fix company level sda targets

[jacobvjk]

closes #445

This PR:

• ensures that target_sda(by_company = TRUE) uses the emission intensity value of the final year of the adjusted_scenario_* to calculate the target_* values for companies when using the SDA approach
• this ensures that - regardless of whether the SDA is applied at company or loan book level - the individual scenario targets converge at the correct scenario value
• the issue had been caused by a right_join, that introduces NAs for every year past the emission intensity projection when running target_sda() at the company level.
• some implementation details:
  • previously, target_summary_groups was adjusted to loan book or company level at the top level of target_sda()
  • it is moved into compute_loanbook_targets(), which gains an argument, by_company, to adjust the grouping variables as needed.
  • the former functionality in compute_loanbook_targets() was one pipe, which is now split into multiple parts.
  • the first part is only the right_join that was previously used too
  • the latter part of the pipe also stays intact
  • we add an intermediate step that is triggered in case targets are calculated on the company level. This optional step ensures the adjusted scenario_with_p values are correctly merged for every individual company, ensuring the relevant convergence value exists when calculating company targets
• the snapshot for the company level target_sda() calculation is updated, as it previously contained NAs and wrong values due to this bug

[jacobvjk]

• checks pass locally
• failing checks are likely due to the same reason mentioned by @jdhoffa here:

[Created on 2023-09-01 with reprex v2.0.2](#442 (comment))

[Antoine-Lalechere]

LGTM