char *CoordSystem_getNameColonVersion(CoordSystem *cs) {
if (cs && cs->nameColonVersion == NULL) {
char tmpStr[1024];
sprintf(tmpStr,"%s:%s",CoordSystem_getName(cs),CoordSystem_getVersion(cs) ? CoordSystem_getVersion(cs):"");
StrUtil_copyString(&cs->nameColonVersion, tmpStr, 0);
cs->lenNameColonVersion = strlen(tmpStr);
}
return cs? cs->nameColonVersion : NULL;
}
Vector *AssemblyMapper_listIdsImpl(AssemblyMapper *am, char *frmSeqRegionName, long frmStart, long frmEnd, CoordSystem *frmCs) {
IDType seqRegionId = AssemblyMapper_getSeqRegionId(am, frmSeqRegionName, frmCs);
AssemblyMapperAdaptor *adaptor = AssemblyMapper_getAdaptor(am);
if ( !CoordSystem_compare(frmCs, AssemblyMapper_getComponentCoordSystem(am) ) ) {
if ( !AssemblyMapper_haveRegisteredComponent(am, seqRegionId) ) {
AssemblyMapperAdaptor_registerComponent(adaptor, am, seqRegionId);
}
// Pull out the 'from' identifiers of the mapper pairs. The we
// loaded the assembled side as the 'from' side in the constructor.
MapperPairSet *mps = Mapper_listPairs( AssemblyMapper_getMapper(am), seqRegionId, frmStart, frmEnd, "component");
return MapperPairSet_getFromIds(mps);
} else if ( !CoordSystem_compare(frmCs, AssemblyMapper_getAssembledCoordSystem(am) ) ) {
AssemblyMapperAdaptor_registerAssembled(adaptor, am, seqRegionId, frmStart, frmEnd);
// Pull out the 'to' identifiers of the mapper pairs we loaded the
// component side as the 'to' coord system in the constructor.
MapperPairSet *mps = Mapper_listPairs( AssemblyMapper_getMapper(am), seqRegionId, frmStart, frmEnd, "assembled");
return MapperPairSet_getToIds(mps);
} else {
fprintf(stderr, "Coordinate system %s %s is neither the assembled nor the component coordinate system of this AssemblyMapper\n",
CoordSystem_getName(frmCs), CoordSystem_getVersion(frmCs) );
exit(1);
}
}
int CoordSystem_compare(CoordSystem *cs1, CoordSystem *cs2) {
/* Dodgy DAS stuff
if (!$cs2 || !ref($cs2) || !$cs2->isa('Bio::EnsEMBL::CoordSystem')) {
if ($cs2->isa('Bio::EnsEMBL::ExternalData::DAS::CoordSystem')) {
return $cs2->equals($self);
}
throw('Argument must be a CoordSystem');
}
*/
if (!EcoString_strcmp(CoordSystem_getVersion(cs1), CoordSystem_getVersion(cs2)) && !EcoString_strcmp(CoordSystem_getName(cs1), CoordSystem_getName(cs2))) {
// Equal, so return 0
return 0;
}
return 1;
}
MapperRangeSet *AssemblyMapper_fastMapImpl(AssemblyMapper *am, char *frmSeqRegionName, long frmStart, long frmEnd, int frmStrand, CoordSystem *frmCs, Slice *toSlice) {
Mapper *mapper = AssemblyMapper_getMapper(am);
CoordSystem *asmCs = AssemblyMapper_getAssembledCoordSystem(am);
CoordSystem *cmpCs = AssemblyMapper_getComponentCoordSystem(am);
AssemblyMapperAdaptor *adaptor = AssemblyMapper_getAdaptor(am);
char *frm;
IDType seqRegionId = AssemblyMapper_getSeqRegionId(am, frmSeqRegionName, frmCs);
// Speed critical section:
// Try to do simple pointer equality comparisons of the coord system
// objects first since this is likely to work most of the time and is
// much faster than a function call.
if ( frmCs == cmpCs
|| ( frmCs != asmCs && !CoordSystem_compare(frmCs, cmpCs)) ) {
if ( !IDHash_contains(AssemblyMapper_getComponentRegister(am), seqRegionId) ) {
AssemblyMapperAdaptor_registerComponent( adaptor, am, seqRegionId);
}
frm = "component";
} else if ( frmCs == asmCs || !CoordSystem_compare(frmCs, asmCs) ) {
// This can be probably be sped up some by only calling registered
// assembled if needed.
