Page Menu
Home
c4science
Search
Configure Global Search
Log In
Files
F96785974
o_set_lens_bayes.c
No One
Temporary
Actions
Download File
Edit File
Delete File
View Transforms
Subscribe
Mute Notifications
Award Token
Subscribers
None
File Metadata
Details
File Info
Storage
Attached
Created
Mon, Dec 30, 20:28
Size
10 KB
Mime Type
text/x-c
Expires
Wed, Jan 1, 20:28 (2 d)
Engine
blob
Format
Raw Data
Handle
23226256
Attached To
R1448 Lenstool-HPC
o_set_lens_bayes.c
View Options
#include<stdlib.h>
#include<math.h>
#include<fonction.h>
#include<constant.h>
#include<dimension.h>
#include<structure.h>
/****************************************************************/
/* nom: o_set_lens_bayes */
/* auteur: Eric Jullo */
/* date: 03/05/06 */
/* place: ESO, Chile *
*
* Set the lens parameters that give the most probable model.
*
* Refer to setBayesModel() for the list of method possibities
*
* For each parameter of the <bayes.dat> assign the median value
*/
void o_set_lens_bayes(int method, int prior, double limit)
{
double **array; // contains the bayes.dat data
int nParam;
long int nVal; //size of array
// Read the bayes.dat file
array = readBayesModels(&nParam, &nVal);
if ( nVal == 0 )
{
#ifdef DEBUG
fprintf(stderr, "ERROR: bayes.dbg.dat file not found.\n");
#else
fprintf(stderr, "ERROR: bayes.dat file not found.\n");
#endif
exit(-1);
}
if ( method >= nVal )
{
fprintf(stderr, "ERROR: %d larger than maximum line index %ld\n", method, nVal - 1);
exit(-1);
}
// set the model
setBayesModel(method, nVal, array);
// Set the lmin and lmax limits to Gaussian 3 sigma error
o_set_limit_bayes(nParam, nVal, array, prior, limit);
free(array);
}
// Set the lmin and lmax values to 1sigma error around the median value
void o_set_limit_bayes(int nParam, long int nVal, double **array,
int prior, double limit)
{
extern struct g_mode M;
extern struct g_grille G;
extern struct g_pot P[NPOTFILE];
extern struct g_cosmo C;
extern struct g_image I;
extern struct pot lmin[];
extern struct pot lmax[];
extern struct z_lim zlim[];
extern struct z_lim zalim;
extern int block[][NPAMAX];
extern int cblock[NPAMAX];
extern int vfblock[NPAMAX];
extern struct sigposStr sigposAs;
int iParam, ipx;
long int i;
double med, std, err, min, max;
// Set the position of the first physical parameter in bayes.dat
iParam = IDPARAM1;
// Set the optimized clumps limits
for ( i = 0; i < G.no_lens; i++ )
{
for ( ipx = CX; ipx <= PMASS; ipx++ )
if ( block[i][ipx] != 0 )
{
block[i][ipx] = prior;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / med;
if ( prior == 3 )
{
min = med;
max = std;
}
else
{
min = med * ( 1 - err );
max = med * ( 1 + err );
}
o_set_lmin( i, ipx, min );
o_set_lmax( i, ipx, max );
if ( ipx == B0 )
{
lmin[i].sigma = min;
lmax[i].sigma = max;
}
}
}
// for the grid clumps
for ( i = G.nmsgrid; i < G.nlens; i++ )
{
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
//std = std > fabs(med) ? fabs(med) : std;
err = std / med;
// uniform prior only
min = med * ( 1 - err );
max = med * ( 1 + err );
if( block[i][B0] != 0 )
{
lmin[i].sigma = min;
lmax[i].sigma = max;
}
else
{
lmin[i].pmass = min;
lmax[i].pmass = max;
}
}
// Set the source parameters
if( M.iclean == 2 )
{
extern int sblock[NFMAX][NPAMAX];
extern struct g_source S;
for ( i = 0; i < S.ns; i++ )
for ( ipx = SCX; ipx <= SFLUX; ipx++ )
