DAS-NONE: Move trajectory subsetter tests to shared_utils

[flamingbear]

Description

I saw a terrible example in the shared_utils readme linking a xarray githash. I cleaned that up, and thought let's just add trajectory subsetter to use the shared_utils since that's where the idea came from and that word was never going to get scheduled.

This PR updates Trajectory Subsetter test to use shared utilies, but also refactors them to remove the old datatree comparison and renames the old compare_results_to_reference_file_new -> compare_results_to_reference_file

This PR also separates out the basic shared utilities into utilities.py and compare.py so that if you're not using xarray comparisons you don't need to include it in your environment for your tests.

Jira Issue ID

None

Local Test Steps

Check out this branch. Build the trajecotry-subseter and nsidc-icesat2 images.

make trajectory-subsetter-image
make nsidc-icesat2-image

Set the appropriate environment variables for UAT and run the notebooks

# UAT
export HARMONY_HOST_URL=https://harmony.uat.earthdata.nasa.gov
export EDL_USER=
export EDL_PASSWORD=

./run_notebooks.sh trajectory-subsetter nsidc-icesat2

Verify they both succeed.

PR Acceptance Checklist

• Acceptance criteria met
• Tests added/updated (if needed) and passing
• Documentation updated (if needed)
• CHANGELOG updated with the changes for this PR

[lyonthefrog]

I ran the tests and both notebooks passed.

I made a few simple simple comments and questions, so when those are resolved I'll approve!

[lyonthefrog]

HDF5 files in Trajectory Subsetter regression test? Also, could this be a shared utility or do we not really use it anywhere else yet?

[owenlittlejohns]

There is something similar in the #95 and, from the documentation string here, I'm going to go out on a limb and suggest maybe there is a similar function in the Swath Projector test utilities.

[flamingbear]

I commented below

I reckon the remove_results_files could be moved into the shared utilities (maybe with an optional argument or two, such as a directory to clean up, or a file extension), but I'm not going to hold the PR up over that.

I prefer using a temporaryDirectory, so I didn't move it into the shared group. I just didn't up date the trajectory subsetter tests to use a tempdir.

[flamingbear]

I could be convinced to move this to shared, but not for this PR. And I'd need a decent reason probably, because I'd bet they're already different. just looked, they are different, this one has the wrong comment on it.

[flamingbear]

I fixed the comment though: e955ef7

[lyonthefrog]

Same question here

[flamingbear]

same answer, just metadata, the version of python used when the notebook was edited.

[owenlittlejohns]

I think this looks great.

It's a shame about the limitations on conditional logic in a Dockerfile. If that was possible, I'd probably recommend just adding something there to install the packages for these utilities when they are to be included.

I reckon the remove_results_files could be moved into the shared utilities (maybe with an optional argument or two, such as a directory to clean up, or a file extension), but I'm not going to hold the PR up over that.

[owenlittlejohns]

There is something similar in the #95 and, from the documentation string here, I'm going to go out on a limb and suggest maybe there is a similar function in the Swath Projector test utilities.

[flamingbear]

I reckon the remove_results_files could be moved into the shared utilities (maybe with an optional argument or two, such as a directory to clean up, or a file extension), but I'm not going to hold the PR up over that.

I prefer using a temporaryDirectory, so I didn't move it into the shared group. I just didn't up date the trajectory subsetter tests to use a tempdir.