Man page for  produce-ag


   produce-ag -corrections [-with-slopes]          [-using <tse>] <campaign> [<campaign> ...]
   produce-ag -residuals   [-with-bias-and-slopes] [-using <tse>] <campaign> [<campaign> ...]
   produce-ag -residuals   [-with-slopes]          [-using <tse>] <campaign> [<campaign> ...]
   produce-ag -results                             [-using <tse>] <campaign> [<campaign> ...]
   foreach c ( $campaigns )
     project-times2tse -S -C 'AUTOSLMEAN U=32' -C+ '${FFMT:[+TSF]} ->W ${WOPT} G=${GOPT:[1000.]} ' \
                       -o msets/msets $c
   setenv MSETSADDG 380.
   produce-ag -own-sets msets/ -use-gaplength 2 <campaign> [<campaign> ...]
                  Processes the mc-files from mc-files.lst for the specified

                  campaigns (usually d/AGSG_...) and produces project-wide
                  ascii files (type tsf), alternately AG-set.txt -like files.
 PREMISE:         project-times2tse as in script final-analysis
                  tsf-files with time stamps at central set-times (irregularly spaced)
                  with set-means, -stdev's and number of drops.
 with option         Result

  -results         - A project-wide ascii file and an mc-file with
                     reduced set-means and uncertainties
                     Reduction: (SCG minus drift) plus AG-slopes
                     Output to ${TSFOUT}<mark>-<project>.set.tsf
                     See ./autoset.tse,SFA or ,SA
                     Nice defaults exist.

  -corrections     - A project-wide ascii file with
                     the series composed of (SCG minus drift) + AG-slopes
               -with-slopes :  slopes (no biases)
                     are composed parametrically.
               -with-bias-and-slopes : both parametrically
                     None of these two options:
                     Biases and slopes from the urtap solution.
                     Output as under -results, but autoset.tse labels
                     SPA or PA apply; check the section for the default
                     values that are proposed.
  -residuals       - A project-wide ascii file with
                     the series is subtracted from the AG series.
                     Both slopes/bias-and-slopes options available as under
                     -corrections .
  -own-sets <dir>/ - Project-wide ascii files in gsoft-form, sets from drop data
                     in <dir>/    (.set.txt with headers)
                     in <dir>/d/  (.set.dat without)
                     The latter can be concatenated e.g. according to
                     platforms and instruments, or single campaigns.
                     Study the scatter: per campaign, all campaigns
                     using e.g. ag-set-wmean (
  -results         - implies  -weighted-means -with-slopes
  -own-sets        - implies  -weighted-means -with-slopes
                     Needs a special set of tse-files (see usage), input
                     from <dir>/
 Additional options (may depend on context)
  -weighted-means  - tsfedit will be instructed to compute weighted
                     means. Default is even weighting (except with -results)
  -in-ag-format    - output set-files furnish values in uGal
                     and their names obtain extension .txt
                     Comments (table header text) will be
                     included (in other context you must manually
                        setenv AGCU 32
                     to write the comment to the file output
                        setenv AGCU 6
                     to print it on the protocol.
  -using <tse-file>[,<trg>]
                   - the default is autoset.tse and a target that
                     is internally devised from the context. To
                     override this, specify the option late on the
                     command line. This is very complex and
                     probably fault-prone --- discouraged.
  -use-gaplength #n
                   - The default is 5 samples. For the Nov. 2009 campaign this is

                     too long. For that project, please use 2.
 This script uses tslist with a tsf-edit file autoset.tse.
 In that file the destination directory is set (e.g. tsfo/ )
 tslist is run under tsl, producing log files in ~/tslist-logs/.../dummy.ts.log
 Residuals with slopes are always treated as with-bias-and-slopes
 the difference being that with-slopes is computed parametrically
 (output is marked _sy)
 and with-bias-and-slopes using the time series generated by urtap
 (output is marked empty)
 Corrections without slopes (without bias-and-slopes) are marked _wobs
 Corrections with-slopes are marked _ws
 In the computation of set values, the corrections are evenly weighted,
 the residuals are weighted sums.
 Results are AG - SCG - slopes, need the errors of the weighted means.
 Results exclude bias terms but include the slopes.

 Script evaluate-b+s is more recent, useful for extracting also the uncertainty of the
 slope parameters. It could be made useful also in this script.