GtUchar gt_Bwtseqcontextiterator_next(GtUword *bound,
Bwtseqcontextiterator *bsci)
{
GtUchar cc;
if (bsci->bound != BWTSeqTerminatorPos(bsci->bwtseq))
{
cc = BWTSeqGetSym(bsci->bwtseq, bsci->bound);
} else
{
cc = SEPARATOR;
}
*bound = bsci->bound = BWTSeqLFMap(bsci->bwtseq, bsci->bound, &bsci->extBits);
return cc;
}
extern Seqpos
BWTSeqLocateMatch(const BWTSeq *bwtSeq, Seqpos pos,
struct extBitsRetrieval *extBits)
{
if (bwtSeq->featureToggles & BWTLocateBitmap)
{
Seqpos nextLocate = pos;
unsigned locateOffset = 0;
while (!BWTSeqPosHasLocateInfo(bwtSeq, nextLocate, extBits))
nextLocate = BWTSeqLFMap(bwtSeq, nextLocate, extBits), ++locateOffset;
EISRetrieveExtraBits(bwtSeq->seqIdx, nextLocate,
EBRF_RETRIEVE_CWBITS | EBRF_RETRIEVE_VARBITS,
extBits, bwtSeq->hint);
{
Seqpos maxPosVal = ((bwtSeq->featureToggles & BWTReversiblySorted)?
(BWTSeqLength(bwtSeq) - 1)
/bwtSeq->locateSampleInterval:
BWTSeqLength(bwtSeq) - 1);
unsigned bitsPerCount = requiredSeqposBits(extBits->len),
bitsPerBWTPos = requiredSeqposBits(extBits->len - 1),
bitsPerOrigPos = requiredSeqposBits(maxPosVal);
BitOffset locateRecordIndex =
gt_bs1BitsCount(extBits->cwPart, extBits->cwOffset,
nextLocate - extBits->start),
locateRecordOffset = ((bwtSeq->featureToggles & BWTLocateCount?
bitsPerBWTPos:0) + bitsPerOrigPos)
* locateRecordIndex
+ ((bwtSeq->featureToggles & BWTLocateCount)?bitsPerCount:0);
Seqpos matchPos =
gt_bsGetSeqpos(
extBits->varPart, extBits->varOffset + locateRecordOffset
+ ((bwtSeq->featureToggles & BWTLocateCount)?bitsPerBWTPos:0),
bitsPerOrigPos);
if (bwtSeq->featureToggles & BWTReversiblySorted)
matchPos = matchPos * bwtSeq->locateSampleInterval;
matchPos += locateOffset;
gt_assert(!(bwtSeq->featureToggles & BWTLocateCount)
|| gt_bsGetSeqpos(extBits->varPart,
extBits->varOffset + locateRecordOffset,
bitsPerBWTPos)
== nextLocate - extBits->start);
return matchPos;
}
}
else if (bwtSeq->featureToggles & BWTLocateCount)
{
BitOffset markOffset;
Seqpos nextLocate = pos, matchPos;
unsigned locateOffset = 0; /* mark is at most locateInterval
* positions away */
unsigned bitsPerOrigPos
= requiredSeqposBits(((bwtSeq->featureToggles & BWTReversiblySorted)?
