Major refactor based on Julia Discourse feedback

[ablaom]

This PR is a major refactor, based on the feedback at this Julia Discourse thread as of November 2023.

There are really too many changes to itemize and I recommend interested parties peruse the new documentation, which is less bloated and substantially rationalised. These changes will not be considered "final", even after merge.

The main change is a move away from methods with return multiple values (such as separate report or state) and towards the design suggested by @CameronBieganek. A basic user workflow looks like this:

X = <some training features>
y = <some training target>
Xnew = <some test or production features>

# Train:
model = fit(forest, X, y)

# Predict probability distributions:
predict(model, Distribution(), Xnew)

# Generate point predictions:
ŷ = predict(model, LiteralTarget(), Xnew) # or `predict(model, Xnew)`

# Apply an "accessor function" to inspect byproducts of training:
LearnAPI.feature_importances(model)

# Slim down and otherwise prepare model for serialization:
small_model = minimize(model)
serialize("my_random_forest.jls", small_model)

# Recover saved model and algorithm configuration:
recovered_model = deserialize("my_random_forest.jls")
@assert LearnAPI.algorithm(recovered_model) == forest
@assert predict(recovered_model, LiteralTarget(), Xnew) == ŷ

I am also handling the previously "optional data interface" in a different way. See the obs method referred to in the docs.

I have not renamed algorithms to strategies in this PR. This was simply going to be too much cognitive dissonance

For later PR's

• refactor "update", "ingest" apparatus (just dumped them for now).
• address Accounting for data objects that only iterate  #15
• Address some other outstanding issues
• Too many traits about input/output types/scitype. Dump some of these?

[codecov]

Codecov Report

Attention: 15 lines in your changes are missing coverage. Please review.

Comparison is base (7d9dae0) 55.29% compared to head (71c57f4) 64.70%.

Files	Patch %	Lines
src/traits.jl	53.33%	14 Missing ⚠️
src/minimize.jl	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##              dev      #25      +/-   ##
==========================================
+ Coverage   55.29%   64.70%   +9.41%     
==========================================
  Files           6        9       +3     
  Lines          85      119      +34     
==========================================
+ Hits           47       77      +30     
- Misses         38       42       +4     

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