FDA SPL Product to Substance (Active/Inactive/Moiety)

[jrlegrand]

Resolves #79

Explanation

Added FDA SPL inactive ingredient, active ingredient, and active moiety staging tables/views.

Rationale

To support the Purdue research use case.

Tests

Ran DAG several times. Tested a query.

select * 
from flatfile.mthspl_product_to_inactive_ingredient
where product_rxcui in (
	select distinct product_rxcui
	from flatfile.mthspl_product_to_active_ingredient
   	where lower(active_ingredient_name) in('cetirizine', 'cetirizine hydrochloride')
)

[jrlegrand]

@Bridg109 - NOTE: I was getting some errors from the api-rxnorm-ingredient step of the rxnorm_full DAG but I think it's unrelated to my build in this PR.

[Bridg109]

This PR runs and returns data that looks like it should be useful to the researcher.

The one thing I would suggest it to change to TRUNCATE tables methods, instead of DROP TABLE. This will add benefit as we don't need to cascade which might cause issues on API and it also allows us to do append data as needed. See NADAC PR for an example

[Bridg109]

Should we consider adding in a API for 11 digit NDC for the active and inactive ingredient? Since we have the info seems like a quick thing to add the API

[Bridg109]

we should also set a start date so this runs with out us triggering it.

[Bridg109]

Could also consider merging the read sql functions into a for loop like done in NADAC. If we start trimming these then making a class and implementing will be easier...or maybe we just wait till we make the class

[Bridg109]

Could consider merging the api keys calls into a single task with the download dataset. This would enforce atomiciticy on the dag...which is a BP

[jrlegrand]

Noting that most of these comments are about the RxNorm DAG as a whole, I would feel better about addressing these when we eventually do a class-based / TRUNCATE overhaul of our DAGs. I added it as a note to other tickets that relate to those topics so we don't forget.