search

E3SM-Project / e3sm-unified

A metapackage for a unified anaconda environment for analyzing results from the Energy Exascale Earth System Model (E3SM).
BSD 3-Clause "New" or "Revised" License
8 stars 8 forks source link

Update to v1.3.0 #55

Closed xylar closed 5 years ago

xylar commented 5 years ago

This version drops python 2.7 support.

Main packages updated:

rpwagner commented 5 years ago

Nevermind, I triggered it. (Had to log in.)

@xylar if you want to trigger Travis CI on this, you may need to do a minor commit to the from branch. It's of limited use until we address #58, though.

xylar commented 5 years ago

@rpwagner, this branch is very much a work in progress so I'm not worried about re-triggering CI. That'll happen naturally when I make changes here.

I'll update CI to use python 3 and drop the pinning on e3sm_unified as soon as I've uploaded new versions. That will at least be a back-stop on #58 for now.

rpwagner commented 5 years ago

Makes sense. I'll stop changing things for now.

sterlingbaldwin commented 5 years ago

The current version of processflow is 2.2.3.

chengzhuzhang commented 5 years ago

We will use updated version of output_viewer 1.3.1

xylar commented 5 years ago

We will use updated version of output_viewer 1.3.1

Sorry about that! I missed that update, I guess. Just updated and I'm testing a build now.

xylar commented 5 years ago

The current version of processflow is 2.2.3.

@sterlingbaldwin, that's good to know, but I'm assuming you might need to have some testing and a new release once e3sm_diags and mpas-analysis are updated. Or do you think it'll just work?

sterlingbaldwin commented 5 years ago

Im going to need to do testing, and if there're problems Im going to have to release a new version, so you might end up using 2.2.4.

xylar commented 5 years ago

@sterlingbaldwin, I'm currently not able to install e3sm_diags (because we've been waiting on a CDAT release) with the rest of e3sm-unified so I've been holding of on that one. Obviously, I can't install processflow without e3sm_diags so I've been holding off on that, too, for now. That's why I haven't bothered to put in a version of processflow yet.

xylar commented 5 years ago

@sterlingbaldwin, are you able to test with a local build of e3sm_diags using this PR? https://github.com/E3SM-Project/e3sm_diags/pull/256 First, do a local build: In the meta.yaml, change the path to .., since the release hasn't been made yet. Then, run:

conda build -c conda-forge -c cdat/label/v82 -c e3sm meta.yaml

If all goes well, you can do:

conda create -n processflow_test -c conda-forge -c cdat/label/v82 -c e3sm python=3.7 e3sm_diags=2.0.0 mpas-analysis=1.2.5 ...

I haven't got the go-ahead yet to release mpas-analysis 1.2.6 so that will need to wait. I don't think you'll have trouble with the latest release, since nothing big has changed about the config files.

xylar commented 5 years ago

With the builds from E3SM-Project/e3sm_diags#256 and https://github.com/E3SM-Project/processflow/pull/165, I was able to build what I hope to be the final version of e3sm-unified 1.3.0 under linux. So as soon as a new processflow package is build (whether a new release or a rebuild of the existing 2.2.3) and e3sm_diags 2.0.0 is released, we should be set!!!

xylar commented 5 years ago

@rpwagner, I updated the definition of the container and CI is failing (timed out in one case, can't find singularity in the other). Could you have a look? It may be better to address this in another PR unless it's easy.

rpwagner commented 5 years ago

@xylar sure, I'll look shortly.

xylar commented 5 years ago

Okay, I think CI is finally fixed. @rpwagner, the fix was to add travis_wait so things don't time out. I also pointed to a specific singularity tag rather than a release branch (which seems to be transient) and fixed a few other small issues.

rpwagner commented 5 years ago

All makes sense to me. Let me know if you need me to look at anything else.