33 add optional arg .by for grouping

[jacobvjk]

closes #33
closes #21

• calculation of net alignment metric does not depend on any grouping information (previously depended on group_id, which represented loan books)
• for all functions in calculate_company_alignment_metric.R this means they now truly return company level information without any references to loan books or other groups previously implied by group_id
• aggregate_alignment_loanbook_exposure() gains argument .by which can be used to define the level of aggregation for loan books
  • Default is NULL which means that aggregation happens without grouping. This corresponds to the meta loan book when multiple loan books are processed.
  • .by accepts character vectors as inputs. Providing more than one variable to group by is possible
  • the groups must be available in matched_prioritized

[jacobvjk]

@jdhoffa I think this PR now covers what we want to achieve here. Given the extent of the changes here, I am keeping it as a draft for now and would ask for your initial review. I expect there are probably a few things that could be changed cosmetically that we may want to add as well

[jdhoffa]

A check-point review and some requested changes

Let me know when this is ready for a more in-depth review and I will test it out in more detail locally

[jdhoffa]

NIT: maybe also change the name of the PR, since the argument is now called .by?

[jdhoffa]

lgtm!! 🚀