static void HitTube(int TubeIndex, int q)
{
TubeState *Tube = Tubes + TubeIndex%MaxActiveTubes;
#if TRACE
Log("HitTube(TubeIndex=%d, q=%d) Count=%d\n",
TubeIndex, q, Tube->Count);
#endif
if (0 == Tube->Count)
{
Tube->Count = 1;
Tube->qLo = q;
Tube->qHi = q;
return;
}
if (q - Tube->qHi > MaxKmerDist)
{
if (Tube->Count >= MinKmersPerHit)
AddHit(TubeIndex, Tube->qLo, Tube->qHi);
Tube->Count = 1;
Tube->qLo = q;
Tube->qHi = q;
return;
}
++(Tube->Count);
Tube->qHi = q;
}
int CLocalSearch::ExecuteSharedFiles(int nMaximum)
{
CQuickLock oLock( Library.m_pSection );
CPtrList* pFiles = Library.Search( m_pSearch, nMaximum );
if ( pFiles == NULL ) return 0;
int nHits = pFiles->GetCount();
while ( pFiles->GetCount() )
{
int nInThisPacket = min( pFiles->GetCount(), (int)Settings.Gnutella.HitsPerPacket );
CreatePacket( nInThisPacket );
int nHitB = 0;
for ( int nHitA = 0 ; nHitA < nInThisPacket ; nHitA++ )
{
CLibraryFile* pFile = (CLibraryFile*)pFiles->RemoveHead();
if ( AddHit( pFile, nHitB ) ) nHitB ++;
}
WriteTrailer();
if ( nHitB > 0 ) DispatchPacket(); else DestroyPacket();
}
delete pFiles;
return nHits;
}
// Called when q=Qlen - 1 to flush any hits in each tube.
static void TubeFlush(int TubeIndex)
{
TubeState *Tube = Tubes + TubeIndex%MaxActiveTubes;
#if TRACE
Log("TubeFlush(TubeIndex=%d) Count=%d\n",
TubeIndex, Tube->Count);
#endif
if (Tube->Count < MinKmersPerHit)
return;
AddHit(TubeIndex, Tube->qLo, Tube->qHi);
Tube->Count = 0;
}
void CLocalSearch::SendHits(const CList< T* >& oFiles)
{
CPacket* pPacket = NULL;
CSchemaMap pSchemas;
BYTE nHits = 0;
for ( POSITION pos = oFiles.GetHeadPosition(); pos; )
{
if ( ! pPacket )
pPacket = CreatePacket();
AddHit( pPacket, pSchemas, oFiles.GetNext( pos ), nHits ++ );
bool bSend = false;
switch ( m_nProtocol )
{
case PROTOCOL_G1:
case PROTOCOL_G2:
bSend = nHits >= Settings.Gnutella.HitsPerPacket ||
pPacket->m_nLength >= MAX_QUERY_PACKET_SIZE;
break;
case PROTOCOL_DC:
// One hit per packet in DC++ protocol
bSend = true;
break;
default:
ASSERT( FALSE );
}
// Send full packet
if ( bSend )
{
WriteTrailer( pPacket, pSchemas, nHits );
DispatchPacket( pPacket );
pPacket = NULL;
nHits = 0;
}
}
if ( nHits )
{
WriteTrailer( pPacket, pSchemas, nHits );
DispatchPacket( pPacket );
pPacket = NULL;
}
ASSERT( pPacket == NULL );
ASSERT( pSchemas.IsEmpty() );
}
// Called when end of a tube is reached
// A point in the tube -- the point with maximal q -- is (Tlen-1,q-1).
