Granules are note loaded
#1
Hello SEAScope team.
The tool you are developing is purely insane,  so thank you Smile
I have decided to dive in by following latest Summer school's jupyter norebooks provided here https://docs.google.com/presentation/d/1...913eda_0_0 .
In the first one (about Lagrangian advection) one has to load some current granule.  
I've encountered the problem when tried to load a granule of a "GlobCurrent geostrofic" data collection.
After drawing a polygon, selecting it in SEAScope Viewer (seascope-processor is running), I push an "Extract granule data" and then a "Python object" buttons and the link to a '.pyo' file is displayed only. There are actually some granules of this data collection in the ../seascope/data/globcurrent_l4_geostrophic/ directory, but they are not displayed in the terminal.
In the aforementioned notebook it is suggested to load a granule using link to the '.pyo' file (I have it), and when I execute the code in cells Jupyter doesn't load any granules for some reason. It loads an empty list instead. So why is that? Maybe this "loading granule via link to .pyo" business doesn't work with stream lines data?

OS is Ubuntu 18.04
Version of Python is 3.6.8

Thank you in advance.
Reply
#2
Hi,

Thank you for the kind words about SEAScope Smile


You guessed right: the extraction method doesn't work with streamlines (it won't work with arrows/barbs either), only with data rendered as raster images (or trajectories).

This information is in the document, but it is quite easy to miss, we will make it more visible in the future versions of the training course instructions:

Quote:Select the current_norm raster variable from the Geostrophic GlobCurrent current collection in SEAScope (you can try different currents later)

Can you try again with the raster and tell me if it works for you?

Sidenote:

There are two ways to load data in the notebook:
  1. extract data and fetch the extraction directly from SEAScope memory (this is the default method in the notebook)
  2. extract data, send it to seascope-processor so that it saves the extraction as a file and then load the file in the notebook
So if you don't modify the notebook and follow the instructions, the notebook will use the first method and you won't have to start seascope-processor at all.
Reply
#3
Sylvain,

Yes. Changing defaultRendering option in the corresponding config.ini file of the data collection to RASETR made things work. 

Thank you.
Reply
#4
The instructions clearly state that the variable you must select is "current_norm raster", not the "current speed streamlines". So if you follow the instructions, it should work out of the box and not configuration changes are required.
Reply
#5
(2019-07-08T06:31 AM)Sylvain Herlédan Wrote: The instructions clearly state that the variable you must select is "current_norm raster", not the "current speed streamlines".

The viewer doesn't display any variables other than "current speed" (rendered as steream lines) in a GlobCurrent collection (see the attached screenshot).
There is no code unit dedicated to different rendering in the corresponding config.ini file of the collection.
So I am then forced to compose it myself.

Oh, my apologies!

I missed the link to the custom configuration file!
That is the greatest blunder I have ever had.

I thought I had a bit of an attentiveness, but ...
I am really sorry for that Sad 
 
Actually, I used the presentation as a source of notebooks only and decided to go through them without considering a mere possibility that something additional would be required.

Thank you for the support and patience.


Attached Files Thumbnail(s)
   
Reply
#6
Hi again,

No problem, now I understand why you struggled! I thought you got some info from students who attended the training course but in fact you really started from the information listed on the website!

There is one very important thing missing on the website because the info had already been passed down to students by email: a link to the data used during the training! We didn't expect people who did not attend the ORSS to follow the course afterwards, so we didn't put the links to the data bundles on the website.

So don't blame yourself, it is quite impressive that you managed this far without this data bundle!

You can download the full SEAScope installation (Linux) used during the training course, including the data, here: https://www.oceandatalab.com/orss2019/or...ope.tar.gz
Reply
#7
Sylvain,

Now everything works: drawing trajectories of arbitrary drifters and swell rays and the rest! That is 
impressive. 

I would have applied for the training course had I monitored oncoming events before. From now on I shall keep it in mind Smile

I have already looked up some case data collections, which are pretty good to discover the matter.
Thank you ever so much.
Reply


Forum Jump: