uofuepibio / epiworld Goto Github PK

The following lack of implementation:

• Model.rm_entity
• Model.add_entity_n
• Model.add_entity_f

A contradiction can occur when a global action changes an agent to a status that depends on a virus, e.g.:

1. Update status moves agent to recovered, and the virus is to be removed.
2. The global action sees the agent still in the model, and moves the agent to infectious.
3. The virus will be removed, but the status is set as the last one, infectious.
4. The agent will then be incorporated into the model as infectious, but with no virus, causing an error.

The rationale to compute transmission should be different.

1. Using contact_rate, sample neighbors from the population.
2. Follow the same steps as in the ModelSIR class, i.e., regular transmission prob.

This alternative should be used whenever any of the follow conditions is true:

• Transmission probability in the virus is a function.
• There are tools involved.

The Person::update_status function uses the IN template function which right now is causing a bit of a bottleneck. Instead of checking set inclusion, we could leverage that statuses are unsigned ints, and thus an updating function could be reached by simply accessing by index, i.e., update_status_funs[p->status_current]();.

Both lack of a print method.

Currently, when reset()=true and there's a backup population, the agents' states are not reset. This should not be the case, as the backup should only preserve network structure (and data), not states, tools, and viruses.