static void TubeEnd(int q)
{
int DiagIndex = CalcDiagIndex(Tlen - 1, q - 1);
int TubeIndex = CalcTubeIndex(DiagIndex);
TubeState *Tube = Tubes + TubeIndex%MaxActiveTubes;
#if TRACE
Log("TubeEnd(%d) DiagIndex=%d TubeIndex=%d Count=%d\n",
q, DiagIndex, TubeIndex, Tube->Count);
#endif
if (Tube->Count >= MinKmersPerHit)
AddHit(TubeIndex, Tube->qLo, Tube->qHi);
Tube->Count = 0;
}
void SingleAlign::SnpAlign(RefSeq &ref, bit32_t mode) {
bit32_t i,j,m, modeindex, cmodeindex, mc, h, read_chain_index_mask;
Hit prefetch_hit, _hHit;
//cout<<_pread->seq<<endl;
if(param.RRBS_flag){ //RRBS mode
for(read_chain_index=0; read_chain_index<2;read_chain_index++){
if(!xflag_chain[read_chain_index]) continue;
cmodeindex=modeindex=xseedindex[read_chain_index][mode].second;
if(read_chain_index) cmodeindex=map_readlen/param.seed_size-1-modeindex;
_seed=xseeds[read_chain_index][modeindex][0];
//cout<<"mode:"<<mode<<" m="<<ref.index[_seed].n1<<" seed:"<<_seed;
//cout<<" "; disp_bfa(param.map3to4(_seed), param.seed_size); cout<<endl;
if((m=ref.index[_seed].n1)==0) continue;
_refloc=ref.index[_seed].loc1; //list of seeded reference locations
h=param.profile[modeindex][0]+cseed_offset*read_chain_index;
read_chain_index_mask=read_chain_index<<24;
for(j=0; j!=m; j++) {
_hHit=_refloc[j];
if(((_hHit.chr^read_chain_index_mask)>>16)!=cmodeindex) continue; // mode or strand not match
_hit.chr=_hHit.chr&0xffff;
if(_hHit.loc<h) continue; //underflow the start of refseq
_hit.loc=_hHit.loc-h;
//cout<<" j="<<j<<" chr"<<(int)_hit.chr<<":"<<_hit.loc<<endl;
CountMismatch(xseq[read_chain_index], (_hit.loc%SEGLEN)<<1, ref.bfa[_hit.chr].s+_hit.loc/SEGLEN);
//cout<<" mis:"<<tmp_snp<<endl;
if(tmp_snp<=snp_thres) {
_ghit=int2hit(ref,_hit,0,0);
if(!param.pairend){
seg_info=ref.CCGG_seglen(_ghit.chr, _ghit.loc, map_readlen); //get fragment information
//cout<<"seg1:"<<seg_info.first<<" seg2:"<<seg_info.second<<endl;
//if(seg_info.second>param.max_insert) continue; // fragment too large
}
if(AddHit(ref, tmp_snp, mode)) return;
}
if(param.gap>0) {
if(GapAlign(ref, mode, h)) return;
}
}
}
}
else{ //WGBS mode
for(read_chain_index=0; read_chain_index<2;read_chain_index++){
void TGgSchRSet::Merge(const PGgSchRSet& RSet){
if (RSet.Empty()){return;}
// create hash table of existing urls
TStrH TitleStrH(GetHits());
for (int HitN=0; HitN<GetHits(); HitN++){
TitleStrH.AddKey(GetHit(HitN)->TitleStr);
}
// merge hits
for (int HitN=0; HitN<RSet->GetHits(); HitN++){
PGgSchRef Ref=RSet->GetHit(HitN);
if (!TitleStrH.IsKey(Ref->TitleStr)){
AddHit(Ref);
TitleStrH.AddKey(Ref->TitleStr);
}
}
// reset fixed fields
PutNextUrlStr("");
PutAllHits(-1);
}
void TRSet::Merge(const PRSet& RSet){
if (RSet.Empty()){return;}
// create hash table of existing urls
TStrH UrlStrH(GetHits());
for (int HitN=0; HitN<GetHits(); HitN++){
UrlStrH.AddKey(GetHitUrlStr(HitN));
}
// merge hits
for (int HitN=0; HitN<RSet->GetHits(); HitN++){
