Cache fim catfim

[shawncrawley]

This PR contains all of the latest CatFIM updates made over the last couple of months per FIM team feedback. The biggest and most notable of these updates is the processing and inclusion of the Stage-Based CatFIM "Interval" sublayers.

[TylerSchrag-NOAA]

Thanks Shawn!

Nick - I'm going to add a little to this, and will review/approve when it's ready.

[nickchadwick-noaa]

Sounds good, thanks for the help Tyler!

[TylerSchrag-NOAA]

Ok, I'm admittedly not super happy about how hacky/redundant some of this had to turn out. - Kind of 3 steps forward, 2 steps back on the whole abstraction of the HAND Processing workflows. But all of CatFIM is [mostly] working now with the new setup. I'll make some specific comments in the code itself.

Generally speaking, I created a 'alternate_hand_preprocessing' folder in the fim-data-prep lambda function, which houses sql files where additional data prep is needed before running HAND. This could theoretically apply to any FIM Config, but at present it's being used for the FIM Configs where we don't utilize the HAND Cache, i.e. AEP FIM and CatFIM, but do want to do some preprocessing/loading, i.e. pull from the Ras2FIM cache.

I shoulda/woulda/coulda thought through these AEP FIM and CatFIM requirements when designing the Cached FIM enhancement, but as a future optimization I would recommend something like:

1. abstracting HAND processing into its own modular component, that essentially acts like a super simple API. Give it a list of features with flows (or stages), and get back those features with geometry extents. I'd let it handle all the FIM crosswalk stuff as well (the fim_caching_templates sql files I setup in the postprocess_sql lambda are a step in this direction).
2. Have an abstract aggregator that exists right before that step in the pipelines, that distills all the features of any set of fim_configs down to that single unique list of features and flows (or stages) to pass off to hand_processing.
3. Adjust the product processing steps of the pipeline to pull from a) the cache, and b) a shared hand_processing output table to get the cumulative fim extents that it needs (it would be good to standardize the fim caches (ras2fim and hand) into a single cache source as well that can play nice in a simple way with this framework).

[TylerSchrag-NOAA]

So I'm sure this has to do with the triple-bracket code you added to viz_classes.py at the top for sql_replace... but I'm getting job_num in the sql_replace dictionary instead of {job_num}.

I just removed the brackets in the sql code for now to get it working... but you'll want to fix that. I'd recommend adding some code to initialize_pipeline to insert those template sql items in the product config into the sql_replace dictionary at the very beginning of a pipeline, so they don't need to be added in fim-data-prep and postprocess-sql separately (that would also be a good opportunity to add my new fim_caching sql_replace params to that same logic in initialize pipeline, per my sloppy code in postprocess_sql, sorry).