Ansible scripts do not reflect the most efficient way to use the system

[mearlboro]

Motivation:
the Ansible scripts do not clearly separate between setup (happening once) and deployment (happening more and more frequently, especially during experiments)
moreover, the scripts for running experiments are not self-contained: the user still needs to run parts of other scripts (checking the latest code is copied off, synchronising clocks, manually starting the observer server) before running the experiment.yml Ansible playbook

Solution:
split the Ansible playbooks between deployment, updates, testing, and experiment use, so that

• only one script needs to be run for initial installation and setup; it should include tasks also for fetching the status of various components (e.g. date, status of external devices, importing python packages) to test the setup was successful
• a script is dedicated to efficiently update code on already setup observer and players - without attempting to install software that is already there, wasting time
• merge in the dev-test branch, and create Ansible script to stress-test the hats and observer (for temperature, battery etc)
• only one script needs to be run to start an experiment, or make the hats do something on the spot (separate from testing per se!)