Full documentation can be found at (graphv section):\n\
http://oss.oetiker.ch/rrdtool/doc/rrdgraph.en.html";
static PyObject *
_rrdtool_graphv(PyObject *Py_UNUSED(self), PyObject *args)
{
PyObject *ret;
rrd_info_t *data;
if (convert_args("graphv", args) == -1)
return NULL;
Py_BEGIN_ALLOW_THREADS
data = rrd_graph_v(rrdtool_argc, rrdtool_argv);
Py_END_ALLOW_THREADS
if (data == NULL) {
PyErr_SetString(rrdtool_OperationalError, rrd_get_error());
rrd_clear_error();
ret = NULL;
} else {
ret = _rrdtool_util_info2dict(data);
rrd_info_free(data);
}
destroy_args();
return ret;
}
static PyObject *
_rrdtool_updatev(PyObject *Py_UNUSED(self), PyObject *args)
{
PyObject *ret;
rrd_info_t *data;
if (convert_args("updatev", args) == -1)
return NULL;
Py_BEGIN_ALLOW_THREADS
data = rrd_update_v(rrdtool_argc, rrdtool_argv);
Py_END_ALLOW_THREADS
if (data == NULL) {
PyErr_SetString(rrdtool_OperationalError, rrd_get_error());
rrd_clear_error();
ret = NULL;
} else {
ret = _rrdtool_util_info2dict(data);
rrd_info_free(data);
}
destroy_args();
return ret;
}
void print(int argc,char **argv){
if(argc<1){
fprintf(stderr,"\t-> Must supply afile.saf.idx files \n");
fprintf(stderr,"\t-> Examples \n");
fprintf(stderr,"\t-> ./ngsPSMC print pop1.saf.idx \n");
fprintf(stderr,"\t-> ./ngsPSMC print pop1.saf.idx -r chr1:10000000-12000000\n");
return;
}
args *pars = getArgs(argc,argv);
writepsmc_header(stderr,pars->perc);
for(myMap::iterator it=pars->perc->mm.begin();it!=pars->perc->mm.end();++it){
if(pars->chooseChr!=NULL)
it = iter_init(pars->perc,pars->chooseChr,pars->start,pars->stop);
else
it = iter_init(pars->perc,it->first,pars->start,pars->stop);
for(size_t s=pars->perc->first;s<pars->perc->last;s++)
fprintf(stdout,"%s\t%d\t%e\t%e\n",it->first,pars->perc->pos[s]+1,pars->perc->gls[2*s],pars->perc->gls[2*s+1]);
if(pars->chooseChr!=NULL)
break;
}
destroy_args(pars);
}
void print(int argc,char **argv){
if(argc<1){
fprintf(stderr,"\t-> Must supply afile.saf.idx files \n");
fprintf(stderr,"\t-> Examples \n");
fprintf(stderr,"\t-> ./realSFS print pop1.saf.idx \n");
fprintf(stderr,"\t-> ./realSFS print pop1.saf.idx -r chr1:10000000-12000000\n");
fprintf(stderr,"\t-> ./realSFS print pop1.saf.idx pop2.saf.idx -r chr2:10000000-12000000\n");
fprintf(stderr,"\t-> You can generate the oldformat by appending the -oldout 1 to the print command like\n");
fprintf(stderr,"\t-> ./realSFS print pop1.saf.idx pop2.saf.idx -oldout 1\n");
return;
}
args *pars = getArgs(argc,argv);
for(int i=0;i<pars->saf.size();i++)
pars->saf[0]->kind = 2;
if(1||pars->saf.size()!=1){
fprintf(stderr,"\t-> Will jump to multisaf printer and will only print intersecting sites between populations\n");
printMulti<T>(pars);
return;
}
writesaf_header(stderr,pars->saf[0]);
float *flt = new float[pars->saf[0]->nChr+1];
for(myMap::iterator it=pars->saf[0]->mm.begin();it!=pars->saf[0]->mm.end();++it){
if(pars->chooseChr!=NULL)
it = iter_init(pars->saf[0],pars->chooseChr,pars->start,pars->stop);
else
it = iter_init(pars->saf[0],it->first,pars->start,pars->stop);
size_t ret;
int pos;
while((ret=iter_read(pars->saf[0],flt,sizeof(float)*(pars->saf[0]->nChr+1),&pos))){
fprintf(stdout,"%s\t%d",it->first,pos+1);
for(int is=0;is<pars->saf[0]->nChr+1;is++)
fprintf(stdout,"\t%f",flt[is]);
fprintf(stdout,"\n");
}
if(pars->chooseChr!=NULL)
break;
}
delete [] flt;
destroy_args(pars);
}
int print_header(int argc,char **argv){
if(argc<1){
fprintf(stderr,"Must supply afile.saf.idx \n");
return 0;
}
args *pars = getArgs(argc,argv);
if(!pars)
return 1;
writepsmc_header(stdout,pars->perc);
destroy_args(pars);
return 0;
}
