Master

Housekeeping:

• More informative comments added, throughout
• Updated on code legibility and style

Bugfixes:

• An issue during evaluation, where the metrics dict key lap_completed was not being properly used
• An issue during evaluation, where the metrics was not being properly initialised
• An issue during evaluation, where a metric was not being calculated correctly
• An issue during training and evaluation, where agents were sometimes not spawned correctly, after failing a track segment