Enhance .cwl file and UKCP18 additions

[ciaransweet]

What this PR is

Syncing up DevSeed changes we've been working on!

Primarily:

• I modified and renamed the .cwl file in the repo to now be hazard_workflow.cwl
• It now has a fleshed out section in the README, an example input, and covers off a few other inputs it didn't previously expose
• I renamed extra_xarray_store to write_xarray_compatible_zarr to make it a bit more obvious what that flag does
• Added support to write the osc inventory.json format, STAC jsons, or both formats
• Modified the Ukcp18 dataset to be able to run on different resolutions from CEDA
• Parameterised threshold_list and threshold_temperature so we can change them per run config
• Parameterised dask_cluster_kwargs so that we can make processing sequential (as needed for memory restricted environments)
• Changed src/hazard/sources/ukcp18.py's fsspec.filesystem to be a filecache ftp implementation. This will store downloaded files under /tmp/ukcp18cache/ so we don't have to download GB files every year of calculations
• Removed collection as a parameter to pass to the UKCP18 dataset, this can be derived from the desired resolution
• Fixed an issue caused by dynamic CRS handling for global 60km data
• Modified the STAC item generation to template out the values of permutations

Thanks!

[joemoorhouse]

Hi @ciaransweet,

Thanks for this! A question about one of the latest commits that separates out the source data info from the definition of the indicator. (I might be missing something but) I don't see where the source data is added back into the inventory item description. Also I think the info about the averaging periods being used is removed, but I'm not sure that is added back in (e.g. programmatically)?

Thanks, Joe

[ciaransweet]

Hey @joemoorhouse - truth be told those commits probably jumped the gun. Joys of GitHub and forks, I'd rather have kept this PR the commits when I hit the button but I guess our other changes into main have come in.

We do indeed plan on moving the source data info it's own property, just so we aren't relying on the combination being hardcoded. Indeed we could probably just have the description be combination of the two separate values.

As to the averaging periods, let me get back to you on that. I suppose it's something we can generate programmatically from the indicators parameters..

Sorry it turned into a big PR, I'll specifically raise a branch to merge into here next time, to save the growing Pr scenario

[ciaransweet]

@joemoorhouse again, sorry for the big PR, this is it for this one 🤞

https://github.com/os-climate/hazard/pull/118/commits/a202d11d9b192366d187f3799319c5512613b77d this commit re-jigs the descriptions for days_tas_above and degree_days (The indicators we've been dealing with up to this point)

I've undone some of the other indicator description edits as we've not dealt with them and there's probably chances to improve those later.

For the ones I have changed, they now generate their description from 3 components:

• The indicator .md
• The source datasets .md
• The scenario + year permutations

This gives us something like:

This frontend pulls the description hazard generates and renders it

Thanks again for taking a look!