static PyObject *PyRRD_graph(
PyObject UNUSED(*self),
PyObject * args)
{
PyObject *r;
char **argv, **calcpr;
int argc, xsize, ysize, i;
double ymin, ymax;
if (create_args("graph", args, &argc, &argv) < 0)
return NULL;
if (rrd_graph(argc, argv, &calcpr, &xsize, &ysize, NULL, &ymin, &ymax) ==
-1) {
PyErr_SetString(ErrorObject, rrd_get_error());
rrd_clear_error();
r = NULL;
} else {
r = PyTuple_New(3);
PyTuple_SET_ITEM(r, 0, PyInt_FromLong((long) xsize));
PyTuple_SET_ITEM(r, 1, PyInt_FromLong((long) ysize));
if (calcpr) {
PyObject *e, *t;
e = PyList_New(0);
PyTuple_SET_ITEM(r, 2, e);
for (i = 0; calcpr[i]; i++) {
t = PyString_FromString(calcpr[i]);
PyList_Append(e, t);
Py_DECREF(t);
rrd_freemem(calcpr[i]);
}
rrd_freemem(calcpr);
} else {
Py_INCREF(Py_None);
PyTuple_SET_ITEM(r, 2, Py_None);
}
}
destroy_args(&argv);
return r;
}
int main( int argc, char *argv[] )
{
Huint i;
argument *arg;
hthread_t tid[WORKERS + DISPATCHERS];
arg = create_args();
for(i = 0;i<WORKERS;i++) tid[i] = create_worker( arg );
for(i = 0;i<DISPATCHERS;i++) tid[i+WORKERS] = create_dispatcher( arg );
for(i = 0;i<WORKERS;i++) hthread_join(tid[i], NULL);
for(i = 0;i<DISPATCHERS;i++) hthread_join(tid[i+WORKERS], NULL);
destroy_args( arg );
printf( "--QED--\n" );
return 1;
}
int print_header(int argc,char **argv){
if(argc<1){
fprintf(stderr,"Must supply afile.saf.idx \n");
return 0;
}
args *pars = getArgs(argc,argv);
if(!pars)
return 1;
if(pars->saf.size()!=1){
fprintf(stderr,"print_header only implemeted for single safs\n");
exit(0);
}
writesaf_header(stdout,pars->saf[0]);
destroy_args(pars);
return 0;
}
static PyObject *
PyRRD_last(PyObject UNUSED(*self), PyObject *args)
{
PyObject *r;
int argc, ts;
char **argv;
if (create_args("last", args, &argc, &argv) < 0)
return NULL;
if ((ts = rrd_last(argc, argv)) == -1) {
PyErr_SetString(ErrorObject, rrd_get_error());
rrd_clear_error();
r = NULL;
} else
r = PyInt_FromLong((long)ts);
destroy_args(&argv);
return r;
}
static PyObject
*PyRRD_flushcached(PyObject UNUSED(*self), PyObject * args) {
PyObject *r;
int argc;
char **argv;
if (create_args("flushcached", args, &argc, &argv) < 0)
return NULL;
if (rrd_flushcached(argc, argv) != 0) {
PyErr_SetString(ErrorObject, rrd_get_error());
rrd_clear_error();
r = NULL;
}
else {
Py_INCREF(Py_None);
r = Py_None;
}
destroy_args(&argv);
return r;
}
static PyObject *
PyRRD_resize(PyObject UNUSED(*self), PyObject *args)
{
PyObject *r;
char **argv;
int argc, ts;
if (create_args("resize", args, &argc, &argv) < 0)
return NULL;
if ((ts = rrd_resize(argc, argv)) == -1) {
PyErr_SetString(ErrorObject, rrd_get_error());
rrd_clear_error();
r = NULL;
} else {
Py_INCREF(Py_None);
r = Py_None;
}
destroy_args(&argv);
return r;
}
static PyObject
*PyRRD_updatev(PyObject UNUSED(*self), PyObject * args) {
PyObject *r;
int argc;
char **argv;
rrd_info_t *data;
if (create_args("updatev", args, &argc, &argv) < 0)
return NULL;
if ((data = rrd_update_v(argc, argv)) == NULL) {
PyErr_SetString(ErrorObject, rrd_get_error());
rrd_clear_error();
r = NULL;
}
else {
r = PyDict_FromInfo(data);
rrd_info_free(data);
}
destroy_args(&argv);
return r;
}
int fst_index(int argc,char **argv){
if(argc<1){
fprintf(stderr,"Must supply afile.saf.idx [chrname, write more info]\n");
return 0;
}
args *arg = getArgs(argc,argv);
if(!arg->fstout){
fprintf(stderr,"\t-> Must supply -fstout for doing fstindex\n");
return 0;
}
std::vector<persaf *> &saf =arg->saf;
//assert(saf.size()==2);
size_t nSites = arg->nSites;
if(nSites == 0){//if no -nSites is specified
nSites = 100000;//<- set default to 100k sites, no need to load everything...
