ILAMB task revision

[forsyth2]

Add ILAMB task. Revision of #197. Resolves #133. Resolves #134. A future pull request will handle #229.

[forsyth2]

It took some debugging to create the single initial commit for this pull request. I've listed what I believe to be the necessary steps below:

$ git fetch upstream e3sm2cmip_ilamb
$ git checkout -b e3sm2cmip_ilamb upstream/e3sm2cmip_ilamb
$ git rebase --rebase-merges -i upstream/main
# There are 8 commits:
# pick
# change to fixup
# change to fixup
# change to fixup
# merge -C
# pick
# change to fixup
# change to fixup
# Now, there are 3 commits.
# See https://stackoverflow.com/questions/32840872/convert-merge-into-rebase-without-having-to-perform-the-merge-again
$ git reset --soft HEAD~2
$ git log
# Now, only the first commit shows up
$ git commit --no-verify # Effectively turns the merge commit (and the commit after) into a single regular commit
$ git rebase -i upstream/main
# There are 3 commits:
# pick
# change to fixup
# change to fixup
# Now, there is 1 commit.
$ git commit --amend
# change commit title to "ILAMB task initial commit"
$ git push forsyth2 e3sm2cmip_ilamb
# Create this pull request
# Confirmed that this pull request adds 1,580 lines and removes 2 lines, just as #197 does.

[forsyth2]

@chengzhuzhang This pull request contains two commits. The first is all the work you did on #197. The second is my revisions: 3210894 -- that's what needs a review. Once we merge this, I'll update the expected files for the integration tests.

Notable changes:

• Added parameters (and defaults) for ts_atm_subsection and ts_land_subsection. This enables multiple ts subtasks to be run (e.g., with different values for ts_fmt).
• Changed location of setting export to ALL.
• Made sure cmip_metadata and ts_fmt were defined in the appropriate places.
• Added code to move visuals to web server.
• Changed tmp directory name to avoid name collisions.

[forsyth2]

Also note I'll handle #229 in a follow-up pull request.

[chengzhuzhang]

could this be [[ atm_monthly_180x360_aave ]]? I think the logic is if ts_fmt=cmip, both nco convention time series and cmip time series are generated.
And this sub session can be both a dependency of both e3sm_diags and ilamb.

[forsyth2]

Thanks, it must have been some issue in debugging that made me think it required a separate subtask. It looks like it works fine using the same subtask.

[chengzhuzhang]

Hey Ryan, the PR looks good, I only has one query left as a single line comment. Thanks!

[forsyth2]

Expected files for integration tests (https://web.lcrc.anl.gov/public/e3sm/diagnostic_output/ac.forsyth2/zppy_test_complete_run_www/) have been updated.