GOES-16 Satellite Image - Move to Back? / Re-order Layers

Post any questions, ideas, or topics related to Jython and Python scripting.
User avatar
joleenf
Posts: 1120
Joined: Mon Jan 19, 2009 7:16 pm

Re: GOES-16 Satellite Image - Move to Back? / Re-order Layers

Post by joleenf »

Hi,

This is happening on a GNU/Linux box. The map is displaying fine on the IOS machine I am using (McIDAS-V version 1.8beta1 Build-date: 2018-08-06 09:30), but not on the GNU/Linux box (McIDAS-V build from today 2018-08-17, however, was happening with the build from August 6th as well). In that case, the map shifts to the bottom of the display layers after the zoom.

Process:
1.) buildWindow at 2160x2160.
2.) display base visible image layer
3.) display sandwich layer (RGB Composite)
4.) set layer to projection
5.) zoom in to fill window with displayed layers

GNU/Linux capture image
GNU/Linux capture image


Joleen
User avatar
bobc
Posts: 938
Joined: Mon Nov 15, 2010 5:57 pm

Re: GOES-16 Satellite Image - Move to Back? / Re-order Layers

Post by bobc »

Hi Joleen,

Here's a script I've been testing with to try replicating the problem:

Code: Select all

m1Parms = dict(
    server = 'lead.unidata.ucar.edu',
    dataset = 'RTGOESR',
    descriptor = 'M1',
    position = -1,
    )

tList = listADDEImageTimes(**m1Parms)
dataTime = tList[0]['time']

vis = loadADDEImage(band=2, size=(500,500), mag=(-4,-4), **m1Parms)
ir = loadADDEImage(band=14, size=(500,500), mag=(1,1), unit='TEMP', **m1Parms)
vis2 = loadADDEImage(band=2, size = 'ALL', **m1Parms)
sandData = sandwich(ir,vis)

panel=buildWindow(2160,2160)

visLayer = activeDisplay().createLayer('Image Display', vis2)
sandLayer = activeDisplay().createLayer('RGB Composite', sandData)

I've been unable to replicate the issue thus far. Could you give this script a try and see if you see the problem of the data overlapping the map? If you do, would you mind running the script in a McV session started with "-tempuserpath" and let us know if you see the problem there? I'm basically trying to track down if it's something with a McV setting or preference that might be causing the problem.

Thanks,
Bob
User avatar
joleenf
Posts: 1120
Joined: Mon Jan 19, 2009 7:16 pm

Re: GOES-16 Satellite Image - Move to Back? / Re-order Layers

Post by joleenf »

Hi Bob,

This does not replicate the problem for me, either in a VNC viewer or in a background script through a vncserver. I was able to replicate the problem in the vnc viewer using the exact scripts I have for the real-time processing of the sandwich product. However, once I set the vertical position of the map upward, the problem disappeared in the viewer, but not in the background scripts. What is odd about that change, is that 5 weeks ago (2018-07-14 17:35:13 -0500), I took out the setVerticalPosition on the map, because it was interfering rather than helping with map visibility on the same machine. During that week, I was using a nightly from July 9th.

However, perhaps the most important information is that this replicates the problem for me when I run from a script in the background through a vncserver and when I view in the VNC viewer (addition of setting projection to fixed grid format):

Code: Select all

m1Parms = dict(
    server = 'lead.unidata.ucar.edu',
    dataset = 'RTGOESR',
    descriptor = 'M1',
    position = -1,
    )

tList = listADDEImageTimes(**m1Parms)
dataTime = tList[0]['time']

vis = loadADDEImage(band=2, size=(500,500), mag=(-4,-4), **m1Parms)
ir = loadADDEImage(band=14, size=(500,500), mag=(1,1), unit='TEMP', **m1Parms)
vis2 = loadADDEImage(band=2, size = 'ALL', **m1Parms)
sandData = sandwich(ir,vis)

panel=buildWindow(2160,2160)

visLayer = activeDisplay().createLayer('Image Display', vis2)
sandLayer = activeDisplay().createLayer('RGB Composite', sandData)

activeDisplay().setProjection(sandLayer)



Joleen
User avatar
joleenf
Posts: 1120
Joined: Mon Jan 19, 2009 7:16 pm

Re: GOES-16 Satellite Image - Move to Back? / Re-order Layers

Post by joleenf »

Hi Bob,

Have you been able to replicate this problem when projecting to the fixed grid?

Joleen
User avatar
bobc
Posts: 938
Joined: Mon Nov 15, 2010 5:57 pm

Re: GOES-16 Satellite Image - Move to Back? / Re-order Layers

Post by bobc »

Hi Joleen,

Sorry for the delayed response to this one. I ran the last script you sent (that includes setProjection()) from the foreground and background several times on one of our Linux 64-bit testing machines as well as my Windows 7 machine and I've still been unable to get the problem to appear. I'm out of the office until Wednesday, but when I get back in I'll spend some more time trying to replicate this on some of our other testing machines (another Linux-64bit one and a macOS laptop).

Thanks,
Bob
User avatar
joleenf
Posts: 1120
Joined: Mon Jan 19, 2009 7:16 pm

Re: GOES-16 Satellite Image - Move to Back? / Re-order Layers

Post by joleenf »

Hi Bob,

I wonder then if this problem is currently only affecting the programs run in a vncserver display. Maybe I can demo the problem on Wednesday.

Joleen
User avatar
bobc
Posts: 938
Joined: Mon Nov 15, 2010 5:57 pm

Re: GOES-16 Satellite Image - Move to Back? / Re-order Layers

Post by bobc »

Hi Joleen,

Thanks for meeting up yesterday and demonstrating the problem. I updated Inquiry 2700 with the latest findings:

  • The problem only exists when running with a VNC viewer, not running locally on the machine.
  • The problem appears when setProjection is used. You can set the projection to 'World' or the projection of the ABI data and you'll see the problem of the data overlapping the map. If auto-set projection is enabled and setProjection is not used, then McV will snap the display to the projection of the data (as expected) and the data does not overlap the map.
  • When setProjection is used, you can work around the problem by bringing the map layer to the front with mapLayer.setVerticalPosition(-.96) or mapLayer.displayableToFront().

If any of this sounds incorrect, or if you have anything to add, let me know. Looking at the GOES-16 sandwich product online, it appears that the script is still running correctly, as the image generated about 10 minutes ago has the map in front of the data.

Thanks again,
Bob
Post Reply