// nSites=nsites(saf,arg);
}
fprintf(stderr,"\t-> nSites: %lu\n",nSites);
std::vector<Matrix<float> *> gls;
for(int i=0;i<saf.size();i++)
gls.push_back(alloc<float>(nSites,saf[i]->nChr+1));
// int ndim= parspace(saf);
if(arg->sfsfname.size()!=choose(saf.size(),2)){
fprintf(stderr,"\t-> You have supplied: %lu populations, that is %d pairs\n",saf.size(),choose(saf.size(),2));
fprintf(stderr,"\t-> You therefore need to supply %d 2dsfs priors instead of:%lu\n",choose(saf.size(),2),arg->sfsfname.size());
exit(0);
}
std::vector<double *> sfs;
int inc =0;
for(int i=0;i<saf.size();i++)
for(int j=i+1;j<saf.size();j++){
size_t pairdim = (saf[i]->nChr+1)*(saf[j]->nChr+1);
double *ddd=new double[pairdim];
readSFS(arg->sfsfname[inc],pairdim,ddd);
normalize(ddd,pairdim);
sfs.push_back(ddd);
inc++;
}
double **a1,**b1;
a1=new double*[choose(saf.size(),2)];
b1=new double*[choose(saf.size(),2)];
inc=0;
for(int i=0;i<saf.size();i++)
for(int j=i+1;j<saf.size();j++){
calcCoef((int)saf[i]->nChr,(int)saf[j]->nChr,&a1[inc],&b1[inc]);
// fprintf(stderr,"a1[%d]:%p b1[%d]:%p\n",inc,&a1[inc][0],inc,&b1[inc][0]);
inc++;
}
BGZF *fstbg = openFileBG(arg->fstout,".fst.gz");
FILE *fstfp = openFile(arg->fstout,".fst.idx");
char buf[8]="fstv1";
bgzf_write(fstbg,buf,8);
fwrite(buf,1,8,fstfp);
#if 0
for(int i=0;i<ndim;i++)
fprintf(stdout,"%f %f\n",a1[i],b1[i]);
exit(0);
#endif
#if 1
size_t nsafs=saf.size();
fwrite(&nsafs,sizeof(size_t),1,fstfp);
for(int i=0;i<nsafs;i++){
size_t clen= strlen(saf[i]->fname);
fwrite(&clen,sizeof(size_t),1,fstfp);
fwrite(saf[i]->fname,1,clen,fstfp);
}
#endif
int asdf = choose(saf.size(),2);
std::vector<double> *ares = new std::vector<double> [choose(saf.size(),2)];
std::vector<double> *bres = new std::vector<double> [choose(saf.size(),2)];
// for(int i=0;i<3;i++)
// fprintf(stderr,"ares.size():%lu bres.size():%lu sfs:%p\n",ares[i].size(),bres[i].size(),&sfs[i][0]);
std::vector<int> posi;
setGloc(saf,nSites);
int *posiToPrint = new int[nSites];
for(myMap::iterator it = saf[0]->mm.begin();it!=saf[0]->mm.end();++it) {
// fprintf(stderr,"doing chr:%s\n",it->first);
if(arg->chooseChr!=NULL){
it = saf[0]->mm.find(arg->chooseChr);
if(it==saf[0]->mm.end()){
fprintf(stderr,"Problem finding chr: %s\n",arg->chooseChr);
break;
}
}
for(int i=0;i<choose(saf.size(),2);i++){
ares[i].clear();
bres[i].clear();
}
posi.clear();
while(1) {
int ret=readdata(saf,gls,nSites,it->first,arg->start,arg->stop,posiToPrint,NULL);//read nsites from data
// fprintf(stderr,"ret:%d glsx:%lu\n",ret,gls[0]->x);
//if(gls[0]->x!=nSites&&arg->chooseChr==NULL&&ret!=-3){
//fprintf(stderr,"continue continue\n");
// continue;
//}
fprintf(stderr,"\t-> Will now do fst temp dump using a chunk of %lu\n",gls[0]->x);
int inc=0;
for(int i=0;i<saf.size();i++)
for(int j=i+1;j<saf.size();j++){
// fprintf(stderr,"i:%d j:%d inc:%d gls[i]:%p gls[j]:%p sfs:%p a1:%p b1:%p\n",i,j,inc,gls[i],gls[j],sfs[i],&a1[inc][0],&a1[inc][0]);
block_coef(gls[i],gls[j],sfs[inc],a1[inc],b1[inc],ares[inc],bres[inc]);
inc++;
}
for(int i=0;i<gls[0]->x;i++)
posi.push_back(posiToPrint[i]);
for(int i=0;i<gls.size();i++)
gls[i]->x =0;
if(ret==-2)//no more data in files or in chr, eith way we break;
break;
}
size_t clen = strlen(it->first);
fwrite(&clen,sizeof(size_t),1,fstfp);
fwrite(it->first,1,clen,fstfp);
size_t nit=posi.size();
assert(1==fwrite(&nit,sizeof(size_t),1,fstfp));
int64_t tell = bgzf_tell(fstbg);
fwrite(&tell,sizeof(int64_t),1,fstfp);
bgzf_write(fstbg,&posi[0],posi.size()*sizeof(int));
int inc =0;
for(int i=0;i<saf.size();i++)