TStr HitUrlStr; TStr HitTitleStr; TStr HitSrcNm; TStr HitCtxStr;
RSet->GetHit(HitN, HitUrlStr, HitTitleStr, HitSrcNm, HitCtxStr);
if (!UrlStrH.IsKey(HitUrlStr)){
AddHit(HitUrlStr, HitTitleStr, HitSrcNm, HitCtxStr);
UrlStrH.AddKey(HitUrlStr);
}
}
// reset fixed fields
PutNextUrlStr("");
PutAllHits(-1);
}
int CLocalSearch::ExecutePartialFiles(int nMaximum)
{
ASSERT( m_nProtocol == PROTOCOL_G2 );
ASSERT( m_pSearch != NULL );
if ( m_pSearch->m_bTiger == FALSE && m_pSearch->m_bSHA1 == FALSE &&
m_pSearch->m_bED2K == FALSE && m_pSearch->m_bBTH == FALSE ) return 0;
CSingleLock pLock( &Transfers.m_pSection );
if ( ! pLock.Lock( 50 ) ) return 0;
int nCount = 0;
m_pPacket = NULL;
for ( POSITION pos = Downloads.GetIterator() ; pos ; )
{
CDownload* pDownload = Downloads.GetNext( pos );
if ( ! pDownload->IsShared() ) continue;
if ( ( m_pSearch->m_bTiger && pDownload->m_bTiger && m_pSearch->m_pTiger == pDownload->m_pTiger )
|| ( m_pSearch->m_bSHA1 && pDownload->m_bSHA1 && m_pSearch->m_pSHA1 == pDownload->m_pSHA1 )
|| ( m_pSearch->m_bED2K && pDownload->m_bED2K && m_pSearch->m_pED2K == pDownload->m_pED2K )
|| ( m_pSearch->m_bBTH && pDownload->m_bBTH && m_pSearch->m_pBTH == pDownload->m_pBTH ) )
{
if ( pDownload->m_bBTH || pDownload->IsStarted() )
{
if ( m_pPacket == NULL ) CreatePacketG2();
AddHit( pDownload, nCount++ );
}
}
}
if ( m_pPacket != NULL )
{
WriteTrailerG2();
DispatchPacket();
}
return nCount;
}
void Tan()
{
// Image file annotated with from-to pile indexes
// Produced by:
// piler2 -trs banded_hits.gff -images mainband_images.gff
const char *HitFileName = RequiredValueOpt("tan");
const char *OutFileName = RequiredValueOpt("out");
const char *PyramidFileName = ValueOpt("pyramid");
const char *MotifFileName = ValueOpt("motif");
const char *strMinHits = ValueOpt("minhits");
const char *strMaxMargin = ValueOpt("maxmargin");
const char *strMinRatio = ValueOpt("minratio");
if (0 != strMinHits)
MIN_HIT_COUNT = atoi(strMinHits);
if (0 != strMaxMargin)
MAX_FRACT_MARGIN = atof(strMaxMargin);
if (0 != strMinRatio)
MIN_RATIO = atof(strMinRatio);
FILE *fInput = OpenStdioFile(HitFileName);
ProgressStart("Initialize piles");
GFFRecord Rec;
int HitCount = 0;
while (GetNextGFFRecord(fInput, Rec))
{
if (0 != strcmp(Rec.Feature, "hit"))
continue;
int QueryPileIndex = -1;
int TargetPileIndex = -1;
ParsePilesAttrs(Rec.Attrs, &QueryPileIndex, &TargetPileIndex);
if (QueryPileIndex != TargetPileIndex)
continue;
char TargetLabel[128];
int TargetStart;
int TargetEnd;
ParseTargetAttrs(Rec.Attrs, TargetLabel, sizeof(TargetLabel), &TargetStart, &TargetEnd);
if (0 != strcmp(Rec.SeqName, TargetLabel))
Quit("Labels don't match");
const int QueryFrom = Rec.Start - 1;
const int QueryTo = Rec.End - 1;
const int TargetFrom = TargetStart - 1;
const int TargetTo = TargetEnd - 1;
const bool Rev = (Rec.Strand == '-');
AddHit(QueryPileIndex, Rec.SeqName, QueryFrom, QueryTo, TargetFrom, TargetTo, Rev);
