Pointlike Parameters

Return to: Part I - Run Pointlike Interactively
On this Page:
DataSpecification
SpectralAnalysis
(Binning Parameters)
ROI
.i_flux
.fit
.TS
.logLikelihood
.localize
plot_sed
.upper_limit
.fit_extension
.tsmap
ROIDisplay
kwargs accepted by id
 

DataSpecification

Note: To view the DataSpecification/SpectralAnalysis code, see: Viewof/pointlike/python/uw/like/pointspec.py.
Name Description
ft1files String pointing directly to an FT1 file or to a list of FT1 files.
ltcube Livetime cube calculated by gtltcube.
binfile

A fits file with the FT1 file binned in healpix. Binfiles are always allsky files, so your binned data is efficiently saved for further use.
Notes:

  • If you have applied gtmktime to the FT1 file, changing its GTI, then these are the GTI used by pointlike when it bins the data; if it is not passed, the data will be automatically binned.
  • If a filename is given but does not exist, after the data is binned it will be saved to the specified file for later.
 
 
 
 
 
 
 
 
Back to Top

SpectralAnalysis (Binning Parameters)

It is strongly recommended that you use an energy binning commensurate with CALDB binning. This can be achieved by making sure 100 MeV appears at the bin edges (an easy way is to set emin=100) and that binsperdec is a multiple of 4. You can easily choose a subset of these bins in the likelihood fitting if you want to narrow the energy bounds.

You can also specify the instrument response function to use during the analysis, and the minimum and maximum ROI (the ROI will scale with the PSF between the two values).

Note:To view the DataSpecification/SpectralAnalysis code, see:
View of/pointlike/python/uw/like/pointspec.py.

Name Default value Description
binsperdec   Number of energy bins per decade of energy.
emin   Minimum energy (in MeV); left edge of the lowest energy bin. Only applied if binned photon data is being created for first time.
 
emax   Maximum energy (in MeV); right edge of the highest energy bin. Only applied on initial binning.
   
irf   Instrument response function.
roi_dir None Center of the region where livetime and exposure is calculated. If this is not set, an allsky exposure/ livetime will be generated.
  roi_dir=180 creates a full sky analysis.
 
exp_radius 20 degree exposure for if roi_dir is set;
allsky if roi_dir is not set.		 Radius (deg) to use if calculate exposure or a ROI.
   
   
   
minROI/maxROI   Minimum and maximum region of interest (ROI) for the analysis (in degrees). Generally, ROI radius is energy-dependent: minROI < r95(e,conversion_type) < maxROI.
   
   
   

That is, the radius is given by the 95% PSF containment
for the band, subject to the constraint that it be
>= minROI and <= maxROI.

Set minROI=maxROI to have a uniform ROI over all energy.
   
   
conv_type -1 Select conversion type (0-front; 1-back; -1 = front+back)
zenithcut 105 Maximum spacecraft pointing angle to allow with respect to zenith.
thetacut 66.4 Cut on photon incidence angle.
event_class 3 Select class level (3-diffuse; 2-source; 1-transient; 0-Monte Carlo)
tstart 0 (no cut on time) Cut on MET > tstart
tstop 0 (no cut on time) Cut on MET < tstop
 
Back to Top

ROI

Note: To view the ROI Object code, see: pointlike/python/uw/like/roi_analysis.py
Name Description
roi_dir The center of the ROI must be specified using as a skymaps.SkyDir object.
xmlfile

Background models and sources can be specified in a source xml file. The format of the xml file is the same as for gtlike, although some of the more complicated spectral models may not yet be implemented.

Note: If you find any inconsistencies in the xml model, please report this to a developer.
 
 
phase_factor If the data has had a phase cut applied to it (for pulsars), this is the fraction of events passing the cut.
 
 

In addition to specifying gtlike style sources, you can also read in a catalog of background sources and add all the background sources to the ROI.

The location of the catalog can be specified with the catalogs flag. Any catalog sources closer than 0.1 degrees of sources already defined in the xml file will not be included, and any catalog source with a similar name to a source specified in xml will not be included.

So it is easiest to define your diffuse model and particular sources of interest in the xml, but get the rest of the catalog sources from the catalog file.

Even if you read a bunch of background sources from the catalog, when the xml file is written out, all of the background sources will be saved into the resulting xml file.

When a catalog is read in with the catalogs flag, by default, only sources within 2 degrees will be fit. Sources farther away will be fixed to their catalog values.

As shown below, you can change this default radius by passing into roi_from_xml instead a FermiCatalog object: 

from uw.like.pointspec_helpers import FermiCatalog    
catalogs = [ FermiCatalog("gll_psc_v03.fit",free_radius=10) ]

Note: Sources from the catalogs argument are added to the model up to a hardcoded radius maxROI+5 from the roi center. In recent versions of pointlike, the XML file is also cut off at the same radius.
catalogs
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fit_emin/fit_emax The minimum and maximum energies to use in spectral fitting. This is generally useful if you want to use the same energy binned data (or keep the energy binning consistent with the CALDB file), but then want to change which of the energy bins are used for a spectral analysis.
 