for(int j=i+1;j<saf.size();j++){
bgzf_write(fstbg,&(ares[inc][0]),ares[inc].size()*sizeof(double));
bgzf_write(fstbg,&(bres[inc][0]),bres[inc].size()*sizeof(double));
inc++;
}
if(arg->chooseChr!=NULL)
break;
}
delGloc(saf,nSites);
destroy(gls,nSites);
destroy_args(arg);
for(int i=0;i<sfs.size();i++)
delete [] sfs[i];
#if 0
fprintf(stderr,"\n\t-> NB NB output is no longer log probs of the frequency spectrum!\n");
fprintf(stderr,"\t-> Output is now simply the expected values! \n");
fprintf(stderr,"\t-> You can convert to the old format simply with log(norm(x))\n");
#endif
bgzf_close(fstbg);
fclose(fstfp);
fprintf(stderr,"\t-> fst index finished with no errors!\n");
return 0;
}
static PyObject *PyRRD_xport(
PyObject UNUSED(*self),
PyObject * args)
{
PyObject *r;
int argc, xsize;
char **argv, **legend_v;
time_t start, end;
unsigned long step, col_cnt;
rrd_value_t *data, *datai;
if (create_args("xport", args, &argc, &argv) < 0)
return NULL;
if (rrd_xport(argc, argv, &xsize, &start, &end,
&step, &col_cnt, &legend_v, &data) == -1) {
PyErr_SetString(ErrorObject, rrd_get_error());
rrd_clear_error();
r = NULL;
} else {
PyObject *meta_dict, *data_list, *legend_list, *t;
unsigned long i, j;
rrd_value_t dv;
unsigned long row_cnt = ((end - start) / step);
r = PyDict_New();
meta_dict = PyDict_New();
legend_list = PyList_New(col_cnt);
data_list = PyList_New(row_cnt);
PyDict_SetItem(r, PyString_FromString("meta"), meta_dict);
PyDict_SetItem(r, PyString_FromString("data"), data_list);
datai = data;
PyDict_SetItem(meta_dict, PyString_FromString("start"), PyInt_FromLong((long) start));
PyDict_SetItem(meta_dict, PyString_FromString("end"), PyInt_FromLong((long) end));
PyDict_SetItem(meta_dict, PyString_FromString("step"), PyInt_FromLong((long) step));
PyDict_SetItem(meta_dict, PyString_FromString("rows"), PyInt_FromLong((long) row_cnt));
PyDict_SetItem(meta_dict, PyString_FromString("columns"), PyInt_FromLong((long) col_cnt));
PyDict_SetItem(meta_dict, PyString_FromString("legend"), legend_list);
for (i = 0; i < col_cnt; i++) {
PyList_SET_ITEM(legend_list, i, PyString_FromString(legend_v[i]));
}
for (i = 0; i < row_cnt; i++) {
t = PyTuple_New(col_cnt);
PyList_SET_ITEM(data_list, i, t);
for (j = 0; j < col_cnt; j++) {
dv = *(datai++);
if (isnan(dv)) {
PyTuple_SET_ITEM(t, j, Py_None);
Py_INCREF(Py_None);
} else {
PyTuple_SET_ITEM(t, j, PyFloat_FromDouble((double) dv));
}
}
}
for (i = 0; i < col_cnt; i++) {
rrd_freemem(legend_v[i]);
}
rrd_freemem(legend_v);
rrd_freemem(data);
}
destroy_args(&argv);
return r;
}
int main_opt(args *arg){
std::vector<persaf *> &saf =arg->saf;
for(int i=0;i<saf.size();i++)
assert(saf[i]->pos!=NULL&&saf[i]->saf!=NULL);
size_t nSites = arg->nSites;
if(nSites == 0){//if no -nSites is specified
nSites=nsites(saf,arg);
}
if(fsizes<T>(saf,nSites)>getTotalSystemMemory())
fprintf(stderr,"\t-> Looks like you will allocate too much memory, consider starting the program with a lower -nSites argument\n");
fprintf(stderr,"\t-> nSites: %lu\n",nSites);
float bytes_req_megs =(float) fsizes<T>(saf,nSites)/1024/1024;
float mem_avail_megs =(float) getTotalSystemMemory()/1024/1024;//in percentile
//fprintf(stderr,"en:%zu to:%f\n",bytes_req_megs,mem_avail_megs);
fprintf(stderr,"\t-> The choice of -nSites will require atleast: %f megabyte memory, that is at least: %.2f%% of total memory\n",bytes_req_megs,bytes_req_megs*100/mem_avail_megs);
std::vector<Matrix<T> *> gls;
for(int i=0;i<saf.size();i++)
gls.push_back(alloc<T>(nSites,saf[i]->nChr+1));
int ndim=(int) parspace(saf);
