dtcenter / METviewer

Tool that creates plots using MET verification statistics output and the R statistical package
http://www.dtcenter.org/met/metviewer/
Apache License 2.0
14 stars 1 forks source link

create a UI for vieweing a scorecard #156

Open TatianaBurek opened 5 years ago

TatianaBurek commented 5 years ago

During the scorecard calculation we can create a XML and an image for each cell. The links to these files can be added to each element so users can view the plot and/or load XML to METviewer/METexpress and recreate the plot. Project key 2783533

scorecaard_ui

TatianaBurek commented 1 year ago
  1. The GSL scorecard that you will see in the video presentation has a web GUI to construct the final grid. In the METviewer case, it would be a big challenge to implement a similar interface. Would it be OK to continue working with XML and the batch mode? -- For now, working with XML and batch mode is good. If we happen to be swimming in money when it is all said and done (ha), we can pursue a different interface.
  2. Where will this application be installed and run? RAL server? Someplace else? -- RAL server would be the primary target. If it gains traction, I could see folks at EMC wanting access to this application as well.
  3. As you know very well :), METviewer uses data that was previously loaded to the database. Which is an extra step that as I understand you try to avoid. I see that you put a lot of effort to use MET Analysis tools working from the command line instead. Would it be OK to leave the current design with the database or the intent is to use "the command line" scorecard which means, of course, extra time and effort. -- Ah, good question. For this scorecard work, we fully intended it to be run the traditional way, where we load the data into a database. We wanted to explore command line options for some of the standard plots we always make, mostly to make the functionality available on an HPC (like Hera), so folks w/o access (or experience) with METviewer could easily make a few plots after running an experiment. Our group still intends on using METviewer for plots and analysis (especially since it is a lot more customizable at this time).
  4. How do you envision the results? Some standalone web page? Part of the METviewer app? -- Honestly, I don't know! What would be easiest or more scalable from your perspective?
  5. Issue #156 lists a lot of ideas that might be not possible to implement for this budget. Would displaying a plot without the XML part be OK? -- Yes, displaying a plot without an XML is fine. We requested more $$, but it got cutback, so let's keep the ideas we don't have money for in mind for the future.

Scorecard example: gfs_rap_amj_i00_scorecard_surface.xml.txt