Add symptome-compartment model and test (#3)

[hfldai]

Logs on branch i3-symptome-compartment

Model specificities

• [x] Simplify the infectious groups
• [x] Include unittest
• [x] Add disease-related deaths
• [x] User-friendly naming for variables and parameters
• [x] Refine symptom compartments: asymptomatic, mild, severe
• [x] Different lambdas for each of the mixings
• [x] Include R0 calculation for the system
• approach: next generation matrix
• [x] Add Incidence and daily deaths into the output
• [ ] Include birth
• [ ] Include natural deaths
• [ ] Thought: Rt calculation throughout the simulation period?

Format (goal - unify model structure across classes; generalize common functions / class definitions):

• [x] Define a general plot_dataframe function outside of the model class (can be applied to any long-format dataframe)
• [x] Output dataframe in long format (as we did in plot function)
• [x] Restructure the model (rename slots, instances and params)
• [x] Moving R0 and output incidences out from the class. The corresponding methods R0() and run() simply return values rather than updating the class.
• [ ] Thought: Move class declarations and related function definitions to objects.R
• [ ] Thought: Move all generic into a single .R file
• [ ] Thought: class inheritance

Others

• [ ] RMD check fails when including plotly and kableExtra in package and vigenettes. -- WHY? Parameter namings - reference list

[hfldai]

Do we want to calculate number of cases and deaths in each model, or include it into the generic plotting function?

I think currently we calculate cumulative disease-caused deaths in each model separately, as each model consists of different infectious groups. The plotting function is defined outside of the model class, which can be called by different models, given they have the same format of output dataframe.

It's interesting to me that waning immunity is not considered when calculating the basic reproduction number. The course of the epidemic is so different with and without waning immunity, yet their reproduction numbers are the same.

Yes, whether or not including waning immunity gives different epidemic curves even though at the same R0, as R0 is the basic reproduction number that only implies the trend at the beginning of the epidemic, under the disease free equilibrium (DFE) condition (S0>0, E0>0, S0>>E0, other compartments = 0); thus R0 does not imply how the pandemic behaves in the later times. I think If we monitor the effective reproduction number R(t) for each time point, we might see the differences caused by waning immunity (and maybe also the effects by other conditions/parameters).