double *sfs=new double[ndim];
//temp used for checking pos are in sync
setGloc(saf,nSites);
while(1) {
int ret=readdata(saf,gls,nSites,arg->chooseChr,arg->start,arg->stop,NULL,NULL);//read nsites from data
int b=0;
//fprintf(stderr,"\t\tRET:%d gls->x:%lu\n",ret,gls[0]->x);
if(ret==-2&&gls[0]->x==0)//no more data in files or in chr, eith way we break;
break;
#if 0
if(saf.size()==1){
if(ret!=-2){
if(gls[0]->x!=nSites&&arg->chooseChr==NULL&&ret!=-3){
// fprintf(stderr,"continue continue\n");
continue;
}
}
}else
#endif
{
if(gls[0]->x!=nSites&&arg->chooseChr==NULL&&ret!=-3){
//fprintf(stderr,"continue continue\n");
continue;
}
}
if(gls[0]->x==0)
continue;
fprintf(stderr,"\t-> Will run optimization on nSites: %lu\n",gls[0]->x);
neverusegoto:
if(arg->bootstrap)
fprintf(stderr,"Will do bootstrap replicate %d/%d\n",b+1,arg->bootstrap);
if(arg->sfsfname.size()!=0)
readSFS(arg->sfsfname[0],ndim,sfs);
else{
if(arg->seed==-1){
for(int i=0;i<ndim;i++)
sfs[i] = (i+1)/((double)(ndim));
}else{
for(int i=0;i<ndim;i++){
double r=drand48();
while(r==0.0)
r = drand48();
sfs[i] = r;
}
}
}
normalize(sfs,ndim);
if(bootstrap==NULL &&arg->bootstrap)
bootstrap = new size_t[gls[0]->x];
if(bootstrap){
for(size_t i=0;i<gls[0]->x;i++)
bootstrap[i] = lrand48() % gls[0]->x;
std::sort(bootstrap,bootstrap+gls[0]->x);
}
double lik;
if(arg->emAccl==0)
lik = em<float>(sfs,arg->tole,arg->maxIter,arg->nThreads,ndim,gls);
else
lik = emAccl<float>(sfs,arg->tole,arg->maxIter,arg->nThreads,ndim,gls,arg->emAccl);
fprintf(stderr,"likelihood: %f\n",lik);
fprintf(stderr,"------------\n");
#if 1
// fprintf(stdout,"#### Estimate of the sfs ####\n");
//all gls have the same ->x. That means the same numbe of sites.
for(int x=0;x<ndim;x++)
fprintf(stdout,"%f ",((double)gls[0]->x)*sfs[x]);
fprintf(stdout,"\n");
fflush(stdout);
#endif
if(++b<arg->bootstrap)
goto neverusegoto;
for(int i=0;i<gls.size();i++)
gls[i]->x =0;
if(ret==-2&&arg->chooseChr!=NULL)
break;
if(arg->onlyOnce)
break;
}
delGloc(saf,nSites);
destroy(gls,nSites);
destroy_args(arg);
delete [] sfs;
fprintf(stderr,"\n\t-> NB NB output is no longer log probs of the frequency spectrum!\n");
fprintf(stderr,"\t-> Output is now simply the expected values! \n");
fprintf(stderr,"\t-> You can convert to the old format simply with log(norm(x))\n");
return 0;
}
int printMulti(args *arg){
fprintf(stderr,"[%s]\n",__FUNCTION__);
std::vector<persaf *> &saf =arg->saf;
for(int i=0;i<saf.size();i++)
assert(saf[i]->pos!=NULL&&saf[i]->saf!=NULL);
size_t nSites = arg->nSites;
if(nSites == 0){//if no -nSites is specified
nSites=nsites(saf,arg);
}
std::vector<Matrix<T> *> gls;
for(int i=0;i<saf.size();i++)
gls.push_back(alloc<T>(nSites,saf[i]->nChr+1));
int ndim=(int) parspace(saf);
double *sfs=new double[ndim];
//temp used for checking pos are in sync
setGloc(saf,nSites);
int *posiToPrint = new int[nSites];
//used for printout old format
FILE **oldfp = NULL;
gzFile oldpos = Z_NULL;
if(arg->oldout){
oldfp = new FILE*[saf.size()];
for(int i=0;i<saf.size();i++){
size_t newlen = strlen(saf[i]->fname)+100;
char *tmp =(char*) calloc(newlen,sizeof(char));
tmp = strncpy(tmp,saf[i]->fname,strlen(saf[i]->fname)-4);
fprintf(stderr,"\t-> Generating outputfile: %s\n",tmp);
oldfp[i] = fopen(tmp,"wb");
free(tmp);
}
size_t newlen = strlen(dirname(saf[0]->fname))+100;
char *tmp = (char*) calloc(newlen,sizeof(char));
snprintf(tmp,newlen,"%s/shared.pos.gz",dirname(saf[0]->fname));
fprintf(stderr,"\t-> Generating outputfile: %s\n",tmp);
oldpos = gzopen(tmp,"wb");
free(tmp);
}
while(1) {
