Add method build_forest(model, ...) to add trees to existing model

[ablaom]

Closes #213.

You can now do this:

features, labels = load_data("iris")
model1 = build_forest(labels, features, 2, 150) # n_trees = 150
model = build_forest(model1, labels, features, 2, 50) # add 50 trees

New docstring below.

Details

build_forest(model, labels, features, options...; keyword_options...)

Return an updated version of model with additional n_trees=options[2] added to the
forest. Here options and keyword_options are as for the regular build_forest method,
which excludes the model argument.

Even if training data is the same in all build_forest calls, it is not practically
possible to guarantee adding trees in steps is identical to adding them all at once,
because of the way random number generators are generated and used. But in all other
respects these approaches are equivalent.

Example

features, labels = load_data("iris")
features = float.(features)
labels = string.(labels)

The call

model = build_forest(labels, features, 2, 200) # n_trees = 200

is approximately equivalent to

model1 = build_forest(labels, features, 2, 150) # n_trees = 150
model = build_forest(model1, labels, features, 2, 50) # n_trees = 50

[rikhuijzer]

Cool feature

[codecov-commenter]

Codecov Report

❌ Patch coverage is 96.15385% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 88.15%. Comparing base (c5df161) to head (a05e4cd).
⚠️ Report is 28 commits behind head on dev.

Files with missing lines	Patch %	Lines
src/classification/main.jl	90.90%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##              dev     #216      +/-   ##
==========================================
+ Coverage   87.99%   88.15%   +0.16%     
==========================================
  Files          10       10              
  Lines        1249     1266      +17     
==========================================
+ Hits         1099     1116      +17     
  Misses        150      150              

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[ablaom]

@rikhuijzer Could you please take a second look?

[rikhuijzer]

@rikhuijzer Could you please take a second look?

Sorry for delayed response. I was away from keyboard in the weekend.

Looks good to me. Thanks for incorporating the comments

[ablaom]

I was away from keyboard in the weekend.

Sounds healthy to me. I am often away from keyboard on weekend. It's just that my weekend finishes earlier than yours 🌐 🥝 😄

Thanks for the prompt review, @rikhuijzer