Sabcom

This version has all elements for the paper on learning

Requirements are now automatically installed when using setup.py

new version of the model on pypi with the option to simulate a second wave with 3 new optional arguments for simulate:

-sec 1000 -sect 167 -scsi random

where -sec means the number of newly infected agents during the second wave, -sect is the period of time at which this new infection happens, -scsi is the scenario which can be either initial (where the new infections happen in those districts where the initial infections happened (e.g. CBD), and random which is a scenario in which new infections occur randomly across the districts.

added sabcom estimate & sabcom sample functionality

- added period_to_become_infected to agent object
- used that to make 1 agent loop in runner. Infecting other agents now works via period_to_become_infected variable to make sure the state of agents does not change within the loop.
- changed truncated normal shocks so that it now draws a large sample only once. This makes the model much faster.