Refactor, polish and document gama-server and its API

[lesquoyb]

Is your request related to a problem? Please describe.
Currently gama-server GUI and gama-server headless are not fully compatible, a client working for one could not be completely working for the other. In addition to this, some parts are just proof of concepts that lack some polish, and are not documented yet.

Describe the improvement you'd like
Here are the problems to tackle:

• in headless we need to use exp_id but the gui only returns the experiment name
• if using exp_id in gui, gama will try to resolve it even though it is meant to work only with the current experiment
• the messages of type simulation status are only sent in the case of gui use
• The use of GS in GUI is undocumented (see this issue)
• The error message RuntimeError is not used anymore
• There's currently no documentation on simulation status, what should be documented is (see this issue):
  • the list of all the different status and their meaning
  • the general explanation of how they work: how to subscribe to them? Can multiple clients get the same messages? What format are they following
• being able to unsubscribe from status messages as described in a previous issue([GS] Being able to subscribe/unsubscribe to simulation status & asking for a simulation status message #15)
• being able to query for the current simulation status as described in this comment
• The compilation errors message is vague and should contain more information ([GS] Add information to compilation errors sent back from gama-server #172)
• in gui mode the server options are not taken into account, every data is sent to the client regardless of the options set in the load function call (runtime, console, inform etc.)

Additional context
This is this repository's first issue 🥳