// BamAlignment.cpp (c) 2009 Derek Barnett
// Marth Lab, Department of Biology, Boston College
// ---------------------------------------------------------------------------
-// Last modified: 4 October 2011 (DB)
+// Last modified: 10 October 2011 (DB)
// ---------------------------------------------------------------------------
// Provides the BamAlignment data structure
// ***************************************************************************
-#include <api/BamAlignment.h>
-#include <api/BamConstants.h>
+#include "api/BamAlignment.h"
+#include "api/BamConstants.h"
using namespace BamTools;
-
-#include <cctype>
-#include <cstdio>
-#include <cstdlib>
-#include <cstring>
-#include <exception>
-#include <iostream>
-#include <map>
-#include <utility>
using namespace std;
/*! \class BamTools::BamAlignment
*/
BamAlignment::~BamAlignment(void) { }
-///*! \fn bool BamAlignment::AddTag(const std::string& tag, const std::string& type, const std::string& value)
-// \brief Adds a field with string data to the BAM tags.
-
-// Does NOT modify an existing tag - use \link BamAlignment::EditTag() \endlink instead.
-
-// \param[in] tag 2-character tag name
-// \param[in] type 1-character tag type (must be "Z" or "H")
-// \param[in] value string data to store
-// \return \c true if the \b new tag was added successfully
-// \sa \samSpecURL for more details on reserved tag names, supported tag types, etc.
-//*/
-
-
-///*! \fn bool AddTag(const std::string& tag, const std::vector<uint8_t>& values);
-// \brief Adds a numeric array field to the BAM tags.
-
-// Does NOT modify an existing tag - use \link BamAlignment::EditTag() \endlink instead.
-
-// \param tag 2-character tag name
-// \param values vector of uint8_t values to store
-
-// \return \c true if the \b new tag was added successfully
-// \sa \samSpecURL for more details on reserved tag names, supported tag types, etc.
-//*/
-
/*! \fn bool BamAlignment::BuildCharData(void)
\brief Populates alignment string fields (read name, bases, qualities, tag data).
// calculate character lengths/offsets
const unsigned int dataLength = SupportData.BlockLength - Constants::BAM_CORE_SIZE;
- const unsigned int seqDataOffset = SupportData.QueryNameLength + (SupportData.NumCigarOperations * 4);
+ const unsigned int seqDataOffset = SupportData.QueryNameLength + (SupportData.NumCigarOperations*4);
const unsigned int qualDataOffset = seqDataOffset + (SupportData.QuerySequenceLength+1)/2;
const unsigned int tagDataOffset = qualDataOffset + SupportData.QuerySequenceLength;
const unsigned int tagDataLength = dataLength - tagDataOffset;
QueryBases.clear();
if ( hasSeqData ) {
QueryBases.reserve(SupportData.QuerySequenceLength);
- for (unsigned int i = 0; i < SupportData.QuerySequenceLength; ++i) {
- char singleBase = Constants::BAM_DNA_LOOKUP[ ( (seqData[(i/2)] >> (4*(1-(i%2)))) & 0xf ) ];
+ for ( size_t i = 0; i < SupportData.QuerySequenceLength; ++i ) {
+ const char singleBase = Constants::BAM_DNA_LOOKUP[ ( (seqData[(i/2)] >> (4*(1-(i%2)))) & 0xf ) ];
QueryBases.append(1, singleBase);
}
}
Qualities.clear();
if ( hasQualData ) {
Qualities.reserve(SupportData.QuerySequenceLength);
- for (unsigned int i = 0; i < SupportData.QuerySequenceLength; ++i) {
- char singleQuality = (char)(qualData[i]+33);
+ for ( size_t i = 0; i < SupportData.QuerySequenceLength; ++i ) {
+ const char singleQuality = static_cast<const char>(qualData[i]+33);
Qualities.append(1, singleQuality);
}
}
for ( ; cigarIter != cigarEnd; ++cigarIter ) {
const CigarOp& op = (*cigarIter);
- switch (op.Type) {
+ switch ( op.Type ) {
// for 'M', 'I', '=', 'X' - write bases
case (Constants::BAM_CIGAR_MATCH_CHAR) :
case (Constants::BAM_CIGAR_HARDCLIP_CHAR) :
break;
- // shouldn't get here
+ // invalid CIGAR op-code
default:
- cerr << "BamAlignment ERROR: invalid CIGAR operation type: "
- << op.Type << endl;
- exit(1);
+ const string message = string("invalid CIGAR operation type: ") + op.Type;
+ SetErrorString("BamAlignment::BuildCharData", message);
+ return false;
}
}
}
TagData.clear();
if ( hasTagData ) {
if ( IsBigEndian ) {
- int i = 0;
- while ( (unsigned int)i < tagDataLength ) {
+ size_t i = 0;
+ while ( i < tagDataLength ) {
i += Constants::BAM_TAG_TAGSIZE; // skip tag chars (e.g. "RG", "NM", etc.)
