Add option of static networks (default are dynamic) (#205)

Static networks are built once at the start of the simulation and then
only update as necessary (i.e. on lockdown or a user adding networks)

If static networks are used, then quarantining is handled by checking
the quarantine status and household of the individual instead of
changing the network

Relative susceptible by interaction (#148)

* Bugfix - remove occupation network from params object once set up

If model is created multiple times from the same params object then we
get different answers. Problem with the Simulation Python interface, but
not the Model Python interface (so was missed in tests).

* Make releative_susceptibility_xxxx per interaction not per day

Remove the normalisation to the based on the mean number of interactions
to the adjusted susceptibility

* Update baseline parameters for asypmtomatic, mild and hospitalised

* Change equation to normalise by mean interactions

Not mean interactions by age

* Update doc with hew def of relative susceptibility

* Update realtive severity infectiousness data with Sun data

* Update the header of set_up_infectious_curves

Remove the text saying we adjust by mean interactions for the age group

* Update with calibrated infectious parameter and lockdown rate

* Increase number of cases in test_disease_transition_times

Test failing with a very small number of cases leading to inaccuracy

* Decrease the infectious rate to reduce saturation effect

test_relative_transmission_update is running in to saturation effect
with the updated params

Merge pull request #82 from p-robot/add_output_of_networks

Add functions for outputting household, random, and occupation networks