::value||is_convertible< TNumeric, Int8 >::value, string >::type NumericToString(TNumeric value, TNumToStringFlags flags=0, int base=10)

Convert numeric value to string.

static string TruncateSpaces(const string &str, ETrunc where=eTrunc_Both)

Truncate whitespace in a string.

static string UInt8ToString(Uint8 value, TNumToStringFlags flags=0, int base=10)

Convert UInt8 to string.

@ fConvErr_NoThrow

Do not throw an exception on error.

@ eNocase

Case insensitive compare.

double Elapsed(void) const

Return time elapsed since first Start() or last Restart() call (in seconds).

void Start(void)

Start the timer.

const TOrg & GetOrg(void) const

Get the Org member data.

list< CRef< CSeq_loc > > TMasks

const TMasks & GetMasks(void) const

Get the Masks member data.

EBlast_filter_program

This defines the possible sequence filtering algorithms to be used in a BLAST database.

const Tdata & Get(void) const

Get the member data.

list< CRef< CBlast_def_line > > Tdata

TMore GetMore(void) const

Get the More member data.

@ eBlast_filter_program_other

bool IsStr(void) const

Check if variant Str is selected.

bool IsSetType(void) const

type of object within class Check if a value has been assigned to Type data member.

const TStr & GetStr(void) const

Get the variant data.

const TType & GetType(void) const

Get the Type member data.

TSet & SetSet(void)

Select the variant.

@ eClass_nuc_prot

nuc acid and coded proteins

list< CRef< CSeqdesc > > Tdata

const TUser & GetUser(void) const

Get the variant data.

bool IsOrg(void) const

Check if variant Org is selected.

const TSource & GetSource(void) const

Get the variant data.

bool IsSource(void) const

Check if variant Source is selected.

const Tdata & Get(void) const

Get the member data.

bool CanGetDescr(void) const

Check if it is safe to call GetDescr method.

const TOrg & GetOrg(void) const

Get the variant data.

void SetDescr(TDescr &value)

Assign a value to Descr data member.

Tdata & Set(void)

Assign a value to data member.

bool IsTitle(void) const

Check if variant Title is selected.

const TDescr & GetDescr(void) const

Get the Descr member data.

bool IsUser(void) const

Check if variant User is selected.

char * dbname(DBPROCESS *dbproc)

Get name of current database.

unsigned int

A callback function used to compare two keys in a database.

static TTaxId s_GetTaxId(const CBioseq &bio)

static bool s_HasTitle(const CBioseq &bio)

static const string kOutput("out")

Command line flag to represent the output.

static string Uint8ToString_DataSize(Uint8 v, unsigned minprec=10)

Converts a Uint8 into a string which contains a data size (converse to NStr::StringToUInt8_DataSize)

void s_ReadObject(CNcbiIstream &file, CFormatGuess::EFormat fmt, CRef< TObj > &obj, const string &msg)

Reads an object defined in a NCBI ASN.1 spec from a stream in multiple formats: binary and text ASN....

int main(int argc, const char *argv[])

static const string kInputSeparators(" ")

Defines token separators when multiple inputs are present.

static const string kInput("in")

Command line flag to represent the input.

const string version

version string

#define ASSERT

macro for assert.

Defines the CNcbiApplication and CAppException classes for creating NCBI applications.

void copy(Njn::Matrix< S > *matrix_, const Njn::Matrix< T > &matrix0_)

Defines BLAST database access classes.

Defines exception class and several constants for SeqDB.

EBlastDbVersion

BLAST database version.

void SeqDB_GetMetadataFileExtension(bool db_is_protein, string &extn)

void SeqDB_SplitQuoted(const string &dbname, vector< CSeqDB_Substring > &dbs, bool keep_quote=false)

Combine and quote list of database names.

void SeqDB_CombineAndQuote(const vector< string > &dbs, string &dbname)

Combine and quote list of database names.

static SLJIT_INLINE sljit_ins msg(sljit_gpr r, sljit_s32 d, sljit_gpr x, sljit_gpr b)

C++ I/O stream wrappers to compress/decompress data on-the-fly.

List of sequence offset ranges.

Structure describing filtered regions created using a particular sequence filtering algorithm.

int algorithm_id

Identifies the algorithm used.

vector< pair< TSeqPos, TSeqPos > > offsets

Start and end offsets of the filtered area.

Defines BLAST database construction classes.

Defines exception class for WriteDB.

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