(BWTSeqLength(bwtSeq) - 1)
/bwtSeq->locateSampleInterval:
BWTSeqLength(bwtSeq) - 1));
while ((markOffset = searchLocateCountMark(bwtSeq, nextLocate,
extBits)) == 0)
{
nextLocate = BWTSeqLFMap(bwtSeq, nextLocate, extBits);
++locateOffset;
}
matchPos = gt_bsGetSeqpos(extBits->varPart, markOffset, bitsPerOrigPos);
if (bwtSeq->featureToggles & BWTReversiblySorted)
matchPos = matchPos * bwtSeq->locateSampleInterval;
matchPos += locateOffset;
return matchPos;
}
/* Internal error: Trying to locate in BWT sequence index without locate
information. */
abort();
return 0; /* shut up compiler */
}
enum verifyBWTSeqErrCode
gt_BWTSeqVerifyIntegrity(BWTSeq *bwtSeq, const char *projectName,
int checkFlags,
GtUword tickPrint, FILE *fp,
GtLogger *verbosity, GtError *err)
{
Suffixarray suffixArray;
struct extBitsRetrieval extBits;
bool suffixArrayIsInitialized = false, extBitsAreInitialized = false;
enum verifyBWTSeqErrCode retval = VERIFY_BWTSEQ_NO_ERROR;
do
{
GtUword seqLen;
gt_assert(bwtSeq && projectName && err);
gt_error_check(err);
initExtBitsRetrieval(&extBits);
extBitsAreInitialized = true;
if (gt_mapsuffixarray(&suffixArray,
SARR_SUFTAB | SARR_ESQTAB, projectName, verbosity, err))
{
gt_error_set(err, "Cannot load reference suffix array project with"
" demand for suffix table file and encoded sequence"
" for project: %s", projectName);
retval = VERIFY_BWTSEQ_REFLOAD_ERROR;
break;
}
suffixArrayIsInitialized = true;
seqLen = gt_encseq_total_length(suffixArray.encseq) + 1;
if (BWTSeqLength(bwtSeq) != seqLen)
{
gt_error_set(err, "length mismatch for suffix array project %s and "
"bwt sequence index", projectName);
retval = VERIFY_BWTSEQ_LENCOMPARE_ERROR;
break;
}
if (checkFlags & VERIFY_BWTSEQ_SUFVAL
&& BWTSeqHasLocateInformation(bwtSeq))
{
GtUword i;
for (i = 0; i < seqLen && retval == VERIFY_BWTSEQ_NO_ERROR; ++i)
{
if (gt_BWTSeqPosHasLocateInfo(bwtSeq, i, &extBits))
{
GtUword sfxArrayValue = gt_BWTSeqLocateMatch(bwtSeq, i,
&extBits);
if (sfxArrayValue != ESASUFFIXPTRGET(suffixArray.suftab,i))
{
gt_error_set(err, "Failed suffix array value comparison"
" at position "GT_WU": "GT_WU" != "GT_WU"",
i, sfxArrayValue,
ESASUFFIXPTRGET(suffixArray.suftab,i));
retval = VERIFY_BWTSEQ_SUFVAL_ERROR;
break;
}
}
if (tickPrint && !((i + 1) % tickPrint))
putc('.', fp);
}
if (tickPrint)
putc('\n', fp);
if (retval != VERIFY_BWTSEQ_NO_ERROR)
break;
}
else if (checkFlags & VERIFY_BWTSEQ_SUFVAL)
{
gt_error_set(err, "check of suffix array values was requested,"
" but index contains no locate information!");
retval = VERIFY_BWTSEQ_SUFVAL_ERROR;
break;
}
else if (!(checkFlags & VERIFY_BWTSEQ_SUFVAL)
&& BWTSeqHasLocateInformation(bwtSeq))
{
fputs("Not checking suftab values.\n", stderr);
}
if (BWTSeqHasLocateInformation(bwtSeq))
{
GtUword nextLocate = BWTSeqTerminatorPos(bwtSeq);
if (suffixArray.longest.defined &&
suffixArray.longest.valueunsignedlong != nextLocate)
{
gt_error_set(err, "terminator/0-rotation position mismatch "GT_WU""
" vs. "GT_WU"", suffixArray.longest.valueunsignedlong,
nextLocate);
retval = VERIFY_BWTSEQ_TERMPOS_ERROR;
break;
}
if ((checkFlags & VERIFY_BWTSEQ_LFMAPWALK)
&& (bwtSeq->featureToggles & BWTReversiblySorted))
{
GtUword i = seqLen;
