Optional pre-parsing of html

[dghelm]

Feature Suggestion: Currently, if you want to know multiple details about a case, you pass the html to be parsed by BS4 for each parse method (attorney, counts, etc) which requires a new parsed soup for each one.

This could be lazily computed on the Case instance (in case it doesn't need to be done at all) and then passed to each parser.

This could also allow for 'lxml' or other parsers to easily be subbed in instead of "html.parser".

I used a messy implementation of this on a project with lots of cases, and between this and the other parser, was able to get good speedups when I wanted to fully parse all details from a case.