const char type = tagData[i]; // get tag type at position i
// swap endian-ness of number of elements in place, then retrieve for loop
BamTools::SwapEndian_32p(&tagData[i]);
- int32_t numElements;
+ uint32_t numElements;
memcpy(&numElements, &tagData[i], sizeof(uint32_t));
i += sizeof(uint32_t);
// swap endian-ness of array elements
- for ( int j = 0; j < numElements; ++j ) {
+ for ( size_t j = 0; j < numElements; ++j ) {
switch (arrayType) {
case (Constants::BAM_TAG_TYPE_INT8) :
case (Constants::BAM_TAG_TYPE_UINT8) :
i += sizeof(uint32_t);
break;
default:
- // error case
- cerr << "BamAlignment ERROR: unknown binary array type encountered: "
- << arrayType << endl;
+ const string message = string("invalid binary array type: ") + arrayType;
+ SetErrorString("BamAlignment::BuildCharData", message);
return false;
}
}
break;
}
- // shouldn't get here
+ // invalid tag type-code
default :
- cerr << "BamAlignment ERROR: invalid tag value type: "
- << type << endl;
- exit(1);
+ const string message = string("invalid tag type: ") + type;
+ SetErrorString("BamAlignment::BuildCharData", message);
+ return false;
}
}
}
// store tagData in alignment
TagData.resize(tagDataLength);
- memcpy((char*)TagData.data(), tagData, tagDataLength);
+ memcpy((char*)(TagData.data()), tagData, tagDataLength);
}
- // clear the core-only flag
+ // clear core-only flag & return success
SupportData.HasCoreOnly = false;
-
- // return success
return true;
}
-///*! \fn bool BamAlignment::EditTag(const std::string& tag, const std::string& type, const std::string& value)
-// \brief Edits a BAM tag field containing string data.
-
-// If \a tag does not exist, a new entry is created.
-
-// \param tag 2-character tag name
-// \param type 1-character tag type (must be "Z" or "H")
-// \param value string data to store
-
-// \return \c true if the tag was modified/created successfully
-
-// \sa BamAlignment::RemoveTag()
-// \sa \samSpecURL for more details on reserved tag names, supported tag types, etc.
-//*/
-
-///*! \fn bool EditTag(const std::string& tag, const std::vector<uint8_t>& values);
-// \brief Edits a BAM tag field containing a numeric array.
-
-// If \a tag does not exist, a new entry is created.
-
-// \param tag 2-character tag name
-// \param value vector of uint8_t values to store
-
-// \return \c true if the tag was modified/created successfully
-// \sa \samSpecURL for more details on reserved tag names, supported tag types, etc.
-//*/
-
-/*! \fn bool BamAlignment::FindTag(const std::string& tag, char*& pTagData, const unsigned int& tagDataLength, unsigned int& numBytesParsed)
+/*! \fn bool BamAlignment::FindTag(const std::string& tag, char*& pTagData, const unsigned int& tagDataLength, unsigned int& numBytesParsed) const
\internal
Searches for requested tag in BAM tag data.
