next up previous contents
Next: Aids to the User Up: Overview Previous: Overview


Xselect is a controlling shell for the Ftools. This means that rather than reading all events data into memory, Xselect holds the data as well as its temporary products in disk files, and processes them by piping these files through various ftools. Though this means more I/O, the core size of the program is kept fairly small.

One consequence of this method of operation is that Xselect performs markedly better if the data is stored on a disk that is directly connected to the cpu that is running Xselect, and can perform very badly over slow networks.

If you want to gain more insight into how Xselect uses the Ftools, and what temporary files are produced, use the ECHO command, which writes all the spawned commands back to the terminal before running them.

Xselect keeps track of five directories. The data directory is the place where the event data is stored. The HK files can be stored in a separate directory, the hk directory. There can be a seperate directory for the auxillary filter files (MKF files), the MKF directory. Xselect will not alter any of these files, and can run when the user has only read access to these directories.

Xselect's scratch files are stored in the current working directory. This is also where the product files go, when they are saved. Finally, there is a directory for storing the user's observation catalogues, which initially defaults to the current working directory. The user must have write access to both these directories for Xselect to function. All these directories can be the same directory although we do not recommend that.

The first stage in the analysis is to set up Xselect's internal state for the particular mission-instrument-datamode combination you are going to study, and the directory wherein your data resides. This is done with the various SET commands.

The next step is to enter some data into the program. The observation catalogue (obscat) is the easiest way to do this. If you do not already have an obscat made, MAKE OBSCAT will create one for you. If you do already have one made, use the LOAD OBSCAT command to use it.

Then you use the CHOOSE command to read data from the obscat list into Xselect's internal list.

You can short-cut these steps by reading the event files directly using READ EVENT. In this case Xselect will automatically set the appropriate mission-instrument-datamode combination.

If the data needs some preliminary conversion, such as Faint to Bright conversion or Fast mode time corrections, this should be done next. These commands process the original files, and produce temporary files that Xselect will work on. If you want to return to the original data, you can CLEAR the conversion, and Xselect will use your originally chosen data. Again, Xselect will not alter your original data, and in fact will even work when you have only read access to the data directory.

Next, you may want to clean your data. There are three sets of commands that perform this function in Xselect. The first are the instrument specific cleaning methods, of which SISCLEAN for the ASCA SIS and GISCLEAN for the ASCA GIS instruments are the only examples.

The second set of commands are the FILTER commands. These commands do no work, they merely enter filters (region, column, detector, grade, timing, phase, pha_cutoff and intensity ) that will be applied to the next run of the extract command.

The third set are the SELECT MKF, HK and EVENTS, FAST, and CHIP commands. These commands all work by applying some Boolean expression to one of the data files. These can be either the event, the HK or the MKF files. Note that FILTER MKF/HK and SELECT MKF/HK are synonyms.

In SELECT EVENTS, the events files are directly reduced, and Xselect will continue to use the products of the selection until you give the command CLEAR SELECTION. An example of this is selecting only one of the four chips of the ASCA SIS, which is accomplished with the command

xsel:ASCA- SIS1 > select event "CCDID==1"

In SELECT MKF and SELECT HK, GTI's are produced by applying the selection expression to parameters in the MKF or HK files, which GTI's will be applied on the next run of EXTRACT. More insight into the nature of these parameters can be gained using the MKFBIN/HKBIN and PLOT MKF/HK commands to examine them.

SELECT FAST and CHIP are special cases. The former selects events based on the ASCA SIS FAST mode area discrimination and the latter selects events for the requested Chandra ACIS or XMM EPIC chips and creates the appropriate GTI extension. Note that SELECT runs the ftool fselect which does not update datamodel subspace keywords for the selection performed.

Next, the filters are applied and the product files produced with the EXTRACT command. The possible products are images, spectra, and lightcurves. You can also output the event list that is the result of all the filtering and selections that were entered.

Finally, these product files can be PLOTted, and SAVEd for input into other analysis packages.

After dispensing with a few odd points, we will take up these topics in more detail.

next up previous contents
Next: Aids to the User Up: Overview Previous: Overview
Keith Arnaud

HEASARC Home | Observatories | Archive | Calibration | Software | Tools | Students/Teachers/Public

Last modified: Monday, 19-Mar-2012 14:16:47 EDT