static char *curChr=NULL;
int ret=readdata(saf,gls,nSites,arg->chooseChr,arg->start,arg->stop,posiToPrint,&curChr);//read nsites from data
// fprintf(stderr,"ret:%d gls->x:%lu\n",ret,gls[0]->x);
if(arg->oldout==0){
for(int s=0;s<gls[0]->x;s++){
if(arg->chooseChr==NULL)
fprintf(stdout,"%s\t%d",curChr,posiToPrint[s]+1);
else
fprintf(stdout,"%s\t%d",arg->chooseChr,posiToPrint[s]+1);
for(int i=0;i<saf.size();i++)
for(int ii=0;ii<gls[i]->y;ii++)
fprintf(stdout,"\t%f",log(gls[i]->mat[s][ii]));
fprintf(stdout,"\n");
}
}else{
for(int s=0;s<gls[0]->x;s++){
if(arg->chooseChr==NULL)
gzprintf(oldpos,"%s\t%d\n",curChr,posiToPrint[s]+1);
else
gzprintf(oldpos,"%s\t%d\n",arg->chooseChr,posiToPrint[s]+1);
for(int i=0;i<saf.size();i++){
double mytmp[gls[i]->y];
for(int ii=0;ii<gls[i]->y;ii++)
mytmp[ii] = log(gls[i]->mat[s][ii]);
fwrite(mytmp,sizeof(double),gls[i]->y,oldfp[i]);
}
}
}
if(ret==-3&&gls[0]->x==0){//no more data in files or in chr, eith way we break;g
//fprintf(stderr,"breaking\n");
break;
}
for(int i=0;i<gls.size();i++)
gls[i]->x =0;
#if 0
if(gls[0]->x!=nSites&&arg->chooseChr==NULL&&ret!=-3){
fprintf(stderr,"continue continue\n");
continue;
}
#endif
if(ret==-2&&arg->chooseChr!=NULL)
break;
if(arg->onlyOnce)
break;
}
delGloc(saf,nSites);
destroy(gls,nSites);
delete [] sfs;
delete [] posiToPrint;
if(arg->oldout==1){
for(int i=0;i<saf.size();i++)
fclose(oldfp[i]);
delete [] oldfp;
gzclose(oldpos);
}
destroy_args(arg);
fprintf(stderr,"\t-> Run completed\n");
return 0;
}
static PyObject *PyRRD_fetch(
PyObject UNUSED(*self),
PyObject * args)
{
PyObject *r;
rrd_value_t *data, *datai;
unsigned long step, ds_cnt;
time_t start, end;
int argc;
char **argv, **ds_namv;
if (create_args("fetch", args, &argc, &argv) < 0)
return NULL;
if (rrd_fetch(argc, argv, &start, &end, &step,
&ds_cnt, &ds_namv, &data) == -1) {
PyErr_SetString(ErrorObject, rrd_get_error());
rrd_clear_error();
r = NULL;
} else {
/* Return :
((start, end, step), (name1, name2, ...), [(data1, data2, ..), ...]) */
PyObject *range_tup, *dsnam_tup, *data_list, *t;
unsigned long i, j, row;
rrd_value_t dv;
row = (end - start) / step;
r = PyTuple_New(3);
range_tup = PyTuple_New(3);
dsnam_tup = PyTuple_New(ds_cnt);
data_list = PyList_New(row);
PyTuple_SET_ITEM(r, 0, range_tup);
PyTuple_SET_ITEM(r, 1, dsnam_tup);
PyTuple_SET_ITEM(r, 2, data_list);
datai = data;
PyTuple_SET_ITEM(range_tup, 0, PyInt_FromLong((long) start));
PyTuple_SET_ITEM(range_tup, 1, PyInt_FromLong((long) end));
PyTuple_SET_ITEM(range_tup, 2, PyInt_FromLong((long) step));
for (i = 0; i < ds_cnt; i++)
PyTuple_SET_ITEM(dsnam_tup, i, PyString_FromString(ds_namv[i]));
for (i = 0; i < row; i++) {
t = PyTuple_New(ds_cnt);
PyList_SET_ITEM(data_list, i, t);
for (j = 0; j < ds_cnt; j++) {
dv = *(datai++);
if (isnan(dv)) {
PyTuple_SET_ITEM(t, j, Py_None);
Py_INCREF(Py_None);
} else {
PyTuple_SET_ITEM(t, j, PyFloat_FromDouble((double) dv));
}
}
}
for (i = 0; i < ds_cnt; i++)
rrd_freemem(ds_namv[i]);
rrd_freemem(ds_namv); /* rrdtool don't use PyMem_Malloc :) */
rrd_freemem(data);
}
destroy_args(&argv);
return r;
}
int fst_stat2(int argc,char **argv){
int pS,pE;//physical start,physical end
int begI,endI;//position in array for pS, pE;
char *bname = *argv;
fprintf(stderr,"\t-> Assuming idxname:%s\n",bname);
perfst *pf = perfst_init(bname);
args *pars = getArgs(--argc,++argv);
fprintf(stderr,"win:%d step:%d\n",pars->win,pars->step);
int *ppos = NULL;
int chs = choose(pf->names.size(),2);
// fprintf(stderr,"choose:%d \n",chs);
double **ares = new double*[chs];