- \param tag requested 2-character tag name
- \param pTagData pointer to current position in BamAlignment::TagData
- \param tagDataLength length of BamAlignment::TagData
- \param numBytesParsed number of bytes parsed so far
+ \param[in] tag requested 2-character tag name
+ \param[in,out] pTagData pointer to current position in BamAlignment::TagData
+ \param[in] tagDataLength length of BamAlignment::TagData
+ \param[in,out] numBytesParsed number of bytes parsed so far
\return \c true if found
bool BamAlignment::FindTag(const std::string& tag,
char*& pTagData,
const unsigned int& tagDataLength,
- unsigned int& numBytesParsed)
+ unsigned int& numBytesParsed) const
{
while ( numBytesParsed < tagDataLength ) {
return false;
}
-/*! \fn bool BamAlignment::GetEditDistance(uint32_t& editDistance) const
- \brief Retrieves value of edit distance tag ("NM").
+/*! \fn int BamAlignment::GetEndPosition(bool usePadded = false, bool closedInterval = false) const
+ \brief Calculates alignment end position, based on its starting position and CIGAR data.
- \deprecated Instead use BamAlignment::GetTag()
- \code
- BamAlignment::GetTag("NM", editDistance);
- \endcode
+ \warning The position returned now represents a zero-based, HALF-OPEN interval.
+ In previous versions of BamTools (0.x & 1.x) all intervals were treated
+ as zero-based, CLOSED.
- \param editDistance destination for retrieved value
-
- \return \c true if found
-*/
-
-// TODO : REMOVE THIS METHOD
-bool BamAlignment::GetEditDistance(uint32_t& editDistance) const {
- return GetTag("NM", (uint32_t&)editDistance);
-}
-
-/*! \fn int BamAlignment::GetEndPosition(bool usePadded = false, bool zeroBased = true) const
- \brief Calculates alignment end position, based on starting position and CIGAR data.
-
- \param usePadded Inserted bases affect reported position. Default is false, so that reported
- position stays 'sync-ed' with reference coordinates.
- \param zeroBased Return (BAM standard) 0-based coordinate. Setting this to false can be useful
- when using BAM data with half-open formats (e.g. BED).
+ \param[in] usePadded Allow inserted bases to affect the reported position. Default is
+ false, so that reported position stays synced with reference
+ coordinates.
+ \param[in] closedInterval Setting this to true will return a 0-based end coordinate. Default is
+ false, so that his value represents a standard, half-open interval.
\return alignment end position
*/
-int BamAlignment::GetEndPosition(bool usePadded, bool zeroBased) const {
+int BamAlignment::GetEndPosition(bool usePadded, bool closedInterval) const {
// initialize alignment end to starting position
int alignEnd = Position;
vector<CigarOp>::const_iterator cigarIter = CigarData.begin();
vector<CigarOp>::const_iterator cigarEnd = CigarData.end();
for ( ; cigarIter != cigarEnd; ++cigarIter) {
- const char cigarType = (*cigarIter).Type;
- const uint32_t& cigarLength = (*cigarIter).Length;
-
- if ( cigarType == Constants::BAM_CIGAR_MATCH_CHAR ||
- cigarType == Constants::BAM_CIGAR_DEL_CHAR ||
- cigarType == Constants::BAM_CIGAR_REFSKIP_CHAR )
- alignEnd += cigarLength;
- else if ( usePadded && cigarType == Constants::BAM_CIGAR_INS_CHAR )
- alignEnd += cigarLength;
+ const CigarOp& op = (*cigarIter);
+
+ switch ( op.Type ) {
+
+ // increase end position on CIGAR chars [DMXN=]
+ case Constants::BAM_CIGAR_DEL_CHAR :
+ case Constants::BAM_CIGAR_MATCH_CHAR :
+ case Constants::BAM_CIGAR_MISMATCH_CHAR :
+ case Constants::BAM_CIGAR_REFSKIP_CHAR :
+ case Constants::BAM_CIGAR_SEQMATCH_CHAR :
+ alignEnd += op.Length;
+ break;
+
+ // increase end position on insertion, only if @usePadded is true
+ case Constants::BAM_CIGAR_INS_CHAR :
+ if ( usePadded )
+ alignEnd += op.Length;
+ break;
+
+ // all other CIGAR chars do not affect end position
+ default :
+ break;
+ }
}
- // adjust for zero-based coordinates, if requested
- if ( zeroBased ) alignEnd -= 1;
+ // adjust for closedInterval, if requested
+ if ( closedInterval )
+ alignEnd -= 1;
// return result
return alignEnd;
}
-/*! \fn bool BamAlignment::GetReadGroup(std::string& readGroup) const
- \brief Retrieves value of read group tag ("RG").
