Process creating NcML file aggregating netCDF files

[huard]

Overview

This process creates two netCDF files, and an NcML aggregating those files. This is intended to be installed on a test environment, so that when the process is called, files are created and picked up by the THREDDS server. Clients can then test that the NcML endpoint works as expected.

[huard]

@cehbrecht Can this be useful in the context of the CDS and Thursday's discussions ?

[cehbrecht]

@huard You can present this as an example for ncml usage in the discussion on Thursday. Maybe you like also to add an example with ncml where the metadata of the files gets "fixed"?

I have put my thoughts for the discussion here:
https://github.com/cehbrecht/climate-operators-prototype/blob/master/README.md

Looking at the ESMValTool we expect fixes that can not be done with ncml. Also the ESGF Future Architecture might not use Thredds/OpenDAP anymore. Just plain file downloading using an Nginx web service. I thought one could use FUSE (File System in User Space) to provide a virtual data layer as a file-system. The FUSE could apply the data fixes on top of the original data when data gets requested. Probably a caching is wanted here as well. But I can not say if this is possible and reasonable ...

[cehbrecht]

See inline comments.

[huard]

Will do, this was just a smoke test. We still need to figure out the client part of the test.

Is Emu the right place to put these types of test processes? I could maintain an Emu branch with those processes if you feel that it will confuse users.

[cehbrecht]

Is Emu the right place to put these types of test processes?

I think Emu is ok. We can move them to other places later if we have the feeling they belong somewhere else.

[huard]

So the process currently returns an URL, but for NcML we'll need an opendap link.

[huard]

Requires geopython/pywps#515 and geopython/pywps#513

[huard]

@cehbrecht I think this is ready to merge.