Enhancements to epidemic_size()

[pratikunterwegs]

This PR addresses issues raised in #190:

1. epidemic_size() now optionally returns a <data.table> of epidemic sizes at specified timepoints via the logical simplify argument; the default behaviour remains returning a numeric vector (for backwards compatibility);
2. Epidemic sizes at multiple timepoints can be requested by passing a vector to stage;
3. Epidemic sizes can be requested at specific timepoints (e.g. $t = 100, 200, 300$) instead of a proportion of epidemic time;
4. A data.table is automatically returned, overriding the simplify argument, if a vector is passed to stage or time.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if e99365a is merged into main:

• ✔️default_ode: 14.5ms -> 14.5ms [-1.23%, +1.16%]
• ✔️default_ode_interventions: 75.8ms -> 75.5ms [-1.11%, +0.39%]
• ✔️default_ode_param_vec: 848ms -> 848ms [-1.15%, +1.06%]
• ✔️default_ode_paramvec_intervs: 6.33s -> 6.34s [-0.71%, +1.08%]
• ✔️ebola: 39.4ms -> 39ms [-4.34%, +1.91%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if e99365a is merged into main:

• ✔️default_ode: 13.9ms -> 14ms [-0.57%, +1.04%]
• ✔️default_ode_interventions: 71.5ms -> 71.7ms [-0.39%, +0.84%]
• ✔️default_ode_param_vec: 823ms -> 821ms [-1.06%, +0.66%]
• ✔️default_ode_paramvec_intervs: 6.16s -> 6.17s [-0.17%, +0.67%]
• ❗🐌ebola: 36.8ms -> 603ms [+1531.63%, +1547.31%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 21ebc61 is merged into main:

• ✔️default_ode: 14.5ms -> 14.4ms [-2.99%, +1.81%]
• ❗🐌default_ode_interventions: 71.7ms -> 72.2ms [+0.02%, +1.44%]
• ✔️default_ode_param_vec: 823ms -> 825ms [-0.37%, +0.96%]
• ✔️default_ode_paramvec_intervs: 6.27s -> 6.24s [-1.61%, +0.78%]
• ✔️ebola: 629ms -> 629ms [-0.89%, +0.77%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[jamesmbaazam]

This looks great, @pratikunterwegs. The enhancements provide more flexibility to the user. It all LGTM. I left some comments, where I thought some changes could be made to your own satisfaction.

[pratikunterwegs]

Thanks @jamesmbaazam, will address these and merge later today.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 21ebc61 is merged into main:

• ✔️default_ode: 11ms -> 10.9ms [-2.27%, +0.45%]
• ✔️default_ode_interventions: 73.4ms -> 72.7ms [-1.94%, +0.02%]
• ✔️default_ode_param_vec: 823ms -> 823ms [-0.42%, +0.46%]
• ✔️default_ode_paramvec_intervs: 6.24s -> 6.25s [-0.26%, +0.6%]
• ✔️ebola: 611ms -> 609ms [-1.09%, +0.35%]
  Further explanation regarding interpretation and methodology can be found in the documentation.