+/*! \fn std::string BamAlignment::GetErrorString(void) const
+ \brief Returns a human-readable description of the last error that occurred
- \deprecated Instead use BamAlignment::GetTag()
- \code
- BamAlignment::GetTag("RG", readGroup);
- \endcode
+ This method allows elimination of STDERR pollution. Developers of client code
+ may choose how the messages are displayed to the user, if at all.
- \param readGroup destination for retrieved value
-
- \return \c true if found
+ \return error description
*/
-
-// TODO : REMOVE THIS METHOD
-bool BamAlignment::GetReadGroup(std::string& readGroup) const {
- return GetTag("RG", readGroup);
+std::string BamAlignment::GetErrorString(void) const {
+ return ErrorString;
}
-///*! \fn bool BamAlignment::GetTag(const std::string& tag, std::string& destination) const
-// \brief Retrieves the string value associated with a BAM tag.
-
-// \param tag 2-character tag name
-// \param destination destination for retrieved value
-
-// \return \c true if found
-//*/
-
-///*! \fn bool BamAlignment::GetTag(const std::string& tag, std::vector<uint32_t>& destination) const
-// \brief Retrieves the numeric array data associated with a BAM tag
-
-// \param tag 2-character tag name
-// \param destination destination for retrieved data
-
-// \return \c true if found
-//*/
-
/*! \fn bool BamAlignment::GetTagType(const std::string& tag, char& type) const
\brief Retrieves the BAM tag type-code associated with requested tag name.
- \param tag 2-character tag name
- \param type destination for the retrieved (1-character) tag type
+ \param[in] tag 2-character tag name
+ \param[out] type retrieved (1-character) type-code
\return \c true if found
\sa \samSpecURL for more details on reserved tag names, supported tag types, etc.
*/
bool BamAlignment::GetTagType(const std::string& tag, char& type) const {
- // make sure tag data exists
- if ( SupportData.HasCoreOnly || TagData.empty() )
+ // skip if alignment is core-only
+ if ( SupportData.HasCoreOnly ) {
+ // TODO: set error string?
+ return false;
+ }
+
+ // skip if no tags present
+ if ( TagData.empty() ) {
+ // TODO: set error string?
return false;
+ }
// localize the tag data
char* pTagData = (char*)TagData.data();
unsigned int numBytesParsed = 0;
// if tag not found, return failure
- if ( !FindTag(tag, pTagData, tagDataLength, numBytesParsed) )
+ if ( !FindTag(tag, pTagData, tagDataLength, numBytesParsed) ){
+ // TODO: set error string?
return false;
+ }
// otherwise, retrieve & validate tag type code
type = *(pTagData - 1);
// unknown tag type
default:
- cerr << "BamAlignment ERROR: unknown tag type encountered: "
- << type << endl;
+ const string message = string("invalid tag type: ") + type;
+ SetErrorString("BamAlignment::GetTagType", message);
return false;
}
}
/*! \fn bool BamAlignment::HasTag(const std::string& tag) const
\brief Returns true if alignment has a record for requested tag.
