[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20020619: setting up ldmd.conf requests, and scripting NEXRCOMP loads

>From: Gilbert Sebenste <address@hidden>
>Organization: NIU
>Keywords: 200206182234.g5IMYN621495 McIDAS-X 7.8 NEXRCOMP ADDE

Gilbert,

>Thanks! I just did the upgrade on weather.admin, and it worked.

Very good.

>Now, questions. I did uncomment out the 1 KM pix on weather2.admin, so those 
>should be processing now via McIDAS.

(For the tracking system) As I pointed out in our exchanges yesterday, the
process of getting McIDAS to display the composite radar images in the
IDD FNEXRAD feed is:

o ingest of the products by the LDM
o decoding of the products by ldm-mcidas decoders
o creating of an ADDE dataset

>I thought the 1 KM imagery would be 
>sucked in and processed with the ldmd.conf entry I made. I just want the 
>1 KM  stuff--unfortunately, if I also ingest the other images, our network 
>admins will have a hissy fit (not to mention the fact that I won't be 
>using them).

I am not sure why your network admins will have a hissy fit.  The other
composites are much smaller than the 1 km National N0R product.  The
2 km N1P composite is 25% the size of the 1 km N0R, and the 4 km NTP
composite is 6.25% the size of the 1 km N0R.

>So, what pattern should I use in my ldmd.conf to request just 
>those I want, the 1 KM images?

Instead of telling you the pattern, I will tell show you how to figure
out the pattern.  What you should do is use the LDM application 'notifyme'
to list out the headers of the products available from your upstream
feed site.  The following notifyme invocation will show all of the
FNEXRAD products that flood.atmos.uiuc.edu (the machine you are feeding
FNEXRAD from):

notifyme -vxl- -f FNEXRAD -o 3600

You can cut down the listing to just the radar composite products by adding
a search pattern to the notifyme invocation:

notifyme -vxl- -f FNEXRAD -o 3600 -p rad

The listing you will get will show you what the headers for all of the
products available from flood.atmos.uiuc.edu that begin with 'rad'.
Your job then is to see which ones of these you want to ingest and
create the regular expression that will request just those.

For instance, to request just the 1 km N0R national composites, I would
use:

request "^rad/NEXRCOMP/1km/n0r" flood.atmos.uiuc.edu

The point is that you can always use notifyme to see what the headers
of any products look like, and then use that information to create
regular expressions to use in a request line for your ~ldm/ldmd.conf file.

>Next, once I get them going, I want to make images. For instance:
>
>SF 1
>EG 1
>EG I 1
>EU REST IMAGE2
>EU REST IMAGE2
>GU MAKE 2 50 0 0
>IMGDISP RTGINI/GE1KVIS STA=ACT 1 WID=9
>REM IMGDISP RTGINI/GW1KVIS STA=DSM 1 WID=9
>REM IMGDISP G8-GHCC/VIS STA=VTN MAG=1 -2 1 WID=9
>MAP VH 2
>EU REST IMAGE2
>EU REST IMAGE2
>GU MAKE 2 50 0 0
>FRMSAVE 1 CFDG
>OS "mv /home/mcidas/workdata/CFDG.GIF /home/httpd/html/cfdg.gif
>MAP SAT 2 COU=ALL 
>FRMSAVE 1 CFDG2
>OS "mv /home/mcidas/workdata/CFDG2.GIF /home/httpd/html/cfdg2.gif

I get the picture.  I only add the following to tidy up your example a
bit:

SF 1
EG B 1
GU MAKE 2 50 0 0
IMGDISP RTGINI/GE1KVIS EU=IMAGE2 STA=ACT
MAP VH 2
FRMSAVE 1 /home/httpd/html/cfdg.gif
MAP SAT 2 COU=ALL 
FRMSAVE 1 /home/httpd/html/cfdg2.gif

The changes are:

1) you can tell EG to complete erase a frame using the 'B' (both) option

2) there is no need to restore an enhancement more than once.  Also,
   you can just tell IMGDISP to use a particular enhancement

3) you are changing graphic level 2 with the 'GU MAKE 2 50 0 0'
   invocation.  If you are doing this alot, you may consider creating
   your own graphic palette and using it

4) you can tell FRMSAVE to save a file into the output directory directly
   (as long as you have write permission for that directory, of course)

If you find that the enhancement doesn't get loaded before the FRMSAVE
is done, I suggest doing the EU REST IMAGE2 once at the very beginning of
the script and explicitly telling it which frame to do the restore in:

EG B 1
EU REST IMAGE2 1
SF 1
GU MAKE 2 50 0 0
IMGDISP RTGINI/GE1KVIS EU=IMAGE2 STA=ACT
MAP VH 2
FRMSAVE 1 /home/httpd/html/cfdg.gif
MAP SAT 2 COU=ALL 
FRMSAVE 1 /home/httpd/html/cfdg2.gif

>I make satellite images from these every 15 minutes.

I was noticing that you were using the ADDE server
adde.unidata.ucar.edu for something like this.  This machine is not
intended to be used by anyone but Unidata since it is taken up and down
all of the time, and I am testing out code modifications on it before
they are released.  I request that you change your DATALOC for the
dataset you are accessing from this machine to one of the public
servers at cooperating server sites:

papagayo.unl.edu
stratus.al.nooa.gov
cacimbo.ggy.uga.edu
pscwx.plymouth.edu
atm.geo.nsf.gov
adde.ucar.edu

>I want to set up a 
>batch job that will make a nice northern IL sector, with counties,
>the bar on the right, etc. I used to be able to look at the McIDAS command 
>mode and learn the commands and what they did by looking at the screen and 
>figuring it out. With the new GUI menu, that's not possible.

Not true.  The MCGUI has a command mode in which is echoed all of the
commands run from the GUI.  The button with the keyboard icon (next to
the button with the big red Z) is what controls posting/hiding of the
MCGUI command mode.  Left click on the button, and you get the MCGUI
command mode in which commands are echoed.  The commands can be
recalled and executed using the same keystrokes used in the original
McIDAS Text and Command window.  Clicking the middle mouse button over
the keyboard iconed button will pop open the original McIDAS Text and
Command window.  In fact, each time you click the middle mouse button
over the GUI keyboard button, another original McIDAS Text and Command
Mode window will be brought up.  Note, however, that these windows do
not echo the commands run from the MCGUI, and those commands are not
recallable from those windows.  Clicking the right mouse button over
the keyboard iconed MCGUI button brings up a second kind of graphical
command window.  Commands from that window are echoed and are
recallable, but these windows do not echo the commands run from the
MCGUI.

>What can I do to help me figure this out?

Open the primary MCGUI command window and play with the commands run to
load the NEXRCOMP composite(s).

>As always, thanks for all this help...I do appreciate it! And I will make 
>those images available to the entire community once I'm done.

I was hoping that you would also make the other NEXRCOMP images
available (e.g. 10KRCM-NAT, 2KN1P-NAT, and 4KNTP-NAT).  This would make
your server offerings the same as the other ones supplying the NEXRCOMP
dataset.

You should note that these images are very small in comparison to the
1KN0R-NAT ones, so they take up considerably less disk space.  The
other thing you could do is ingest these images and only keep an hour's
worth (a couple to three for the 10KRCM-NAT ones as they are very
small).  You would need to scour these differently than the 1KN0R-NAT
ones, of course.

Tom