double **bres = new double*[chs];
double unweight[chs];
double wa[chs];
double wb[chs];
size_t nObs =0;
for(myFstMap::iterator it=pf->mm.begin();it!=pf->mm.end();++it){
if(pars->chooseChr!=NULL){
it = pf->mm.find(pars->chooseChr);
if(it==pf->mm.end()){
fprintf(stderr,"Problem finding chr: %s\n",pars->chooseChr);
break;
}
}
fprintf(stderr,"nSites:%lu\n",it->second.nSites);
if(it->second.nSites==0&&pars->chooseChr!=NULL)
break;
else if(it->second.nSites==0&&pars->chooseChr==NULL)
continue;
bgzf_seek(pf->fp,it->second.off,SEEK_SET);
ppos = new int[it->second.nSites];
bgzf_read(pf->fp,ppos,sizeof(int)*it->second.nSites);
for(int i=0;i<it->second.nSites;i++)
ppos[i]++;
for(int i=0;i<choose(pf->names.size(),2);i++){
ares[i] = new double[it->second.nSites];
bres[i] = new double[it->second.nSites];
bgzf_read(pf->fp,ares[i],sizeof(double)*it->second.nSites);
bgzf_read(pf->fp,bres[i],sizeof(double)*it->second.nSites);
}
if(pars->type==0)
pS = ((pars->start!=-1?pars->start:ppos[0])/pars->step)*pars->step +pars->step;
else if(pars->type==1)
pS = (pars->start!=-1?pars->start:ppos[0]);
else if(pars->type==2)
pS = 1;
pE = pS+pars->win;
begI=endI=0;
// fprintf(stderr,"ps:%d\n",pS);exit(0);
if(pE>(pars->stop!=-1?pars->stop:ppos[it->second.nSites-1])){
fprintf(stderr,"end of dataset is before end of window: end of window:%d last position in chr:%d\n",pE,ppos[it->second.nSites-1]);
// return str;
}
while(ppos[begI]<pS) begI++;
endI=begI;
while(ppos[endI]<pE) endI++;
//fprintf(stderr,"begI:%d endI:%d\n",begI,endI);
while(1){
for(int i=0;i<chs;i++)
unweight[i] = wa[i] = wb[i] =0.0;
nObs=0;
fprintf(stdout,"(%d,%d)(%d,%d)(%d,%d)\t%s\t%d",begI,endI-1,ppos[begI],ppos[endI-1],pS,pE,it->first,pS+(pE-pS)/2);
for(int s=begI;s<endI;s++){
#if 0
fprintf(stdout,"%s\t%d",it->first,ppos[s]+1);
for(int i=0;i<choose(pf->names.size(),2);i++)
fprintf(stdout,"\t%f\t%f",ares[i][s],bres[i][s]);
fprintf(stdout,"\n");
#endif
for(int i=0;i<choose(pf->names.size(),2);i++){
unweight[i] += ares[i][s]/bres[i][s];
wa[i] += ares[i][s];
wb[i] += bres[i][s];
}
nObs++;
}
double fstW[chs];
for(int i=0;nObs>0&&i<chs;i++){
fstW[i] = wa[i]/wb[i];
fprintf(stdout,"\t%f\t%f",unweight[i]/(1.0*nObs),fstW[i]);
}
if(chs==3){
//if chr==3 then we have 3pops and we will also calculate pbs statistics
calcpbs(fstW);//<- NOTE: the pbs values will replace the fstW values
for(int i=0;i<3;i++)
fprintf(stdout,"\t%f",fstW[i]);
}
fprintf(stdout,"\n");
pS += pars->step;
pE =pS+pars->win;
if(pE>(pars->stop!=-1?pars->stop:ppos[it->second.nSites-1]))
break;
while(ppos[begI]<pS) begI++;
while(ppos[endI]<pE) endI++;
}
for(int i=0;i<choose(pf->names.size(),2);i++){
delete [] ares[i];
delete [] bres[i];
}
delete [] ppos;
if(pars->chooseChr!=NULL)
break;
}
delete [] ares;
delete [] bres;
destroy_args(pars);
perfst_destroy(pf);
return 0;
}
int fst_print(int argc,char **argv){
char *bname = *argv;
fprintf(stderr,"\t-> Assuming idxname:%s\n",bname);
perfst *pf = perfst_init(bname);
writefst_header(stderr,pf);
args *pars = getArgs(--argc,++argv);
int *ppos = NULL;
fprintf(stderr,"choose:%d \n",choose(pf->names.size(),2));
double **ares = new double*[choose(pf->names.size(),2)];
double **bres = new double*[choose(pf->names.size(),2)];
for(myFstMap::iterator it=pf->mm.begin();it!=pf->mm.end();++it){
if(pars->chooseChr!=NULL){
it = pf->mm.find(pars->chooseChr);
if(it==pf->mm.end()){
fprintf(stderr,"Problem finding chr: %s\n",pars->chooseChr);
break;
}
}
if(it->second.nSites==0)
continue;
bgzf_seek(pf->fp,it->second.off,SEEK_SET);
ppos = new int[it->second.nSites];
