A quick look at a night of acm data
Last Updated: Aug 15, 2019

Often I want to make a simple surevy of the acm images for a particular night.

Method 1: Original (Slow!) method

originally I built a script set that allowed to build two files (BaseDir and Date) that specified the night and where it is located on disk. The script I would next run (acm_table) built a table file (ACM.table) that contained a large variety of header and image quantities. I could quickly summarize these images with acm_table_qc to learn how many bias adn sky images were present. I could view these properties in an interactive plot and select acm images for viewing with point_selector.

On sco2019: 
wheredata Y                    # /media/sco/DataDisk1/sco/AD/HET_work/acm_nights,  20190724
acm_table 10.0 N               # 509 images, take about 4 seconds on sco2019  
% acm_table_qc ACM Y N
  0   0   509   0   (Nbias,Ndark,Nopen,Nnone)  
% point_selector ACM timeH ImNum N

On mcs: 
[sco@mcs ~/tmp]$ wheredata Y                    # /hetdata/data, 20190724
[sco@mcs ~/tmp]$ acm_table 10.0 N               # 536 images, take about 3 minutes on mcs  
[sco@mcs ~/tmp]$ acm_table_qc ACM Y N     
  10   10   512   4   (Nbias,Ndark,Nopen,Nnone)  
[sco@mcs ~/tmp]$ point_selector ACM timeH ImNum N

This routine is easy to run, but is slow to run. This is particularly true on mcs, the very machine where we often want to summarize an entire night of acm images. Nevertheless, I could run this procedure on the two computers above and see in short order that my image set on sco2019 (my home machine) was incomplete: the bias frames are missing. This is a common occurrence since the bias image are usually taken late the next day in RA Ops, and I probably transferred the images in the morning. Hence, the prident next step would be to re-transfer the acm image set.

Method 2: A faster approach with acm_table_markII

I now have the script anmed acm_table_markII that will perform the above job very quickly. A quick summary os useful. Here we build a complete list of the available acm images with the script pas_imlistgen (this makes the file list.IMAGES). Note that we still need the [BaseDir,Date] files. Next, a much faster routine (survey_headers.sh) is used to generate our table file of image data. The survey_headers.sh is built to recognize the case of acm images. When acm images are found, certain useful image header data are written to the file ACMSUP.dat. This file is ised (with the script 1d2table) to generate a table file of acm data named ACMDAT.table. The important thing is that this table file contains the "imgtype" parameter. We can then use the acm_table_qc script to summarize our image types.

On sco2019: 
% wheredata Y                    # /media/sco/DataDisk1/sco/AD/HET_work/acm_nights,  20190724
% acm_table_markII N  
  10   0   455   54   (Nbias,Ndark,Nopen,Nnone)  
acm_table_markII processing time for 519 acm images = 1.000000 (seconds)
Done. Recall that a good way to review the images is: 
   point_selector ACMDAT uthrs im N  

On mcs: 
[sco@mcs ~/tmp]$ wheredata Y                    # /hetdata/data, 20190724
[sco@mcs ~/tmp]$  acm_table_markII N  
  10   10   455   61   (Nbias,Ndark,Nopen,Nnone)  
acm_table_markII processing time for 536 acm images = 14.000000 (seconds)
Done. Recall that a good way to review the images is: 
   point_selector ACMDAT uthrs im N

The important thisng to note here is the speed gain. Here it took us 14 seconds mcs to survey a night. The old methond took 180 seconds to do the same job!

Back to calling page