// Add() is typically used when creating a GLIX from GFF records.
// It can be called multiple times with the same label, the
// maximum position determines the sequence length.
void GLIX::Add(const char *Label, int Pos)
{
SEQDATA *IE = SeqIndexLookup(Label);
if (0 == IE)
AddToIndex(Label, Pos + 1);
else
{
if (Pos + 1 > IE->Length)
IE->Length = Pos + 1;
}
}
const TFile& TFileCollection::Merge(const TFile& file) {
// if (!file.FindInGeneratedPool())
// return;
TRef myfile = FindByAlias(file);
if (!myfile) {
myfile = new TFile(file);
Files.push_back(myfile);
} else
myfile->MergeAliases(file);
AddToIndex(*myfile);
return *myfile;
}
OGRErr OGRMILayerAttrIndex::IndexAllFeatures( int iField )
{
poLayer->ResetReading();
OGRFeature *poFeature = nullptr;
while( (poFeature = poLayer->GetNextFeature()) != nullptr )
{
const OGRErr eErr = AddToIndex( poFeature, iField );
delete poFeature;
if( eErr != OGRERR_NONE )
return eErr;
}
poLayer->ResetReading();
return OGRERR_NONE;
}
OGRErr OGRMILayerAttrIndex::IndexAllFeatures( int /* iField */ )
{
OGRFeature *poFeature;
poLayer->ResetReading();
while( (poFeature = poLayer->GetNextFeature()) != NULL )
{
OGRErr eErr = AddToIndex( poFeature );
delete poFeature;
if( eErr != CE_None )
return eErr;
}
poLayer->ResetReading();
return OGRERR_NONE;
}
void FilterB(int Tlen_, char *B_, int Qlen_, const FilterParams &FP, int Diameter,
bool Comp)
{
if (Comp)
Quit("-diameter requires -fwdonly (this needs to be fixed!)");
SeqQ = B_;
Tlen = Tlen_;
Qlen = Qlen_;
MinMatch = FP.SeedLength;
MaxError = FP.SeedDiffs;
TubeOffset = FP.TubeOffset;
const int Kmask = pow4(k) - 1;
// Ukonnen's Lemma
MinKmersPerHit = MinMatch + 1 - k*(MaxError + 1);
// Maximum distance between SeqQ positions of two k-mers in a match
// (More stringent bounds may be possible, but not a big problem
// if two adjacent matches get merged).
MaxKmerDist = MinMatch - k;
TubeWidth = TubeOffset + MaxError;
if (TubeOffset < MaxError)
{
Log("TubeOffset < MaxError\n");
exit(1);
}
if (MinKmersPerHit <= 0)
{
Log("MinKmersPerHit <= 0\n");
exit(1);
}
MaxActiveTubes = (Tlen + TubeWidth - 1)/TubeOffset + 1;
Tubes = all(TubeState, MaxActiveTubes);
zero(Tubes, TubeState, MaxActiveTubes);
// Ticker tracks cycling of circular list of active tubes.
int Ticker = TubeWidth;
// Initialize index to cover first window
int StartKmerValidPos = 0;
int EndKmerValidPos = 0;
AllocateIndex(Diameter, k);
const int KmerCount = Diameter - k + 1;
int Start;
int End;
for (Start = 0; Start < KmerCount; ++Start)
{
int StartKmer = GetKmer(SeqQ, Start);
AddToIndex(StartKmer, Start);
}
#if DEBUG
ValidateIndex(SeqQ, 0, Diameter - k);
#endif
// Scan entire sequence.
// Start is coordinate of first base in first k-mer in sliding window
// End is coordinate of first base in last k-mer in sliding window
Start = 0;
End = Start + Diameter - k;
int StartKmer = GetKmer(SeqQ, Start);
if (StartKmer == -1)
{
StartKmer = 0;
StartKmerValidPos = Start + k + 1;
}
int EndKmer = GetKmer(SeqQ, End);
if (EndKmer == -1)
{
EndKmer = 0;
EndKmerValidPos = End + k + 1;
}
for (; End < Qlen - k; ++Start, ++End)
{
#if DEBUG
if (Start%10000 == 0)
fprintf(stderr, "%d\n", Start);
//if (Start%1000 == 0)
// ValidateIndex(SeqQ, Start, End);
#endif
if (Start >= StartKmerValidPos)
{
assert(StartKmer == GetKmer(SeqQ, Start));
for (DeclareListPtr(p) = GetListPtr(StartKmer); NotEndOfList(p); p = GetListNext(p))
{
int HitPos = GetListPos(p);
CommonKmer(Start, HitPos);
}
DeleteFirstInstanceFromIndex(StartKmer, Start);
}
if (0 == --Ticker)
{
TubeEnd(Start);
Ticker = TubeOffset;
}
{
char c = SeqQ[Start + k];
int x = CharToLetter[c];
if (x < 0)
{
StartKmer = 0;
StartKmerValidPos = Start + k + 1;
}
else
StartKmer = ((StartKmer << 2) | x) & Kmask;
}
{
char c = SeqQ[End + k];
int x = CharToLetter[c];
if (x < 0)
{
EndKmer = 0;
EndKmerValidPos = End + k + 1;
}
else
EndKmer = ((EndKmer << 2) | x) & Kmask;
if (End+1 >= EndKmerValidPos)
{
assert(EndKmer == GetKmer(SeqQ, End+1));
AddToIndex(EndKmer, End+1);
}
}
}
#if DEBUG
ValidateIndex(SeqQ, Start, Qlen - k);
#endif
// Special case for end of sequence, don't slide the index
// for the last Diameter bases.
for (; Start < Qlen - k; ++Start)
{
{
char c = SeqQ[Start + k];
int x = CharToLetter[c];
if (x < 0)
{
StartKmer = 0;
StartKmerValidPos = Start + k + 1;
}
else
StartKmer = ((StartKmer << 2) | x) & Kmask;
}
if (Start >= StartKmerValidPos)
{
for (DeclareListPtr(p) = GetListPtr(StartKmer); NotEndOfList(p); p = GetListNext(p))
{
int HitPos = GetListPos(p);
CommonKmer(Start, HitPos);
}
}
if (0 == --Ticker)
{
TubeEnd(Start);
Ticker = TubeOffset;
}
}
TubeEnd(Qlen - 1);
int DiagFrom = CalcDiagIndex(Tlen - 1, Qlen - 1) - TubeWidth;
int DiagTo = CalcDiagIndex(0, Qlen - 1) + TubeWidth;
int TubeFrom = CalcTubeIndex(DiagFrom);
if (TubeFrom < 0)
TubeFrom = 0;
int TubeTo = CalcTubeIndex(DiagTo);
for (int TubeIndex = TubeFrom; TubeIndex <= TubeTo; ++TubeIndex)
TubeFlush(TubeIndex);
freemem(Tubes);
}
