int eseqclusterData::getOTU(float dist,earray<eintarray>& otus,int size)
{
eintarray otuarr;
int i,j;
otuarr.init(size);
for (i=0; i<size; ++i)
otuarr[i]=i;
// point at the seq we merged with
for (i=0; i<mergearr.size() && mergearr[i].dist>=dist; ++i)
otuarr[mergearr[i].y]=mergearr[i].x;
eintarray otuind;
otuind.init(otuarr.size(),-1);
// translate all the merged seq ids to the last seq id of the otu
for (i=0; i<otuarr.size(); ++i){
for (j=otuarr[i]; j!=otuarr[j]; j=otuarr[j]);
otuarr[i]=j;
if (otuind[j]==-1){
otuind[j]=otus.size();
otus.add(eintarray());
}
}
for (i=0; i<otuarr.size(); ++i)
otus[otuind[otuarr[i]]].add(i);
return(otus.size());
}
void eseqclustersingle::init(INDTYPE count,const estr& ofilename,const estr& seqsfile,const earray<ebasicarray<INDTYPE> >& dupslist) {
ofile.open(ofilename,"w");
ofile.write("# seqsfile: "+seqsfile+"\n");
ofile.write("# OTU_count Merge_distance Merged_OTU_id1 Merged_OTU_id2\n");
long i,j;
mergecount=0;
scount.reserve(count);
scluster.reserve(count);
smerge.reserve(count);
incluster.reserve(count);
for (i=0; i<count; ++i){
scount.add(1);
scluster.add(i);
smerge.add(-1);
incluster.add(list<INDTYPE>());
incluster[i].push_back(i);
}
for (i=0; i<dupslist.size(); ++i){
for (j=1; j<dupslist[i].size(); ++j){
++mergecount;
ofile.write(estr(scluster.size()-mergecount)+" 1.0 "+dupslist[i][0]+" "+dupslist[i][j]+"\n");
clusterData.mergearr.add(eseqdist(dupslist[i][0],dupslist[i][j],1.0));
}
}
cout << "# initializing cluster with: "<< count<< " seqs" << endl;
}
void enter_file_indicator (char *f)
{
int last;
if (!f [0]) return;
if (Filez.nr) {
last = Filez.nr - 1;
if (!strcmp (f, Filez.x [last].file)) return;
if (Filez.x [last].indx == indx + 1) {
delete [] Filez.x [last].file;
Filez.x [last].file = StrDup (f);
return;
}
}
last = Filez.alloc ();
Filez.x [last].indx = indx + 1;
Filez.x [last].file = StrDup (f);
}
void eseqclusteravg::init(INDTYPE count,const estr& filename,const estr& seqsfile,const earray<ebasicarray<INDTYPE> >& dupslist,float _thres,float (_fdist)(const estr&,const estr&,int),estrarray& _seqarr,int _seqlen)
{
thres=_thres;
seqarr=&_seqarr;
seqlen=_seqlen;
fdist=_fdist;
ofile.open(filename,"w");
ofile.write("# seqsfile: "+seqsfile+"\n");
ofile.write("# OTU_count Merge_distance Merged_OTU_id1 Merged_OTU_id2\n");
long i,j;
incmaxdist=1.0;
incmaxit=smatrix.end();
cf=0;
lastdist=0.0;
scount.reserve(count);
scluster.reserve(count);
smerge.reserve(count);
inter.reserve(count);
incluster.reserve(count);
mergecount=0;
for (i=0; i<count; ++i){
scount.add(1);
scluster.add(i);
smerge.add(-1);
incluster.add(list<INDTYPE>());
incluster[i].push_back(i);
inter.add(list<INDTYPE>());
}
for (i=0; i<dupslist.size(); ++i){
for (j=1; j<dupslist[i].size(); ++j){
++mergecount;
ofile.write(estr(scluster.size()-mergecount)+" 1.0 "+dupslist[i][0]+" "+dupslist[i][j]+"\n");
clusterData.mergearr.add(eseqdist(dupslist[i][0],dupslist[i][j],1.0));
}
}
cout << "# initializing cluster with: "<< count<< " seqs" << endl;
cout << "# initializing smatrix with: " << (long)(count)*(long)(count)/20000l/2l<< " elements" << endl;
smatrix.reserve((long)(count)*(long)(count)/20000l/2l);
