WARES 2nd IF code needed

[teuben]

Code is needed on the data acquisition side as well as pipeline side when the 2nd IF is in place.
This will use roach4..7 files.

[teuben]

New experimental code is now in the branch "IF2". It has been tested by using symlinks and faking the 2nd IF to be the first. This seemed to work.

More sanity testing on filename convention OBSNUM_BANK (where bank=0,1) is needed.

[teuben]

See also #18

[teuben]

The current scheme for obsnum_bank is cumbersome if different flagging, mapping is needed. Although it remains useful to have a before and after view of the pipeline results, now this is the only sane way to reduce 2IF data. Example for M17:

 SLpipeline.sh obsnum=108768 dv=50 dw=50 restart=1 
 SLpipeline.sh obsnum=108768 bank=0 pix_list=-0,5,12,13,14,15 
 SLpipeline.sh obsnum=108768 bank=1 pix_list=-0,2,3,4,12,13   

not to mention that the resulting image appears to be suspiciously swapped in the RA direction.

[teuben]

Additional requirement: any of the two restfreq's could be a 0, in which case there will be no data from the corresponding roach (0..3 or 4..7) boards. numbands however will be 2, so the pipeline will need to deal with this setup.