New simulation model

[GernotMaier]

This is a very large pull request with several major changes:

• changing names of telescopes to official CTAO naming (e.g., MSTN-01). This required a re-writing of the model database and many changes in the code
  • open point: update all docstrings and documentation for new telescope naming
• new organisation of model parameters with a base model_parameter class and to specialized classes called telescope_model and site_model (allows to add in future e.g., illuminator_model classes)
• reading of simulation model parameters from json files (either from file from from repository)
  • directory / repository structure outlined in simulation_model README
• all site and telescope-related parameters are now in the database (or in the simulation model repository)
  • this affects especially the data in the data/parameters/corsika_parameters.yml file and the auxiliary data in the metadata headers of the array element / telescope lists.
• additional changes for more consistent naming (e.g., in db instead of units use unit (as in astropy)

Given that telescope naming has changed, this PR required a lot of name-but-not-functional changes in code, tests, and data files used.

Temporary items to be fixed best in a future PR:

• simulation_model is expected to be on GitHub (planned to be moved to CTAO gitlab)
• CIs for unit- and integration tests required the cloning of the simulation model (see here and here). This should go when all parameters are in the simulation model database
• old/new names are currently duplicated in names.py (see here. This also requires changes in the database.
• missing documentation updates (this is probably a more general issue)

Suggestion for review:

• a lot of files changed.
  • suggest first to do the data files in tests/resources: changes are in telescope names (e.g., from LST-1 to LSTN-01) and removal of corsika-related parameters from the metadata header (now included in db_handler)
  • data/parameters/corsika_parameters.yml : all site/telescope related parameters are no in the DB
• simple changes: db_handler is now in db/db_handler (as there are more db-related modules). This obviously changes all imports on db_handler
• for actual code review: best maybe to start with names.py, then db_handler and then the model parameter classes (model_parameter, telescope_model, site_model).
• after that, look at the rest of the modules and tests (lots of little and annoying changes)

[orelgueta]

Very nice!

I made just a few comments and only one of them worries me a bit (about the Structure/Camera readout from the DB).

[orelgueta]

I don't think I see it fixed (doesn't start with _)

[orelgueta]

Why do you need this at all? if you call the function from the base class without any changes anyway, you can just remove it and you will get the function from the base class

[orelgueta]

Since to write these dictionaries you looked these up, perhaps you can add comments on each of this to say what it is cause I definitely cannot tell from the short name. But only if you can do it off the top of your head, no need to put effort into it.

[GernotMaier]

You mean what ILLN North means? I think I remember.

I anyway plan to remove all those lines and move them into the json schema (some parts are already duplicated, e.g., the list of telescopes) - names.py will read then the schema files to get the naming dicts.

[orelgueta]

Yes, ILLN, RLDN, STPN, etc. I don't remember what all of them means and I think it would be useful to have the non-abbreviated names as well. Also if it's in the schema files it would be useful.

[GernotMaier]

Github does not allow me to reply on that: 'fixed'.

[GernotMaier]

Propagated this to issue #828 (and again I cannot answer to this comment)

[GernotMaier]

I don't know enough about python to answer this question. However, I find that this line increases readability of the code dramatically and would prefer to keep it.

[GernotMaier]

@orel - thanks for the review!! I've addressed all of your comments.

I will need your help on the db_handler:_get_model_parameters_mongo_db for the selection of the applicable parameters. What is wrong with the implementation now? Is it inefficient, as we get the full telescope model instead of camera or structure only? Or is there another issue which I don't see?

[orelgueta]

Changes look good. Let's discuss the thing about the camera/optics in a quick call instead, it would be quicker.

The only other point here is the super() function which you claim helps readability. I find that surprising cause I suggest to just remove that function completely, so there's nothing to read =)
Perhaps I am wrong though, I am not an expert on python inheritance.