Refactor auth mechanism in sync-experiment

[genematx]

Context: Some beamlines (e.g. SMI) collect data on multiple proposals during the same run to make use of the same environment conditions between the experiments (e.g. temperature); changing these conditions often would unnecessarily require too much downtime for the instruments.
Our current sync-experiment workflow assumes that a user (or a beamline scientist) switches to a new proposal at the beginning of acquisition (possibly even before opening bsui). This would be impractical for highly dynamic or automated plans as above.

This PR Separates the authentication mechanism in the internal sync_experiment function from the updating of the redis dictionary. This would allow us to switch between multiple proposals within a profile_collection plan (using the new switch_redis_proposal(proposal_number, beamline, username, prefix) function -- SMI already have something very similar that they extensively use in their profile collection, and the idea is to just replace it). This function will still check if a user has the rights to collect data on a proposal, but wouldn't ask for their password. We will need that before running bsui or in the base step of profile collection, the user properly authenticates with sync-experiment (this is effectively an honor system, but writing the data in a correct location is is in their interests, because otherwise they won't be able to read it).

These refactoring changes will not break any existing deployments.