AbstractDynodeRunner code reorg and naming revamp

[arik-shurygin]

this PR does the following:

• save_inference_timeseries now takes particle tuples as a list instead of randomly selecting, allowing for better user control over its behavior. The method will also return the saved dataframe instead of a str of where it was saved to. If user passes None to save_filename the csv will not be saved at all, in case that is required.
• moves some functions hidden within the AbstractDynodeRunner out into utils so they may be more easily referenced by other code, as well as declutter the class itself.
• Rename all references of timelines to timeseries when referring to the saving of metrics like vaccination or infection counts of a simulation.
• Renaming the azure_visualizer_timelines.csv auto-generated file to simulation_timeseries.csv so it is easier to understand what it contains. Also the csv is used in much more than just the shiny app visualizers...

CLOSES #311
while this PR closes #311 some lasting software design decisions have made the request for parallelized loading of posterior particles not possible without substantial changes. I will create an issue #315 to track that