Yannick opened the meeting by announcing the focused topic of the day, which is the ongoing migration to ioda version 2 and the upcoming release.
We have discussed this in the past but this is one more chance for clarifying what will happen (hopefully) next week (probably) - namely the merging into develop of the ioda-v2 feature branches of ioda and ufo.
Steve proceeded to present the following slides
During the presentation, JJ asked "Is there documentation of which files need "-n" argument during conversion?" (referring to slide 7). Steve and Ryan answered that this is only needed for a very small minority of files. They recommended running the file upgrader without the "-n" option. Then, only use if it that gives you an error. There is a diversity of content and file formatting in the older ioda files and the upgrader should work for nearly all. If it does not, let Ryan and Steve know.
Jake asked "previous meetings mentioned some slowness issue, is it fully resolved?"
Steve responded that the ioda-v2 branches are now running with about the same time as the develop branches for the tier 1 unit tests. The next step is to run the JEDI-GDAS testbed application on Orion and compare the timings with develop. He plans to do that today.
Mike Cooke asked: "Are the output files used in testing going to stay in netcdf format or move the h5?"
Steve answered that they will stay in netcdf4 format and we have taken care to ensure that the ioda v2 files can be read in with either hdf5 or netcdf APIs. For example, you can still use the netcdf python module to read them in.
Yannick then asked to clarify the timing. We hope to be ready to merge the feature/ioda-v2 branches into ioda and ufo by next week. After the merge, we will likely need at least a week for code clean-up and documentation, and another week of freezing the code before the release.