- \param tag 2-character tag name
+
+ \param[in] tag 2-character tag name
\return \c true if alignment has a record for tag
*/
bool BamAlignment::HasTag(const std::string& tag) const {
return ( (AlignmentFlag & Constants::BAM_ALIGNMENT_READ_2) != 0 );
}
-/*! \fn bool BamAlignment::IsValidSize(const string& tag, const string& type) const
+/*! \fn bool BamAlignment::IsValidSize(const std::string& tag, const std::string& type) const
\internal
Checks that tag name & type strings are expected sizes.
- \a tag should have length
- \a type should have length 1
- \param tag BAM tag name
- \param type BAM tag type-code
-
- \return \c true if both \a tag and \a type are correct sizes
+ \param tag[in] BAM tag name
+ \param type[in] BAM tag type-code
+ \return \c true if both input strings are valid sizes
*/
-bool BamAlignment::IsValidSize(const string& tag, const string& type) {
+bool BamAlignment::IsValidSize(const std::string& tag, const std::string& type) const {
return (tag.size() == Constants::BAM_TAG_TAGSIZE) &&
(type.size() == Constants::BAM_TAG_TYPESIZE);
}
-/*! \fn bool BamAlignment::RemoveTag(const std::string& tag)
+/*! \fn void BamAlignment::RemoveTag(const std::string& tag)
\brief Removes field from BAM tags.
- \return \c true if tag was removed successfully (or didn't exist before)
+ \param[in] tag 2-character name of field to remove
*/
-bool BamAlignment::RemoveTag(const std::string& tag) {
+void BamAlignment::RemoveTag(const std::string& tag) {
// if char data not populated, do that first
if ( SupportData.HasCoreOnly )
// skip if no tags available
if ( TagData.empty() )
- return false;
+ return;
// localize the tag data
char* pOriginalTagData = (char*)TagData.data();
unsigned int newTagDataLength = 0;
unsigned int numBytesParsed = 0;
- // if tag not found, simply return true
+ // skip if tag not found
if ( !FindTag(tag, pTagData, originalTagDataLength, numBytesParsed) )
- return true;
+ return;
// otherwise, remove it
- char* newTagData = new char[originalTagDataLength];
+ RaiiBuffer newTagData(originalTagDataLength);
// copy original tag data up til desired tag
pTagData -= 3;
numBytesParsed -= 3;
const unsigned int beginningTagDataLength = numBytesParsed;
newTagDataLength += beginningTagDataLength;
- memcpy(newTagData, pOriginalTagData, numBytesParsed);
+ memcpy(newTagData.Buffer, pOriginalTagData, numBytesParsed);
// attemp to skip to next tag
const char* pTagStorageType = pTagData + 2;
// squeeze remaining tag data
const unsigned int skippedDataLength = (numBytesParsed - beginningTagDataLength);
const unsigned int endTagDataLength = originalTagDataLength - beginningTagDataLength - skippedDataLength;
- memcpy(newTagData + beginningTagDataLength, pTagData, endTagDataLength );
+ memcpy(newTagData.Buffer + beginningTagDataLength, pTagData, endTagDataLength );
// save modified tag data in alignment
- TagData.assign(newTagData, beginningTagDataLength + endTagDataLength);
+ TagData.assign(newTagData.Buffer, beginningTagDataLength + endTagDataLength);
}
+}
- // clean up & return success
- delete[] newTagData;
- return true;
+/*! \fn void BamAlignment::SetErrorString(const std::string& where, const std::string& what) const
+ \internal
+
+ Sets a formatted error string for this alignment.
+
+ \param[in] where class/method where error occurred
+ \param[in] what description of error
+*/
+void BamAlignment::SetErrorString(const std::string& where, const std::string& what) const {
+ static const string SEPARATOR = ": ";
+ ErrorString = where + SEPARATOR + what;
}
/*! \fn void BamAlignment::SetIsDuplicate(bool ok)
else AlignmentFlag |= Constants::BAM_ALIGNMENT_MATE_UNMAPPED;
}
-/*! \fn void BamAlignment::SetIsMateUnmapped(bool ok)
- \brief Complement of using SetIsMateMapped().