AssemblyMapperAdaptor_registerAssembled( adaptor, am, seqRegionId, frmStart, frmEnd);
frm = "assembled";
} else {
fprintf(stderr,"Coordinate system %s %s is neither the assembled nor the component coordinate system of this AssemblyMapper\n",
CoordSystem_getName(frmCs), CoordSystem_getVersion(frmCs) );
}
return Mapper_fastMap( mapper, seqRegionId, frmStart, frmEnd, frmStrand, frm );
}
Vector *PredictionTranscriptAdaptor_fetchAllBySlice(PredictionTranscriptAdaptor *pta, Slice *slice, char *logicName, int loadExons) {
//my $transcripts = $self->SUPER::fetch_all_by_Slice($slice,$logic_name);
Vector *transcripts = BaseFeatureAdaptor_fetchAllBySlice((BaseFeatureAdaptor *)pta, slice, logicName);
// if there are 0 or 1 transcripts still do lazy-loading
if ( ! loadExons || Vector_getNumElement(transcripts) < 2 ) {
return transcripts;
}
// preload all of the exons now, instead of lazy loading later
// faster than 1 query per transcript
// get extent of region spanned by transcripts
long minStart = 2000000000;
long maxEnd = -2000000000;
int i;
for (i=0; i<Vector_getNumElement(transcripts); i++) {
PredictionTranscript *t = Vector_getElementAt(transcripts, i);
if (PredictionTranscript_getSeqRegionStart((SeqFeature*)t) < minStart) {
minStart = PredictionTranscript_getSeqRegionStart((SeqFeature*)t);
}
if (PredictionTranscript_getSeqRegionEnd((SeqFeature*)t) > maxEnd) {
maxEnd = PredictionTranscript_getSeqRegionEnd((SeqFeature*)t);
}
}
Slice *extSlice;
if (minStart >= Slice_getStart(slice) && maxEnd <= Slice_getEnd(slice)) {
extSlice = slice;
} else {
SliceAdaptor *sa = DBAdaptor_getSliceAdaptor(pta->dba);
extSlice = SliceAdaptor_fetchByRegion(sa, Slice_getCoordSystemName(slice), Slice_getSeqRegionName(slice),
minStart, maxEnd, Slice_getStrand(slice), CoordSystem_getVersion(Slice_getCoordSystem(slice)), 0);
}
// associate exon identifiers with transcripts
IDHash *trHash = IDHash_new(IDHASH_MEDIUM);
for (i=0; i<Vector_getNumElement(transcripts); i++) {
PredictionTranscript *t = Vector_getElementAt(transcripts, i);
if ( ! IDHash_contains(trHash, PredictionTranscript_getDbID(t))) {
IDHash_add(trHash, PredictionTranscript_getDbID(t), t);
}
}
IDType *uniqueIds = IDHash_getKeys(trHash);
char tmpStr[1024];
char *qStr = NULL;
if ((qStr = (char *)calloc(655500,sizeof(char))) == NULL) {
fprintf(stderr,"Failed allocating qStr\n");
return transcripts;
}
int lenNum;
int endPoint = sprintf(qStr, "SELECT prediction_transcript_id, prediction_exon_id, exon_rank FROM prediction_exon WHERE prediction_transcript_id IN (");
for (i=0; i<IDHash_getNumValues(trHash); i++) {
if (i!=0) {
qStr[endPoint++] = ',';
qStr[endPoint++] = ' ';
}
lenNum = sprintf(tmpStr,IDFMTSTR,uniqueIds[i]);
memcpy(&(qStr[endPoint]), tmpStr, lenNum);
endPoint+=lenNum;
}
qStr[endPoint++] = ')';
qStr[endPoint] = '\0';
free(uniqueIds);
StatementHandle *sth = pta->prepare((BaseAdaptor *)pta,qStr,strlen(qStr));
sth->execute(sth);
IDHash *exTrHash = IDHash_new(IDHASH_MEDIUM);
ResultRow *row;
while ((row = sth->fetchRow(sth))) {
IDType trId = row->getLongLongAt(row,0);
IDType exId = row->getLongLongAt(row,1);
int rank = row->getIntAt(row,2);
if (! IDHash_contains(exTrHash, exId)) {
Vector *vec = Vector_new();
Vector_setFreeFunc(vec, PredictionTranscriptRankPair_free);
IDHash_add(exTrHash, exId, vec);
}
Vector *exVec = IDHash_getValue(exTrHash, exId);
PredictionTranscriptRankPair *trp = PredictionTranscriptRankPair_new(IDHash_getValue(trHash, trId), rank);
Vector_addElement(exVec, trp);
}
IDHash_free(trHash, NULL);
sth->finish(sth);
PredictionExonAdaptor *pea = DBAdaptor_getPredictionExonAdaptor(pta->dba);
Vector *exons = PredictionExonAdaptor_fetchAllBySlice(pea, extSlice);
// move exons onto transcript slice, and add them to transcripts
for (i=0; i<Vector_getNumElement(exons); i++) {
PredictionExon *ex = Vector_getElementAt(exons, i);
// Perl didn't have this line - it was in GeneAdaptor version so I think I'm going to keep it
if (!IDHash_contains(exTrHash, PredictionExon_getDbID(ex))) continue;
PredictionExon *newEx;
if (slice != extSlice) {
newEx = (PredictionExon*)PredictionExon_transfer((SeqFeature*)ex, slice);
if (newEx == NULL) {
fprintf(stderr, "Unexpected. Exon could not be transferred onto PredictionTranscript slice.\n");
exit(1);
}
} else {
newEx = ex;
}
Vector *exVec = IDHash_getValue(exTrHash, PredictionExon_getDbID(newEx));
int j;
for (j=0; j<Vector_getNumElement(exVec); j++) {
PredictionTranscriptRankPair *trp = Vector_getElementAt(exVec, j);
PredictionTranscript_addExon(trp->transcript, newEx, &trp->rank);
}
}
IDHash_free(exTrHash, Vector_free);
free(qStr);
return transcripts;
}
