Read in descriptions for nested properties

[guineveresaenger]

Uses a nested path lookup for properties whose descriptions come to us via entityDocs.Attributes.

Allows for the regex to match against . in nested output names.

Fixes #1774.

This is a draft so far as there's a few issues here.

update: Please see a build of the AWS provider docs, for a schema built with these changes where specific missing fields are now correctly populated:

MQ Broker Instance before
MQ Broker Instance after

Lambda Function Snapstart Before
Lambda Function Snapstart After

Elasticsearch DomainVPCOptions Before
Elasticsearch DomainVPCOptions After

** Notes**

1. I discovered that a majority of nested property descriptions (at least in AWS) come from what is described as a "fallback" in the code comment, which I edited to add that information.

2. This fixes this bug where extra bullet points in a property description were attached to the previously-parsed entry.

Before:

After (localhost):

3. However, this may need some more work to properly map multi-nested documentation as seen in mq_broker.instances which still results in missing descriptions on our end

[iwahbe]

This looks directionally good.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 60.24%. Comparing base (3a2ed13) to head (0117474).
Report is 10 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1798      +/-   ##
==========================================
- Coverage   60.60%   60.24%   -0.37%     
==========================================
  Files         303      310       +7     
  Lines       42348    42682     +334     
==========================================
+ Hits        25667    25713      +46     
- Misses      15208    15496     +288     
  Partials     1473     1473              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[iwahbe]

Where do these lists come from? Are they always .0, never .N or .1?

[guineveresaenger]

This is an example of a TF schema.TypeList which I believe only ever uses the .0 to access the list elements. I'm admittedly a bit fuzzy on this so I'm happy to be corrected here.
I also searched the p-aws schema as well as upstream/website for any indications of !=.0 and came up with nothing. I know that's not proof and if you want to make this digit-agnostic, we can.

[iwahbe]

I trust your check. This is the type of information that would make an excellent comment on line 974.