++HitCount;
}
ProgressDone();
Progress("%d hits, %d piles", HitCount, PileCount);
ProgressStart("Allocate piles");
for (int PileIndex = 0; PileIndex < PileCount; ++PileIndex)
{
TanPile &Pile = Piles[PileIndex];
Pile.Hits = all(HitData, Pile.HitCount);
Pile.HitCount = 0;
}
ProgressDone();
ProgressStart("Assign hits to piles");
Rewind(fInput);
while (GetNextGFFRecord(fInput, Rec))
{
if (0 != strcmp(Rec.Feature, "hit"))
continue;
int QueryPileIndex = -1;
int TargetPileIndex = -1;
ParsePilesAttrs(Rec.Attrs, &QueryPileIndex, &TargetPileIndex);
if (QueryPileIndex != TargetPileIndex)
continue;
char TargetLabel[128];
int TargetStart;
int TargetEnd;
ParseTargetAttrs(Rec.Attrs, TargetLabel, sizeof(TargetLabel), &TargetStart, &TargetEnd);
if (0 != strcmp(Rec.SeqName, TargetLabel))
Quit("Labels don't match");
const int QueryFrom = Rec.Start - 1;
const int QueryTo = Rec.End - 1;
const int TargetFrom = TargetStart - 1;
const int TargetTo = TargetEnd - 1;
const bool Rev = (Rec.Strand == '-');
AssignHit(QueryPileIndex, Rec.SeqName, QueryFrom, QueryTo, TargetFrom, TargetTo, Rev);
}
ProgressDone();
fOut = OpenStdioFile(OutFileName, FILEIO_MODE_WriteOnly);
fPyramid = (0 == PyramidFileName ? 0 : OpenStdioFile(PyramidFileName, FILEIO_MODE_WriteOnly));
fMotif = (0 == PyramidFileName ? 0 : OpenStdioFile(MotifFileName, FILEIO_MODE_WriteOnly));
ProgressStart("Find pyramids");
for (int PileIndex = 0; PileIndex < PileCount; ++PileIndex)
FindPyramids(PileIndex);
int PyramidCount = PyramidIndex;
ProgressDone();
Progress("%d pyramids", PyramidCount);
}
PyObject * PyHit_PushWord(PyObject * pSelf, PyObject* args, PyObject* kwargs)
{
/*
= You can change the iPos & iPhrase setting without add a real hit by push a Empty word("").
@param:
- pos:
- phrase:
*/
csfHelper_HitCollectorObject *self = (csfHelper_HitCollectorObject *)pSelf;
static char *kwlist[] = {"word", "pos", "phrase", "fieldindex", NULL};
const char* sWord = NULL;
int iPos = -1;
int iPhrase = -1;
int iField = -1;
int ok = PyArg_ParseTupleAndKeywords( args, kwargs, "z|iii", kwlist, &sWord,
&iPos , &iPhrase, &iField ); //not inc the value refer
if(!ok) {
if(PyErr_Occurred()) {
PyErr_Print();
PyErr_Clear();
}
return NULL;
}
if(iPos < 1 ) iPos = -1; // 0 and < is invalid.
if(iPhrase <1) iPhrase = -1;
//fast setting
if(!sWord) {
if(iPos != -1)
self->iPos = iPos;
if(iPhrase == -1)
self->iPhrase = iPhrase; //refresh iPos
RET_PYNONE;
}
if(iPos == -1)
iPos = self->iPos;
else
{
if(iField == -1)
self->iPos = iPos; //refresh iPos
}
// no more than 24, if greater than 24, just leave field as current.
if(iField > SPH_MAX_FIELDS || iField < 0)
iField = -1;
if(iPhrase == -1)
iPhrase = self->iPhrase;
else
{
if(iField == -1)
self->iPhrase = iPhrase; //refresh iPos
}
if(iField == -1)
iField = self->iField; //NO, NO, u can NOT change field via a filled parameters' pushWord call...
