Hi All,
I was climbing through the ERDDAP code and the erddapTest repository. I discovered some ruby code that I could not run as the ruby-netcdf package is not apparently available in the Ubuntu 22.X version of the OS. So, I went as far as creating a docker image that allowed access to a Ruby 2 ruby-netcdf package that allowed me to run those scripts again.
Doing some deeper digging, a ruby-netcdf package will appear in Ubuntu 24.04 with support for Ruby 3 allowing that code to work again. However, this seems to be internally maintained by Debian. The main source code site still only supports Ruby 2.
As an exercise, I slowly creating companion python/xarray versions of that code. What may transpire is creation of a generic repo of netcdf examples for testing with ERDDAP and various compliance checkers. This will include generation of some examples using the OG-1.0 format specification. From my current investigations, ERDDAP holds the best diversity of examples at the moment.
I wanted to ask what was the plan for the ERDDAP test suite and examples? It looks like there might be three copies of examples?
1. Examples embedded in the java code that include the example and how to use the tools to generate a dataset.xml.
2. Ruby code and examples as part of the errdap repo.
3. Ruby code and examples in the erddapTest repo.
Where is the best place to look for current authoritative examples of the three locations mentioned above?
Thank you,
Rob
PS> This exercise largely helped me formulate a pathway for the acoustic dataset. I had a devil of a time figuring out an appropriate structure. Once I was able to stand up a "CF trajectoryProfile Multidimentional SingleTrajectory H.6.2" example, I was set.
The _acoustic version is a flattened version that works, but not 100%. The _CF represents a better structure. The _OG is the OG-1.0 format. The only real difference at the moment between CF and OG is the variable names are CAPITAL and the recommended dimension names are different.