Loving ISIS - Confessions of a Former XSPEC User
Functions in My .isisrc Files:
Below I printout the brief descriptions found at that top of each of
my .isisrc files. The order roughly follows that of the
example analysis discussed on these web pages. Once my .isisrc
files are loaded, most functions will return a brief usage message if
they are called without any arguments. (One of these days I will get
around to writing real help files for these.)
No guarantees that any of these are particularly intelligently programmed. (In fact, I can guarantee that some most assuredly are not.) But again, they get the job done well-enough.
Loading Data (.isisrc_data):
% list_all : List all the data, arfs, and rmfs % load_hexte : Load HEXTE data % load_pca : Load PCA (or any data with a set background file) % load_all : alias of load_pca % read_radio : Read radio data from an ascii file, load to struct. % load_radio : Take the structure above, and load it into data
Grouping Data (.isisrc_group):
% keV_note : Notice positive valued bins based upon energy % cuts. (Positive value restriction ties into % radio data functions later on.) % chan_note : Notice positive valued bins based upon channels % grppha : Rebin data from lowest keV bound to a min counts/bin % grppha_cut : Rebin data from specified keV bound to a min counts/bin % grppha_min : ... and add in a min number of channels per bin % grppha_sn : ... from specified bound, to a min S/N per bin % grppha_sys : ... rebin to a min S/N, while including a specified % (uniform) fractional systematic error % grppha_sn_min : ... from specified bound, to a min S/N per bin, & % minimum number of channels per bin
Errors Bars & Custom Statistics (.isisrc_conf):
% sep_grid : Set the eval method to SEPARATE_GRID % mrg_grid : Set the eval method to MERGED_GRID % usr_grid : Set up logarithmically spaced USER_GRIDs % fstat_mod : Set the fit statistic to sigma=model % fstat_gehr : Set the fit statistic to sigma=gehrels % fstat_norm : Set the fit statistic to sigma=data % lmdif : Set the fit method to lmdif % minim : Set the fit method to minim % subplex : Set the fit method to subplex % marq : Set the fit method to marquardt % conf_map_fail_hook : Switch between minim/lmdif when doing % error contours, and the fit fails % conf_map : Call conf_map_counts with conf_map_fail_hook % add_syst : The new rebinning error hook function, if % systematic errors are set % set_systematic : Set systematic errors on the data % error_loop : Run confidence contours on fit, with refits % when lower chi^2 is found % error_save : Save error bars to a parameter file with best % fit parameters. % bayes_nconf : Bayesian estimate of lower/upper counts bound % given observed counts & significance % bayes_sconf : Bayesian estimate of lower/upper significance bound % given observed counts & desired lower/upper bounds
Custom Fit Functions (.isisrc_fitfun):
% simple_gpile_fit : Gratings Pileup correction function % (New, better version than in "ABC Guide to Pileup") % mgauss_fit : Negative gaussian that won't go < 0 (multiplicative) % qpo_fit : Lorentzian, normalized to QPO rms (counts/bin) % zfc_fit : Zero-frequency centered Lorentzian (counts/bin) % plaw_fit : Counts/bin power-law % bkn_plaw_fit : Counts/bin broken power-law % dbkn_plaw_fit : Counts/bin doubly broken power-law % rms_gauss_fit : Counts/bin Guassian, normalized to RMS. % three_chip_fit : Renorm for specific ACIS chips % four_chip_fit : Renorm for specific ACIS chips % chip() : Combining the above twoNote: A number of the above fit functions were designed for use in timing analysis with ISIS. Normally, ISIS and XSPEC functions calculate a fit function as a bin integral quantity. For timing analysis, I effectively define the average power spectrum in terms of counts per bin, and want to compute models as bin averaged. Hence the models above. (I also find a number of these models, especially the Lorentzian in
qpo_fit();, to be more stable when the
integral over the bin is done analytically, rather than numerically.)
Note also that ISIS has intrinsic fit functions, bin_width
and bin_width_en that can be combined with XSPEC fit
functions to similar effect.
Flux Calculation Functions (.isisrc_flux):
% kev_flux : Model flux density over a given energy span % kev_flux_pc : Integrated model flux, given assumed distance in pc % lambda_flux : Model flux density over a given wavelength span % lambda_flux_pc : Integrated model flux, given assumed distance in pc
Plotting Functions (.isisrc_plots):
% Plot_Unit : My variation on ISIS's plot_unit. Same choices as % in ISIS, but this saves a variable that the below % routines will use. % plot_bsub_counts : Plot background subtracted data as counts/bin, % with choices of three kinds of residuals, or none % at all. % plot_bsub_all : Main plotting routine, to plot background sub- % tracted data in detector counts space. Counts % per second per Plot_Unit, with choices of three % kinds of residuals, or none at all. % plot_bsub_model : An hplot of just the background subtracted model. % oplot_bsub_model : An ohplot of just the background subtracted model. % plot_residuals : A plot of just the data residuals % plot_unfold : A plot of the unfolded spectra, or energy/freq./ % wavelength times unfolded spectra. % pg_color : Make a few nice colors for pgplot % pg_info : Writes out useful info about plotting choices. % msg : A message function used a lot below. % open_print : A version of open_plot that invokes pg_color first % close_print : A version of close_plot that will then display the % hardcopy via a chosen system utility (e.g., gv)
To actually download the .isisrc files, look here.
