Using a CSC Save File in CIAO

Identify ASCII format

The CXC Datamodel can read and write various ASCII file formats. Many of these can be automatically recognized by first few lines in the file; however, the Tab-Separated-Values (TSV) returned by CSCView needs to be explicitly identified.

To use any CSCView TSV format file users need to specify a kernel option which takes the form:

[opt kernel=text/tsv]

So for example if the output from CSCView was saved as m81.tsv, then users can use this in CIAO using the full Virtual File Syntax would be

"m81.tsv[opt kernel=text/tsv]"

Quotes are likely going to be needed when commands are run on the UNIX command line but should be omitted when a CIAO tool prompts for the value.

Note

The VOTable (XML) format output is not supported by CIAO tools; however, one can use the Python XML parsing routines or other third-party modules to read such files into Python applications and can be imported by DS9. One example is the astropy.io.votable subpackage of Astropy.

Using with CIAO tools

With the virtual file specification discussed above, all the CIAO utilities are now available for use.

Suppose we would like to select all CSC 2.0 master sources from the Orion Nebula Cluster as observed in the ACIS broad band (0.5-7.0 keV). Suppose further that we have used the CSC2.0 WWT application to identify the particular ACIS-based master source stack ID acisfJ0535165m052323_001, which includes all the sources we want to consider. If we search for all master sources within this stack using CSCView, we get a table of 1539 unique sources in this region, which we have saved in the tab-separated-value file orion.tsv. We can now use the various CIAO tools to inspect and analyze these data.

We can start with simply running dmlist to see what columns we selected:

unix% dmlist orion.tsv"[opt kernel=text/tsv]" cols

--------------------------------------------------------------------------------
Columns for Table Block orion.tsv
--------------------------------------------------------------------------------
 
ColNo  Name                 Unit        Type             Range
   1   name                              String[22]                          Source name in the format '2CXO Jhhmmss.s{+|-}ddmmss[X]'
   2   detect_stack_id                   String[24]                          Stack Source identifier
   3   ra                                String[12]                          Source position, ICRS right ascension
   4   dec                               String[11]                          Source position, ICRS declination
   5   instrument                        String[4]                           Instrument used for the stacked observations; 'ACIS' or 'HRC'
   6   region_id                         Int4           -                    Detection region identifier (component number)
   7   match_type                        String[11]                          Type of match
   8   src_cnts_aper_b      count        Real8          -Inf:+Inf            Aperture-corrected detection net counts inferred from the source region aperture
   9   flux_aper_b          erg/s/cm^2   Real4          -Inf:+Inf            Aperture-corrected net energy flux inferred from the source region aperture, bes
  10   flux_significance_b               Real4          -Inf:+Inf            Significance of the stacked-observation detection determined from the ratio of t
  11   ks_intra_prob_b                   Real4          -Inf:+Inf            Intra-observation Kolmogorov-Smirnov test variability probability (highest value
  12   theta_mean           arcmin       Real4          -Inf:+Inf            Mean source region aperture off-axis angle from all stacked observations

unix% dmlist orion.tsv"[opt kernel=text/tsv]" counts
1539

In the first example we see which columns were selected, including the units and comments, as well as the number of rows (i.e. number of sources).

If we wanted to check to see how many sources are unambiguous (i.e. not confused with other sources in the stack) we can simply use the dmfiltering syntax.

unix% dmlist orion.tsv"[opt kernel=text/tsv][match_type=u]" counts
1515

which shows us that 1515 out of 1539 sources are not confused.

If we wanted to do a population study of the sources, i.e. a \(\log{(N)}/\log{(S)}\) analysis, we can use the TSV file with several CIAO tools as shown below:

unix% dmstat orion.tsv"[opt kernel=text/tsv][cols flux_significance_b]"
flux_significance_b
    min:        0             @:        10
    max:        460.8999939           @:        771
   mean:        28.99595175
  sigma:        39.520637864
    sum:        43261.960011
   good:        1492
   null:        47

unix% dmextract orion.tsv"[opt kernel=text/tsv][bin flux_significance_b=1:461:1]" src_sig.fits op=generic clob+
# dmextract (CIAO 4.5): WARNING: Keywords LIVETIME and EXPOSURE not found in file 'orion.tsv'
        LIVETIME set to default 1.0

sherpa In [1]: load_data("src_sig.fits[cols flux_significance_b,counts]")
sherpa In [2]: set_source(exp.ee)
sherpa In [3]: ignore(0,3)
sherpa In [4]: set_method("moncar")
sherpa In [5]: fit()
Dataset               = 1
Method                = moncar
Statistic             = chi2gehrels
Initial fit statistic = 878.783
Final fit statistic   = 100.296 at function evaluation 1906
Data points           = 458
Degrees of freedom    = 455
Probability [Q-value] = 1
Reduced statistic     = 0.220431
Change in statistic   = 778.487
   ee.offset      -0.0334259 
   ee.coeff       -0.0525678 
   ee.ampl        72.8004    
sherpa In [6]: plot_fit()
sherpa In [7]: log_scale(XY_AXIS)
sherpa In [8]: limits(Y_AXIS,0.1,AUTO)
sherpa In [9]: set_plot_xlabel("Broad Band Source Significance")
sherpa In [10]: set_plot_ylabel("Number of Sources")
sherpa In [11]: set_plot_title("CSC ACIS Stacked Results for Orion Nebula Cluster")
sherpa In [12]: limits(X_AXIS,AUTO,120.)

Here we have determined the range of the broad band source significance values using dmstat, used that range to extract a histogram of the number of sources with flux significances between 1 and 461 with dmextract and then input the data into Sherpa to fit the profile with an exponential model, as shown in Figure 1.

Figure 1: The result of plot_fit()

[Version: full-size]

Figure 1: The result of plot_fit()

Obviously this is not a great fit to these data, but the purpose of this example is to illustrate the types of analysis users can perform.

Using with SAOImage DS9

Returning to our CSCView search, we can select one or more of the sources under the Search Results tab to find and download CSC 2.0 data products, and in this case we choose a Stack Events Image including all of these sources, namely acisfJ0535165m052323_001N020_b_img3.fits. We can then import our orion.tsv table file into DS9's catalog tool either on the command line:

unix% ds9 acisfJ0535165m052323_001N020_b_img3.fits -catalog import tsv orion.tsv &

or by going to Analysis→Catalog Tool ..., then in the Catalog Tool window go to File→Import→Tab-Separated-Value→orion.tsv. The result is shown in Figure 2

Figure 2: Using DS9

[Version: full-size]

Figure 2: Using DS9

DS9 displaying the image with sources overlain and a tabular view.

A CIAO-compatible format REGION file may be generated from orion.tsv using some strategy with dmlist and an awk command as follows:

unix% dmlist "cscresults.tsv[opt kernel=text/tsv][cols RA,DEC]" data,clean | tail -1539 | awk '{print "circle("$1":"$2":"$3","$4":"$5":"$6",2.5\")"}' > orion_all.reg

where dmlist has been used to dump the RA and DEC columns in a clean format, tail has been used to eliminate the column header text line, and awk has been employed to put the RA and Dec into CIAO-compatible region file format to specify circles centered on each RA and Dec with 2.50 arcsecond radii. Loading the regions into DS9 produces the image here.

The catalog tool can then be used to select sources, create plots of source properties and participate in SAMP enabled analysis sessions with other SAMP aware applications.