- \deprecated For sake of symmetry with the query methods
- \sa IsMateMapped(), SetIsMateMapped()
-*/
-void BamAlignment::SetIsMateUnmapped(bool ok) {
- SetIsMateMapped(!ok);
-}
-
/*! \fn void BamAlignment::SetIsMateReverseStrand(bool ok)
\brief Sets "alignment's mate mapped to reverse strand" flag to \a ok.
*/
else AlignmentFlag &= ~Constants::BAM_ALIGNMENT_REVERSE_STRAND;
}
-/*! \fn void BamAlignment::SetIsSecondaryAlignment(bool ok)
- \brief Complement of using SetIsPrimaryAlignment().
- \deprecated For sake of symmetry with the query methods
- \sa IsPrimaryAlignment(), SetIsPrimaryAlignment()
-*/
-void BamAlignment::SetIsSecondaryAlignment(bool ok) {
- SetIsPrimaryAlignment(!ok);
-}
-
/*! \fn void BamAlignment::SetIsSecondMate(bool ok)
\brief Sets "alignment is second mate on read" flag to \a ok.
*/
else AlignmentFlag &= ~Constants::BAM_ALIGNMENT_READ_2;
}
-/*! \fn void BamAlignment::SetIsUnmapped(bool ok)
- \brief Complement of using SetIsMapped().
- \deprecated For sake of symmetry with the query methods
- \sa IsMapped(), SetIsMapped()
-*/
-void BamAlignment::SetIsUnmapped(bool ok) {
- SetIsMapped(!ok);
-}
-
-/*! \fn bool BamAlignment::SkipToNextTag(const char storageType, char*& pTagData, unsigned int& numBytesParsed)
+/*! \fn bool BamAlignment::SkipToNextTag(const char storageType, char*& pTagData, unsigned int& numBytesParsed) const
\internal
Moves to next available tag in tag data string
- \param storageType BAM tag type-code that determines how far to move cursor
- \param pTagData pointer to current position (cursor) in tag string
- \param numBytesParsed report of how many bytes were parsed (cumulatively)
+ \param[in] storageType BAM tag type-code that determines how far to move cursor
+ \param[in,out] pTagData pointer to current position (cursor) in tag string
+ \param[in,out] numBytesParsed report of how many bytes were parsed (cumulatively)
\return \c if storageType was a recognized BAM tag type
- \post \a pTagData will point to the byte where the next tag data begins.
+
+ \post \a pTagData will point to the byte where the next tag data begins.
\a numBytesParsed will correspond to the cursor's position in the full TagData string.
*/
bool BamAlignment::SkipToNextTag(const char storageType,
char*& pTagData,
- unsigned int& numBytesParsed)
+ unsigned int& numBytesParsed) const
{
switch (storageType) {
// read number of elements
int32_t numElements;
- memcpy(&numElements, pTagData, sizeof(uint32_t)); // already endian-swapped if necessary
+ memcpy(&numElements, pTagData, sizeof(uint32_t)); // already endian-swapped, if needed
numBytesParsed += sizeof(uint32_t);
pTagData += sizeof(uint32_t);
bytesToSkip = numElements*sizeof(uint32_t);
break;
default:
- cerr << "BamAlignment ERROR: unknown binary array type encountered: "
- << arrayType << endl;
+ const string message = string("invalid binary array type: ") + arrayType;
+ SetErrorString("BamAlignment::SkipToNextTag", message);
return false;
}
}
default:
- cerr << "BamAlignment ERROR: unknown tag type encountered"
- << storageType << endl;
+ const string message = string("invalid tag type: ") + storageType;
+ SetErrorString("BamAlignment::SkipToNextTag", message);
return false;
}
- // return success
+ // if we get here, tag skipped OK - return success
return true;
}
projects / bamtools.git / blobdiff