 
 
 
Back to Top

.i_flux

Name Default value Description
emin 100 Lower bound in MeV.
emax inf Upper bound in MeV.
e_weight 0 Energy power by which to scale dN/dE.
cgs False If True, energy in ergs, else in MeV.
error False If True, return value is a tuple with flux and estimated error.
   
two_sided False If True, return value is a triple with flux, estimated high and low errors.
   
 
Back to Top

.fit

Name Default value Description
method simplex Choose between 'simplex', 'powell', and 'minuit'.
tolerance 0.01 Approximate absolute tolerance of log likelihood value.
save_values True Store fit values in ROI after fit.
do_background True Calculate errors on the diffuse source's spatial parameters.
fit_bg_first False Before doing the global fit, freeze all point sources and first fit the diffuse sources.
estimate_errors True Estimate th05/24/2011 error_for_steps False Use previously estimated errors for the fitting algorithm's initial step size.
   
use_gradient False

Use an analytic gradient of the spatial parameters when doing a spectral fit. Only supported for some spectral models.

Note: The analytic gradient is only used for spectral fitting (not localizing or extension fitting).
   
   
   
   
gtol 1e-1 When using the analytic gradient, Gradient norm must be less than gtol before successful termination.
   
 
Back to Top

.TS

Name Default value Description
quick   The TS function takes the parameter quick which defaults to True.
   
  If quick=True, the likelihood is calculated with source flux set to 0.
  If quick=False, the code does a full refit of all other free sources for the null hypothesis.
   
    Note: quick=False is the same as setting reoptimize=True in gtlike.
   
which   Name of the source.
method   Fitting algorithm method.
bandfits   Calculate the TS of your source without assuming a spectral model by fitting the flux in each energy bin independently.
   
   
 
Back to Top

.logLikelihood

Name Description
ll

Note: You can get the total loglikelihood using:

ll=-1*roi.logLikelihood(roi.parameters())

This function returns the negative log likelihood (suitable for minimization) so must be multiplied by -1.
 
 
 
 
 
 
Back to Top

.localize

Name Default value Description
which central source Name of the source to localize. For sources not at the ROI center, make sure the ROI is large enough.
   
tolerance 1e-3 Maximum difference in degrees between two successive best fit positions.
   
update False If True, update localization internally after fit, i.e., recalculate point source contribution.
   
verbose False Print out more information during the fit.
bandfits False If True, use a band-by-band (model independent) spectral fit; otherwise, use broadband fit with the spectral model of the source.
   
   
seedpos False If set, use this position instead of the source position.
 
Back to Top

plot_sed

Name Default value Description
roi   ROIAnalysis object.
which 0 Index of source to plot.
axis None 1e2, 1e5, 1e-8, 1e-2 depending on use_ergs
butterfly True Plot model with a butterfly outline.
use_ergs True Convert to ergs in the flux units and use GeV on the x-axis.
   
outdir None If set, save sed into <outdir>/<source_name>_sed.png if outdir is a directory; save into filename=<outdir> if not.
   
   
galmap True Plot position on galactic map if set.
printout* False If True, print the sed points to stdout.
 
Back to Top

.upper_limit

Name Default value Description
which   Index of the point source for which to compute the limit.
   
confidence .95 Desired confidence level of the upper limit.
integral_min -15 Lower limit of the likelihood integral in log space.
integral_max -8 Upper limit of the likelihood integral in log space.
simps_points 10 Number of evaluation points per decade for Simpson's rule.
 
Back to Top

.fit_extension

Name Default value Description
tolerance 0.05 This is passed into Minuit while fitting the extension parameters.
   
bandfits False Whether to use a spectral independent model when fitting extension.
   
use_gradient False If bandfits is False, whether to use the analytic gradient when fitting spectral parameters (during an iteration for a given extension).
   
   
 
Back to Top

.tsmap

Name Default value Description
size 2 Size, in degrees, of the image.
pixelsize 0.1 Size, in degrees, of pixels
proj 'ZEA' Projection type.
galactic False Use galactic or equatorial coords.
center   coords of image center.
Default will be the ROI Center.
   
photon_index 2 Assumed spectral index of the new source used in the tsmap.
   
spectral_model None Instance of spectral model (from Models.py) to use for TS calc. This will override photon_index.
   
 
Back to Top

ROIDisplay

Name Default value Description
figsize ((12,8)) Physics Size of the image.
fignum 3 matplotlib figure number.
pixelsize 0.25 Size of each image pixel.
size 10 Size of the field of view.
nticks 5 Number of axes tick marks.
log_counts_min 1.0 Counts scale minimum.
log_counts_max 2.5 Counts scale maximum.
label_sources False Label sources during plot.
galactic True Coordinate system for plot.
 
Back to Top

kwargs accepted by id

Name Description
cpt_class A counterpart class (or list of classes) to use for associations; should have a python module named for it in the classes directory, e.g., 'crates' will use info from the crates.py module). If None, will use all available.
 
 
name Name of the LAT source. If provided, the association will be saved in the dictionary sa.sources, with name as the key (where sa is the instance of SourceAssociation). Otherwise, the return dict will not be saved.
 
 
trap_mask If true, uses a slightly different selection of sources for computing counterpart densities.
 
  Note: This is here only for compatibility with gtsrcid.
unique If true, compute probabilities for unique associations (one counterpart per LAT source).
 

Owned by:Joshua Lande

 
Last updated by: Chuck Patterson 05/24/2011
Back to Top