QA Review Summary

A lot has changed from v2 to v3 processing; this document provides an overview of process for the QA reviewers.

General Documentation References

The following pages should be useful for general reference.

Version 3 QA Reviewer Files to Edit

There are now two different files that QA reviewers may need to edit during the review process: an expanded YYMMDDH.qagrade file and a new YYMMDDH.qacat file. In either case, after editing the file the reviewer should then execute the "run_scansci" script from the night's datequal directory to update the quality web pages and .sit file.

YYMMDDH.qagrade

This file is the point of change for any grade overrides, as with version 2. It also contains a series of additional flags that may be set as warnings for specific types of anomalies. Negative values are placeholders indicating no override has been entered. A sample file looks like:

\ Reviewer Quality Overrides for 980322s
\ Date created: 08/03/01 19:05:54
|scan  |score|fct1 |fct2 |fct3 |fct4 |fct5 |fct6 |fct7 |S00|S01|S02|S03|S04|S05|S06|S07|S08|S09|S10|S11|W  |
|      |     |phot |sens |seesh|seetr|airgl|     |untrf|hgl|cld|olp|   |   |   |   |   |   |   |   |   |not|
  007    -10  -1.0  -1.0  -1.0  -1.0  -1.0  -1.0  -1.0  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1   0
  008    -10  -1.0  -1.0  -1.0  -1.0  -1.0  -1.0  -1.0  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1   0
  009    -10  -1.0  -1.0  -1.0  -1.0  -1.0  -1.0  -1.0  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1  -1   0

fct1-fct5 are the standard QA factors. fct6 is not used.

fct7 is our eyeball-verified untracked seeing flag. It defaults to 0 whenever the seetracker qafactor goes to 0.1. It's primary utility is to indicate scans with questionable seeing problems that are not bad enough to warrant a quality factor downgrade by the rules but still looks by eye to be questionable (in this case fct4 will still be 1.0, but fct7 will be 0.0).

Several additional flags have been added for certain rare cases that should be set by the QA reviewer when appropriate. By default they are set to 0, but if a reviewer notes a problem they should be reset to 1:

S00 - H electronics glitch (shows up as brightly noisy artifact in coadd)
S01 - Clouds (discovered in v3 that were missed in v2; most likely to be found in earlier Northern data)
S02 - Overlap offsets (mostly identified during v2 overlaps checking of the working databse; these are scans that show a consistent zero-point offset from their neighbors, often taken on different nights, suggesting an improperly determined zero-point offset for a portion of the night)
W - Special note field, indicating the QA review has specific info about this special scan (probably to be set by request of Rae)

Note that "run_scansci" must be run any time any of these fields have been updated. If the grades have changed, then the appropriate sections should be copied into the .QAreview from the new .QAreview.tmpl file. Do not ever change grades by hand in the .QAreview file, since this could lead to inconsistencies between the .qua, .sit, .qagrade, and .QAreview files.

YYMMDDH.qacat

This "Coadd Anomaly Table" is a new file designed to leave a machine-readable record of anomalous problems in coadds that are not otherwise represented in the QA review. During catalog generation this data can be used to set artifact flags at a coadd level. A sample file with a sample entry added by a reviewer looks like:

\ 980322s Coadd Anomaly Table
\ Date created: 08/03/01 19:06:17; Flg codes:
\ a - airplane     g - elect. gltch s - satellite
\ b - bug trail    h - star halo    u - untr. seeing
\ d - diff. spike  m - meteor       x - other
|Date  |H|Scn|Coa|Flg|Comment                                                    |
|  int |c|int|int|c  |c                                                          |
 980322 s 031 009 b   K-bright bug shows up

The standard anomaly codes are summarized in the header. If additional codes are needed they will be added to this document.

Note that this table is by default blank; it contains only header lines. If a coadd shows an anomaly that should be logged, the reviewer should add the appropriate line(s). A separate line should be added for each affected coadd. The special case of coadd number "999" indicates that the entire scan is affected.

Last updated 8/6/01 by R. Hurt