bgzf_read(pf->fp,ppos,sizeof(int)*it->second.nSites);
for(int i=0;i<choose(pf->names.size(),2);i++){
ares[i] = new double[it->second.nSites];
bres[i] = new double[it->second.nSites];
bgzf_read(pf->fp,ares[i],sizeof(double)*it->second.nSites);
bgzf_read(pf->fp,bres[i],sizeof(double)*it->second.nSites);
}
int first=0;
if(pars->start!=-1)
while(ppos[first]<pars->start)
first++;
int last=it->second.nSites;
if(pars->stop!=-1&&pars->stop<=ppos[last-1]){
last=first;
while(ppos[last]<pars->stop)
last++;
}
fprintf(stderr,"pars->stop:%d ppos:%d first:%d last:%d\n",pars->stop,ppos[last-1],first,last);
for(int s=first;s<last;s++){
fprintf(stdout,"%s\t%d",it->first,ppos[s]+1);
for(int i=0;i<choose(pf->names.size(),2);i++)
fprintf(stdout,"\t%f\t%f",ares[i][s],bres[i][s]);
fprintf(stdout,"\n");
}
for(int i=0;i<choose(pf->names.size(),2);i++){
delete [] ares[i];
delete [] bres[i];
}
delete [] ppos;
if(pars->chooseChr!=NULL)
break;
}
delete [] ares;
delete [] bres;
destroy_args(pars);
perfst_destroy(pf);
return 0;
}
int fst_stat(int argc,char **argv){
char *bname = *argv;
fprintf(stderr,"\t-> Assuming idxname:%s\n",bname);
perfst *pf = perfst_init(bname);
args *pars = getArgs(--argc,++argv);
int *ppos = NULL;
int chs = choose(pf->names.size(),2);
// fprintf(stderr,"choose:%d \n",chs);
double **ares = new double*[chs];
double **bres = new double*[chs];
double unweight[chs];
double wa[chs];
double wb[chs];
size_t nObs[chs];
for(int i=0;i<chs;i++){
unweight[i] = wa[i] = wb[i] =0.0;
nObs[i] = 0;
}
for(myFstMap::iterator it=pf->mm.begin();it!=pf->mm.end();++it){
if(pars->chooseChr!=NULL){
it = pf->mm.find(pars->chooseChr);
if(it==pf->mm.end()){
fprintf(stderr,"Problem finding chr: %s\n",pars->chooseChr);
break;
}
}
if(it->second.nSites==0)
continue;
bgzf_seek(pf->fp,it->second.off,SEEK_SET);
ppos = new int[it->second.nSites];
bgzf_read(pf->fp,ppos,sizeof(int)*it->second.nSites);
for(int i=0;i<choose(pf->names.size(),2);i++){
ares[i] = new double[it->second.nSites];
bres[i] = new double[it->second.nSites];
bgzf_read(pf->fp,ares[i],sizeof(double)*it->second.nSites);
bgzf_read(pf->fp,bres[i],sizeof(double)*it->second.nSites);
}
int first=0;
if(pars->start!=-1)
while(ppos[first]<pars->start)
first++;
int last=it->second.nSites;
if(pars->stop!=-1&&pars->stop<=ppos[last-1]){
last=first;
while(ppos[last]<pars->stop)
last++;
}
// fprintf(stderr,"pars->stop:%d ppos:%d first:%d last:%d\n",pars->stop,ppos[last-1],first,last);
for(int s=first;s<last;s++){
#if 0
fprintf(stdout,"%s\t%d",it->first,ppos[s]+1);
for(int i=0;i<choose(pf->names.size(),2);i++)
fprintf(stdout,"\t%f\t%f",ares[i][s],bres[i][s]);
fprintf(stdout,"\n");
#endif
for(int i=0;i<choose(pf->names.size(),2);i++){
if(bres[i][s]!=0){
unweight[i] += ares[i][s]/bres[i][s];
nObs[i]++;
}
wa[i] += ares[i][s];
wb[i] += bres[i][s];
}
}
for(int i=0;i<choose(pf->names.size(),2);i++){
delete [] ares[i];
delete [] bres[i];
}
delete [] ppos;
if(pars->chooseChr!=NULL)
break;
}
double fstUW[chs];
double fstW[chs];
for(int i=0;i<chs;i++){
fstUW[i] = unweight[i]/(1.0*nObs[i]);
fstW[i] = wa[i]/wb[i];
fprintf(stderr,"\t-> FST.Unweight[nObs:%lu]:%f Fst.Weight:%f\n",nObs[i],fstUW[i],fstW[i]);
fprintf(stdout,"%f %f\n",fstUW[i],fstW[i]);
}
if(chs==3){
//if chr==3 then we have 3pops and we will also calculate pbs statistics
calcpbs(fstW);//<- NOTE: the pbs values will replace the fstW values
for(int i=0;i<3;i++)
fprintf(stderr,"\t-> pbs.pop%d\t%f\n",i+1,fstW[i]);
}
delete [] ares;
delete [] bres;
destroy_args(pars);
perfst_destroy(pf);
return 0;
}