if ( sblock[i][ipx] != 0 )
{
sblock[i][ipx] = prior;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / med;
if ( prior == 3 )
{
min = med;
max = std;
}
else
{
min = med * ( 1 - err );
max = med * ( 1 + err );
}
o_set_lmin(i, ipx, min);
o_set_lmax(i, ipx, max);
}
}
// Set the cosmo
for ( ipx = OMEGAM; ipx <= WA; ipx++ )
if ( cblock[ipx] != 0 )
{
cblock[ipx] = 1;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / med;
o_set_lmin(0, ipx, med * ( 1 - err ));
o_set_lmax(0, ipx, med * ( 1 + err ));
}
// rescale the z_m_limit redshifts
for ( i = 0; i < I.nzlim; i++ )
if ( zlim[i].opt != 0 )
{
zlim[i].bk = prior;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / med;
if ( prior == 3 )
{
zlim[i].min = med;
zlim[i].max = std;
}
else
{
zlim[i].min = med * ( 1 - err );
zlim[i].max = med * ( 1 + err );
}
}
// rescale the z_a_limit redshift
if( zalim.bk != 0 )
{
zalim.bk = prior;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / med;
if ( prior == 3 )
{
zalim.min = med;
zalim.max = std;
}
else
{
zalim.min = med * ( 1 - err );
zalim.max = med * ( 1 + err );
}
}
// Set the velocity field
for ( ipx = VFCX; ipx <= VFSIGMA; ipx++ )
if ( vfblock[ipx] != 0 )
{
vfblock[ipx] = 1;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / med;
o_set_lmin(0, ipx, med * ( 1 - err ));
o_set_lmax(0, ipx, med * ( 1 + err ));
}
// Set the potfile limits
for( i = 0; i < G.npot; i++ )
if ( P[i].ftype != 0 )
{
if ( P[i].ircut != 0 )
{
P[i].ircut = prior;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / med;
if ( prior == 3 )
{
min = med;
max = std;
}
else
{
min = med * ( 1 - err );
if( min < 0 ) min = 0.;
max = med * ( 1 + err );
}
P[i].cut1 = min;
P[i].cut2 = max;
P[i].cutkpc1 = P[i].cut1 * (d0 / C.h * distcosmo1(P[i].zlens));
P[i].cutkpc2 = P[i].cut2 * (d0 / C.h * distcosmo1(P[i].zlens));
}
if ( P[i].isigma != 0 )
{
P[i].isigma = prior;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / med;
if ( prior == 3 )
{
min = med;
max = std;
}
else
{
min = med * ( 1 - err );
if( min < 0 ) min = 0.;
max = med * ( 1 + err );
}
P[i].sigma1 = min;
P[i].sigma2 = max;
}
if ( P[i].islope != 0 )
{
P[i].islope = 1;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / err;
P[i].slope1 = med * ( 1 - err );
if( P[i].slope1 < 0 ) P[i].slope1 = 0.;
P[i].slope2 = med * ( 1 + err );
}
if ( P[i].ivdslope != 0 )
{
P[i].ivdslope = 1;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / err;
P[i].vdslope1 = med * ( 1 - err );
if( P[i].vdslope1 < 0 ) P[i].vdslope1 = 0.;
P[i].vdslope2 = med * ( 1 + err );
}
if ( P[i].ivdscat != 0 )
{
P[i].ivdscat = 1;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / err;
P[i].vdscat1 = med * ( 1 - err );
if( P[i].vdscat1 < 0 ) P[i].vdscat1 = 0.;
P[i].vdscat2 = med * ( 1 + err );
}
if ( P[i].ircutscat != 0 )
{
P[i].ircutscat = 1;
med = median(nVal, array[iParam]);
std = limit * stddev(nVal, array[iParam++]);
std = std > fabs(med) ? fabs(med) : std;
err = std / err;
P[i].rcutscat1 = med * ( 1 - err );
if( P[i].rcutscat1 < 0 ) P[i].rcutscat1 = 0.;
P[i].rcutscat2 = med * ( 1 + err );
}
}
// rescale the noise
if ( sigposAs.bk != 0 )
{
med = median(nVal, array[iParam]);
sigposAs.min = med;
sigposAs.max = stddev(nVal, array[iParam++]);
}
if ( I.dsigell != -1. )
{
med = median(nVal, array[iParam++]);
I.dsigell = med;
}
}
Event Timeline
Log In to Comment