Sabcom

- removed unused agent parameters for probability to switch compartments
- removed excess files
- streamlined repo structure

- change measurement folder name to output_data
- remove excess parameters including lockdown days
- inefficient lockdown is now the baseline scenario that runs if you run the parameter json

make it possible for policies change over time during the simulation

- made probabilities to transition between compartments change in the runner rather than the environment
- added 100k agent notebook to run from pickle

- updated calibration of policy parameters
- added parameter estimate for transmission probability (thanks Tina!)
- removed parameter for connections to delete under self-isolation because could not be calibrated. Self-isolation now means you will not see anyone else then your household members.

Has a new form of learning