// cout << "# smatrix._hashitems = " << smatrix._hashitems << endl;
}
void finduniq(ebasicarray<INDTYPE>& uniqind,earray<ebasicarray<INDTYPE> >& dupslist)
{
ebasicstrhashof<long> duphash;
ebasicstrhashof<long>::iter it;
if (!ignoreUnique){
duphash.reserve(arr.size());
for (long i=0; i<arr.size(); ++i){
if (i%1000==0)
fprintf(stderr,"\r%li/%li",i,(long)arr.size());
it=duphash.get(arr.values(i));
if (it==duphash.end())
{ uniqind.add(i); duphash.add(arr.values(i),uniqind.size()-1); dupslist.add(ebasicarray<INDTYPE>(i)); }
else
dupslist[it.value()].add(i);
}
fprintf(stderr,"\r%li\n",(long)arr.size());
}else{
uniqind.init(arr.size());
for (long i=0; i<uniqind.size(); ++i)
uniqind[i]=i;
}
cout << endl;
}
void make_norm ()
{
if (string) _enter_token (enter_string ());
used_builtins ();
//
C_Ntok = tmpCODE.nr ();
CODE = new int [C_Ntok + 3];
tmpCODE.copy (&CODE);
CODE [C_Ntok] = FORCEERROR;
CODE [C_Ntok + 1] = ';';
CODE [C_Ntok + 2] = FORCEERROR;
tmpCODE.destroy ();
//
C_Syms = new char* [C_Nsyms = symtree.nnodes - nreserved];
symtree.deltree (symtoarray);
Filez.freeze ();
C_Files = Filez.x;
C_Nfiles = Filez.nr;
C_Nlines = Linez.nr ();
Linez.copy (&C_Lines);
Linez.destroy ();
C_Strings = new char* [C_Nstrings = strtree.nnodes];
strtree.deltree (strtoarray);
//
TFloat.copy (&C_Floats);
TFloat.destroy ();
TsInt8.copy (&C_Chars);
TsInt8.destroy ();
TsInt16.copy (&C_Shortints);
TsInt16.destroy ();
TsInt32.copy (&C_Ints);
TsInt32.destroy ();
TuInt32.copy (&C_Unsigned);
TuInt32.destroy ();
}
friend bool operator == (const earray& l, const earray& r)
{
return std::equal(l.begin(), l.end(), r.begin());
}
int eseqclusterData::getTree(const efloatarray& dists,earray<earray<eintarray> >& otus,earray<eintarray>& otusanc,int size)
{
if (dists.size()==0) return(0);
eintarray otuarr;
int i,j,k;
otuarr.init(size);
for (i=0; i<size; ++i)
otuarr[i]=i;
otus.init(dists.size());
otusanc.init(dists.size());
// point at the seq we merged with
for (k=0; k<mergearr.size() && mergearr[k].dist>=dists[0]; ++k)
otuarr[mergearr[k].y]=mergearr[k].x;
eintarray otuind,otuind2;
otuind.init(otuarr.size(),-1);
// translate all the merged seq ids to the last seq id of the otu
for (i=0; i<otuarr.size(); ++i){
for (j=otuarr[i]; j!=otuarr[j]; j=otuarr[j]);
otuarr[i]=j;
if (otuind[j]==-1){
otuind[j]=otus[0].size();
otus[0].add(eintarray());
}
otus[0][otuind[j]].add(i);
}
otusanc[0].init(otus[0].size());
for (i=0; i<otusanc[0].size(); ++i)
otusanc[0][i]=i;
int l;
eintarray otuarr2;
for (l=1; l<dists.size(); ++l){
otusanc[l].init(otus[0].size());
otuarr2.init(otus[l-1].size());
for (i=0; i<otuarr2.size(); ++i)
otuarr2[i]=i;
for (; k<mergearr.size() && mergearr[k].dist>=dists[l]; ++k)
otuarr2[otusanc[l-1][otuind[otuarr[mergearr[k].y]]]]=otusanc[l-1][otuind[otuarr[mergearr[k].x]]];
otuind2.init(otuarr2.size(),-1);
for (i=0; i<otuarr2.size(); ++i){
for (j=otuarr2[i]; j!=otuarr2[j]; j=otuarr2[j]);
otuarr2[i]=j;
if (otuind2[j]==-1){
otuind2[j]=otus[l].size();
otus[l].add(eintarray());
}
otus[l][otuind2[j]].add(i);
}
for (i=0; i<otus[0].size(); ++i){
j=otuarr2[otusanc[l-1][i]];
otusanc[l][i]=otuind2[j];
}
}
return(otus[0].size());
}