20020816: purdue's pnga2area...

>From:  Benjamin Cotton <address@hidden>
>Organization:  Purdue
>Keywords:  200208161723.g7GHNbK29603 ldm-mcidas pnga2area

Anne and Ben,

>I'm trying to debug Ben Cotton's problems at Purdue. One of them is the
>one I described to you earlier.  Here are the messages that are
>appearing in the [system] log:
>
>Aug 16 07:29:08 anvil /kernel: pid 32205 (pnga2area), uid 61000: exited
>on signal 11 (core dumped)
>
>I ran pnga2area from the command line.  I didn't get a core dump, but
>got the following:
>
>(anvil.eas.purdue.edu) [/project/ldm/data/mcidas]% pnga2area -l - -vx
>temp < 02073000_sie.png
>Aug 16 17:05:25 pnga2area[52901]: Starting Up
>       in GetFileName
>       mask: 8
>       band: 4
>       in GetSatAnnotInfo
>Aug 16 17:05:25 pnga2area[52901]: Error: unable to find SATANNOT: No
>such file or directory
>       in GetSatBandInfo
>Aug 16 17:05:25 pnga2area[52901]: Error: unable to find SATBAND: No such
>file or directory
>       cres: 4km
>       in ReplaceToken
>       in ReplaceToken
>       in ReplaceToken
>Aug 16 17:05:25 pnga2area[52901]: output file pathname: temp
>       datoff: 2816
>       datlen: 2560
>       cmtlen: 240
>Aug 16 17:05:26 pnga2area[52901]: unPNG::   961865   2422256  2.5183
>Aug 16 17:05:26 pnga2area[52901]: Exiting
>(anvil.eas.purdue.edu) [/project/ldm/data/mcidas]% ls -l temp
>-rw-r--r--  1 ldm  ldm  2422256 Aug 16 12:05 temp

OK, I think I see the problem.  pnga2area must be able to find the files
SATANNOT and SATBAND in order to run correctly.  This is discussed in
the ldm-mcidas 'Building and Installing' and 'Using LDM-McIDAS Decoders'
sections of the ldm-mcidas web pages:

ldm-mcidas
http://www.unidata.ucar.edu/packages/mcidas/mcidd/

  Building and Installing
  http://www.unidata.ucar.edu/packages/mcidas/mcidd/ldm-mcidas-build.html
  
  Using LDM-McIDAS Decoders
  http://www.unidata.ucar.edu/packages/mcidas/mcidd/ldm-mcidas-use.html
  
The Building and Installing section shows that one should copy
SATANNOT and SATBAND from the ldm-mcidas 'etc' directory to the
'etc' directory of the user running the LDM, ~ldm/etc.  After the
files are in a fixed location, the decoder needs to be told where
they are.  This is done by  specifying the '-a' and '-b' flags
supported by the decoder.  Your example above should, therefore,
be changed to:

pnga2area -vxl - -a /usr/local/ldm/etc/SATANNOT -b /usr/local/ldm/etc/SATBAND 
temp < 02073000_sie.png

>I assume the input file was an appropriate choice - I chose the input
>file based on the following two pqact.conf entries:
>
>
>MCIDAS  ^pnga2area Q1 UI 
>        PIPE    -close /project/ldm/decoders/pnga2area -l
>        /project/ldm/logs/png.log -vx
>        /project/ldm/data/mcidas/%y%m%d%H_sie.mca  
>MCIDAS  ^pnga2area Q1 UI 
>        FILE    -close data/mcidas/%y%m%d%H_sie.png -l
>        /project/ldm/logs/png.log
>

These pqact.conf entries should also be modified to tell pnga2area where
it can find the files SATANNOT and SATBAND:

MCIDAS   ^pnga2area Q1 UI
        PIPE     -close /project/ldm/decoders/pnga2area 
        -vxl /project/ldm/logs/png.log
        -a etc/SATANNOT
        -b etc/SATBAND
        /project/ldm/data/mcidas/%y%m%d%H_sie.mca

Also, I would change the entries to use the date/time in the product header
instead of the date (%y%m%d) and time (%H) from the system clock:

MCIDAS  ^pnga2area Q. (..) (.*) (.*) (.*) (.*) (........) (....)
        PIPE     -close /project/ldm/decoders/pnga2area 
        -vl /project/ldm/logs/png.log
        -a etc/SATANNOT
        -b etc/SATBAND
        /project/ldm/data/mcidas/\6\7_sie.mca

By the way, this is all presented by way of example in the 
ldm-mcidas-pqact.conf file distributed with the ldm-mcidas package
and referenced from the Using LDM-McIDAS Decoders page of the
ldm-mcidas web site.

>Can you shed any light on this?  (We do not yet have any results in
>/project/ldm/logs/png.log.  I'm trying to get that working...)

I feel very certain that the problems being experienced are a result of
the pnga2area entries in pqact.conf not referencing the SATANNOT and SATBAND
files.

>Anne

Date: Fri, 16 Aug 2002 13:58:02 -0600
From: Anne Wilson <address@hidden>
Subject: Re: 20020816: purdue's pnga2area...

Thanks, Tom!

There was also a problem in his pqact.conf that may have impacted
pnga2area, namely, the final argument, apparently the output file name,
may not have been read because it was on a separate line that did not
start with a tab.  'ldmadmin pqactcheck' complained about it.  It seems
possible that pqact might have started pnga2area without that required
argument, which  might have created the error 11 and core dump (which I
didn't see when running it from the command line). 

Anne