XMM-Newton Science Analysis System


ommergelists (ommergelists-1.2) [xmmsas_20230412_1735-21.0.0]

Testing

The aim of the test is to check that ommergelists

1. Correctly constructs the new output source-list file by
   • Checking that the merged SRCLIST table contains all the columns in the SRCLIST table of the first source-list file plus the columns unique to the SRCLIST table of the second input source-list file.
   • Correctly finds the identical sources in the two input source-list files.
   • The data in each row of each column of the merged SRCLIST data is correct.
   • Correctly sorts the merged table into order of increasing Right Ascension.

A program createtest creates the test files and works as follows-

1. It creates an observation source-list file (test1.fits that contains the columns RA_CORR, DEC_CORR, U_RAW_RATE, U_CORR_RATE and U_EXTENSION_FLAG.

2. A randow number of rows, between 10 and 1000, is added to the table.
3. The table is filled by generating random values for each row of each column.

4. A second observation source-list file (test2.fits is created that contains the columns RA_CORR, DEC_CORR, V_RAW_RATE, V_CORR_RATE and V_EXTENSION_FLAG.

5. A randow number of rows, between 10 and 1000, is added to the table.

6. The table is filled by selecting random rows from the first SRCLIST table and by filling the columns of the unfilled rows unfilled with random values.

7. A third source-list file test3.fits is then formed from test1.fits and all the unique sources in test2.fits with raw count-rates no greater than 5.0 are added to it. This file will therefore contain columns for filters U and V.

8. The SRCLIST tables in the files test1.fits, test2.fits and test3.fits are sorted into increasing RA order.

9. ommergelists is then executed with test1.fits and test2.fits as input source-list files and it produces a merged source-list file test4.fits.

10. A progam checkoutput is then run to check the following
    • That the SRCLIST table in test3.fits contains the same number of rows as the SRCLIST table in test4.fits.
    • That the RA values in the SRCLIST table of test3.fits are in order of increasing RA.

    • That each value in each row of each of the columns RA_CORR, DEC_CORR, U_RAW_RATE and B_RAW_RATE in the SRCLIST table of test3.fits are identical to the corresponding values in test4.fits to within a tolerance (0.0001 arcsecs for RA_CORR and DEC_CORR and 1e-6 for the remainder)-between any bigger difference is reported as an error and the test is deemed to have failed.
    • That the algorithm for computing the angular distance between 2 sources is correct. One thousand sources with random RAs and Decs are produced and a second list of sources is produced from this having small known angular separations. The distances between each source in the first list and its corresponding one in the second list is computed using the algorithm that ommergelists and any deviation greater then 0.001 arcsecs from the known deviation is reported as an error and the test deemed to have failed.

The actual testing is controlled by a shell program in the ommergelists test directory and is initiated by typing the command make test. To do the testing in low memory mode, you should set the environment variable SAS_MEMORY_MODEL equal to low. Note that at the daily builds at Villspa the testing of all SAS software packages is done in both high and low memory modes (referring to the memory model used by the SAS Data Access Layer).

For more demanding development testing, the number of SWS files can be set to any number by editing the program createfiles- tests have been done by setting the number of rows in the SRCLIST table in test1.fits to 1,000,000.