/* handle first symbol specially because the encseq
* will not return the terminator symbol */
{
Symbol sym = BWTSeqGetSym(bwtSeq, nextLocate);
if (sym != UNDEFBWTCHAR)
{
gt_error_set(err, "symbol mismatch at position "GT_WU": "
"%d vs. reference symbol %d", i - 1, (int)sym,
(int)UNDEFBWTCHAR);
retval = VERIFY_BWTSEQ_LFMAPWALK_ERROR;
break;
}
--i;
nextLocate = BWTSeqLFMap(bwtSeq, nextLocate, &extBits);
}
while (i > 0)
{
Symbol symRef =
gt_encseq_get_encoded_char(suffixArray.encseq,
--i,
suffixArray.readmode);
Symbol symCmp = BWTSeqGetSym(bwtSeq, nextLocate);
if (symCmp != symRef)
{
gt_error_set(err, "symbol mismatch at position "GT_WU": "
"%d vs. reference symbol %d", i, symCmp, symRef);
retval = VERIFY_BWTSEQ_LFMAPWALK_ERROR;
break;
}
nextLocate = BWTSeqLFMap(bwtSeq, nextLocate, &extBits);
}
if (retval != VERIFY_BWTSEQ_NO_ERROR)
break;
}
else if ((checkFlags & VERIFY_BWTSEQ_LFMAPWALK)
&& !(bwtSeq->featureToggles & BWTReversiblySorted))
{
gt_error_set(err, "requested complete backwards regeneration in index"
" without regeneration capability");
retval = VERIFY_BWTSEQ_LFMAPWALK_IMP_ERROR;
break;
}
}
if (checkFlags & VERIFY_BWTSEQ_CONTEXT)
{
BWTSeqContextRetriever *bwtSeqCR =
gt_BWTSeqCRLoad(bwtSeq, projectName, CTX_MAP_ILOG_AUTOSIZE);
if (!bwtSeqCR)
{
gt_error_set(err, "cannot load BWT sequence context access table"
" for project %s", projectName);
retval = VERIFY_BWTSEQ_CONTEXT_LOADFAIL;
break;
}
fputs("Checking context regeneration.\n", stderr);
{
GtUword i, start, subSeqLen,
maxSubSeqLen = MIN(MAX(MIN_CONTEXT_LEN, seqLen/CONTEXT_FRACTION),
MAX_CONTEXT_LEN),
numTries = MIN(MAX_NUM_CONTEXT_CHECKS,
MAX(2, seqLen/CONTEXT_INTERVAL));
Symbol *contextBuf = gt_malloc(sizeof (Symbol) * MAX_CONTEXT_LEN);
GtEncseqReader *esr =
gt_encseq_create_reader_with_readmode(suffixArray.encseq,
suffixArray.readmode,
0);
for (i = 0; i < numTries && retval == VERIFY_BWTSEQ_NO_ERROR; ++i)
{
GtUword j, end, inSubSeqLen;
subSeqLen = random()%maxSubSeqLen + 1;
start = random()%(seqLen - subSeqLen + 1);
end = start + subSeqLen;
inSubSeqLen = subSeqLen - ((end==seqLen)?1:0);
gt_BWTSeqCRAccessSubseq(bwtSeqCR, start, subSeqLen, contextBuf);
gt_encseq_reader_reinit_with_readmode(esr, suffixArray.encseq,
suffixArray.readmode, start);
for (j = 0; j < inSubSeqLen; ++j)
{
Symbol symRef = gt_encseq_reader_next_encoded_char(esr);
Symbol symCmp = contextBuf[j];
if (symCmp != symRef)
{
gt_error_set(err, "symbol mismatch at position "GT_WU": "
"%d vs. reference symbol %d", start + j, (int)symCmp,
(int)symRef);
retval = VERIFY_BWTSEQ_CONTEXT_SYMFAIL;
break;
}
}
while (j < subSeqLen)
{
Symbol symRef = UNDEFBWTCHAR;
Symbol symCmp = contextBuf[j];
if (symCmp != symRef)
{
gt_error_set(err, "symbol mismatch at position "GT_WU": "
"%d vs. reference symbol %d", start + j, (int)symCmp,
(int)symRef);
retval = VERIFY_BWTSEQ_CONTEXT_SYMFAIL;
break;
}
++j;
}
}
if (retval == VERIFY_BWTSEQ_NO_ERROR)
fputs("Context regeneration completed successfully.\n", stderr);
gt_encseq_reader_delete(esr);
gt_free(contextBuf);
}
gt_deleteBWTSeqCR(bwtSeqCR);
}
} while (0);
if (suffixArrayIsInitialized) gt_freesuffixarray(&suffixArray);
if (extBitsAreInitialized) destructExtBitsRetrieval(&extBits);
return retval;
}