Factor out experiments

[epeters3]

• Factored out the code for each separate experiment (ensemble, straight, metafeatures, etc.) from experimenter.py into its own module.
• Decoupled the Experimenter class from needing to be passed command line arguments, to avoid the need for dependency injection when testing.
• Renamed a couple methods so the experiments would all fit the requirements of the Experiment abstract parent class.
• Added stub class for the stacked and random pipeline experiments.

[codecov-io]

Codecov Report

Merging #60 into develop will increase coverage by 5.56%. The diff coverage is 77.46%.

@@             Coverage Diff             @@
##           develop      #60      +/-   ##
===========================================
+ Coverage    65.38%   70.95%   +5.56%     
===========================================
  Files           20       26       +6     
  Lines         1355     1422      +67     
===========================================
+ Hits           886     1009     +123     
+ Misses         469      413      -56

Impacted Files	Coverage Δ
test/test_pipeline_generator.py	100% <ø> (+46.42%)	:arrow_up:
experimenter/pipeline_builder.py	96.05% <100%> (+2.3%)	:arrow_up:
test/test_ensembling_pipelines_generation.py	97.01% <100%> (+0.09%)	:arrow_up:
test/test_ensembling_pipeline_runs.py	98.11% <100%> (+0.03%)	:arrow_up:
experimenter/experiments/metafeatures.py	47.61% <47.61%> (ø)
experimenter/experimenter.py	59.61% <63.15%> (+1.76%)	:arrow_up:
experimenter/experiments/straight.py	68.51% <68.51%> (ø)
experimenter/experiments/stacked.py	75% <75%> (ø)
experimenter/experiments/random.py	75% <75%> (ø)
experimenter/experiments/experiment.py	80% <80%> (ø)
... and 11 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update fd8f64b...daf52ca. Read the comment docs.