Add intervention & restructure simulate_tree_from_pop()

[jamesmbaazam]

This PR introduces two key changes:

• It partly addresses Implement interventions in models #3 by adding a feature to scale R0 through the mean of offspring_dist. The PR introduces the intvn_scale_r0() function for achieving this. The function only works for poison and negative binomial offspring distributions because their mean represents $R_0$. An error is thrown when r0_reduction > 0 but offspring_dist is neither pois nor nbinom.

• In the process of including this feature, it became necessary to restructure the simulate_tree_from_pop() function. It now uses ... to pass extra arguments to pois and nbinom. The offspring_disp and offspring_mean arguments have been removed. This was not an issue previously but became necessary to streamline how intvn_scale_r0() works.

[sbfnk]

Nice work!

[codecov-commenter]

Codecov Report

Merging #97 (7e80d51) into main (a3cbf4a) will increase coverage by 0.10%.
Report is 26 commits behind head on main.
The diff coverage is 100.00%.

❗ Current head 7e80d51 differs from pull request most recent head 036aa88. Consider uploading reports for the commit 036aa88 to get more accurate results

@@            Coverage Diff             @@
##             main      #97      +/-   ##
==========================================
+ Coverage   98.80%   98.91%   +0.10%     
==========================================
  Files           8        9       +1     
  Lines         419      459      +40     
==========================================
+ Hits          414      454      +40     
  Misses          5        5              

Files	Coverage Δ
R/intervention.R	100.00% <100.00%> (ø)
R/simulate.r	98.99% <100.00%> (+0.19%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[pratikunterwegs]

Just taking a look at this after hearing about it in the developer meeting now, would be good to see where there are commonalities with {epidemics} interventions. If the intervention here is primarily affecting a single value, then it's pretty similar to a <rate_intervention> - happy to discuss how these could be brought closer together.