Docstrings for all epimodel types and generate_latent_infs plus named arguments for constructors

[SamuelBrand1]

This is a draft PR for on going work on documenting AbstractEpiModel subtypes and methods that dispatch on them.

Current focus and decisions:

• As recommended here I am focusing documentation at the highest/most abstract level. This means that I'm going to give the example usages and most extensive documentation to generate_latent_infs dispatch on generic AbstractEpiModel.
• I don't think the METHODLIST was useful in our doc template and has been removed.
• Example usage of a DirectInfections goes beyond just constructing the struct, but also its main usage with generate_latent_infs.
• Example usage is not a doc test; it has a complicated output that works ok with a spread of testitems. However, a next step could be to combine as per discussed in Doctests with @testitem #48. I've readvised this... simple doctests will be included.

This is aimed at addressing #137 which is part of #35 and #97 . It sets my current expectation of docstrings #135 and #126 .

Extra activity

• When it is absolutely straightforward to create a default constructor with named arguments using @kwdef I am doing it, because this also makes example usage easier. This moves towards addressing Make all constructors have named and not positional arguments. #133 , but won't close this issue.
• Advantage of above is that default priors are added directly aiming to address Revisit handling of default priors #131 .

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.61%. Comparing base (649567f) to head (cae1144).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #138   +/-   ##
=======================================
  Coverage   97.61%   97.61%           
=======================================
  Files          13       13           
  Lines         168      168           
=======================================
  Hits          164      164           
  Misses          4        4           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[SamuelBrand1]

I've done everything in the opening description except for the example usage of the EpiModels and generate_latent_infs is not a doctest.

The reason for this is that the pattern of doc testing against an expected result didn't seem to meet what we want from the examples and might lead into complexity around different random seed implementation on different OS etc. An alternative approach which just checks the examples run didn't seem to do anything that our testing CI doesn't do anyway.

Happy to get pushed back on this.

[seabbs]

This is nice. It doesn't look like you have implemented field or return information for functions or made use of DocStringExtensions. I think we need at least a clear return field but perhaps in another issue.