Truncate era dd eds

[kyleredilla]

This PR has changes to the degree day preprocessing notebook with the main change being a truncation of the input years used for ERA-Interim to 1980-2009.

To test, use the new draft geospatial conda env for SNAP geoprocessing pipelines. I.e., use the environment.yml file in the add_conda_env branch of the snap-geo repo (here) to create a conda environment for testing this notebook in.

Make sure the preprocessing notebook runs following instructions and the files import as expected.

This pushes us further towards completing #51 but does not quite close it until we get the new coverages imported into Zeus.

[charparr]

Wow I let this go a bit stale! Sorry about that - looking now. @cstephen what version of ipykernel worked to build the env? I'm hitting what is likely the same issue via a lengthy attempt by conda to resolve the conflicts.

[cstephen]

@charparr Downgrading ipykernel to 6.15.2 in environment.yml is what worked for me. Kyle fixed up the add_conda_env branch of the snap-geo repo to reflect this:

https://github.com/ua-snap/snap-geo/commit/89b458b1c766bdbcf121ef047a6f0caa5f6ca3e2

[charparr]

Whoop! I obviously missed that commit, thanks. I'm attempting to build from that version but still getting stuck in an endless conflict resolve attempt. I'm not sure what's up but I may just nuke + reinstall my conda on Atlas.

[kyleredilla]

Hey @charparr are you trying to create the conda environment on the login node or a compute node?

[charparr]

@kyleredilla I'm trying from a compute node on Atlas. I wiped and re-installed conda. I get to here and then it resolves conflicts for an hour plus and ultimately fails. I've seen conda take a long time to create and resolve environments but this feels a little different.

[kyleredilla]

Dang! This is happening to me now too in trying to create this env. I do not remember bumping into any issues with conflicts previously. Maybe conda updated dependencies or dropped something, I have no idea. I will reconstruct this env spec with one that works and ping here when that's done.

[charparr]

Wow! Yeah I tried building with a fresh conda (seemed like a good way to start 2023) and it had no impact. The fact that you and Craig were able to build, but I couldn't, and now you can't either, definitely makes me suspect some upstream dependency change that now cannot be reconciled.

[kyleredilla]

Okay there is a new version of the environment.yml file in the add_conda_env branch of the snap-geo repo. This one seemed to create in under 10 minutes I think, so let me know if things are taking much longer than that (or if it doesn't work at all, of course).

[charparr]

Thanks Kyle - his .yml worked with about the same ~10 minute build time. Onward!

[charparr]

After a classic Python env kerfuffle I've executed the cells in this notebook! The notebook performs as expected and the resulting files are successful test ingest ingredients on Apollo.