{
CSphSource_Python* pPySource = dynamic_cast<CSphSource_Python*>(self->m_pSource);
BYTE sBuf [ 16+3*SPH_MAX_WORD_LEN + 1024 ];
if(pPySource){
CSphDict* pDict = pPySource->GetDict();
bool bWordDict = pDict->GetSettings().m_bWordDict;
bool bGlobalPartialMatch = !bWordDict && ( pPySource->m_iMinPrefixLen > 0 || pPySource->m_iMinInfixLen > 0 );
{
SphWordID_t iWord = pDict->GetWordID ((BYTE*) sWord );
iPos = HITMAN::Create(iField, iPos);
AddHit(self, iWord, iPos, iPhrase);
}
if ( bGlobalPartialMatch )
{
int iBytes = strlen ( (const char*)sWord );
memcpy ( sBuf + 1, sWord, iBytes );
sBuf[0] = MAGIC_WORD_HEAD;
sBuf[iBytes+1] = '\0';
AddHit(self, pDict->GetWordIDWithMarkers ( sBuf ), iPos, iPhrase);
}
if ( pPySource->m_bIndexExactWords )
{
int iBytes = strlen ( (const char*)sWord );
memcpy ( sBuf + 1, sWord, iBytes );
sBuf[0] = MAGIC_WORD_HEAD_NONSTEMMED;
sBuf[iBytes+1] = '\0';
AddHit(self, pDict->GetWordIDNonStemmed ( sBuf ), iPos, iPhrase);
}
self->iPos++ ;
}
}
RET_PYNONE;
}
PyObject * PyHit_Push(PyObject * pSelf, PyObject* args, PyObject* kwargs)
{
//int iPos = HIT_PACK ( iField, iStartPos );
csfHelper_HitCollectorObject *self = (csfHelper_HitCollectorObject *)pSelf;
static char *kwlist[] = {"wordid", "pos", "phrase", "fieldindex", NULL};
SphWordID_t iWord;
int iPos = -1;
int iPhrase = -1;
int iField = -1;
#if USE_64BIT
int ok = PyArg_ParseTupleAndKeywords( args, kwargs, "K|iii", kwlist, &iWord,
#else
int ok = PyArg_ParseTupleAndKeywords( args, kwargs, "k|iii", kwlist, &iWord,
#endif
&iPos , &iPhrase, &iField ); //not inc the value refer
if(!ok) {
if(PyErr_Occurred()) {
PyErr_Print();
//PyErr_Clear();
}
return NULL;
}
if(iPos < 1 ) iPos = -1; // 0 and < is invalid.
if(iPhrase <1) iPhrase = -1;
// fast setting
if(!iWord) {
if(iPos != -1)
self->iPos = iPos;
if(iPhrase == -1)
self->iPhrase = iPhrase; //refresh iPos
RET_PYNONE;
}
// push hit
if(iPos == -1)
iPos = self->iPos;
else
{
if(iField == -1)
self->iPos = iPos; //refresh iPos
}
// no more than 24, if greater than 24, just leave field as current.
if(iField > SPH_MAX_FIELDS || iField < 0)
iField = -1;
if(iPhrase == -1)
iPhrase = self->iPhrase;
else
{
if(iField == -1)
self->iPhrase = iPhrase; //refresh iPos
}
if(iField == -1)
iField = self->iField; //NO, NO, u can NOT change field via a filled parameters' pushWord call...
{
iPos = HITMAN::Create(iField, iPos);
AddHit(self, iWord, iPos, iPhrase);
self->iPos++ ;
}
RET_PYNONE;
}
