Add dictionary support for oversamplers

[bchen1116]

fix #2142

[codecov]

Codecov Report

Merging #2288 (f6ca020) into main (3faffc8) will increase coverage by 0.1%.
The diff coverage is 100.0%.

@@           Coverage Diff           @@
##            main   #2288     +/-   ##
=======================================
+ Coverage   99.9%   99.9%   +0.1%     
=======================================
  Files        280     280             
  Lines      24360   24490    +130     
=======================================
+ Hits       24333   24463    +130     
  Misses        27      27             

Impacted Files	Coverage Δ
evalml/automl/automl_search.py	100.0% <100.0%> (ø)
...s/components/transformers/samplers/base_sampler.py	100.0% <100.0%> (ø)
.../automl_tests/test_automl_search_classification.py	100.0% <100.0%> (ø)
evalml/tests/component_tests/test_components.py	100.0% <100.0%> (ø)
evalml/tests/component_tests/test_oversamplers.py	100.0% <100.0%> (ø)

---

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 3faffc8...f6ca020. Read the comment docs.

[bchen1116]

Decided to default to the sampler_balanced_ratio value that was provided for AutoMLSearch to set the sampling_ratio, which means it overrides the sampling_ratio that a user might input to the pipeline_params via

pipeline_parameters = {sampler : {"sampling_ratio": 0.5}}

Updated the docstring above to mention this

[ParthivNaresh]

Very cool! I don't think you need an elif here, it can just be an else since I don't think we'd have another logic path but total nitpick.

Also I like the way you set the logic for the elif block. You could do the same to the if block and move setting the self._frozen_pipeline_parameters after:

if self._sampler_name not in parameters:
     parameters[self._sampler_name] = {"sampling_ratio": self.sampler_balanced_ratio}
else:
     parameters[self._sampler_name].update({"sampling_ratio": self.sampler_balanced_ratio})
self._frozen_pipeline_parameters[self._sampler_name] = parameters[self._sampler_name]

[chukarsten]

Nice Bryan, this is good work. I think you and I discussed a bunch of things in slack parallel to this and I like the addition of docs and the test for the string keys. I still personally think the binary case with the key/val in the dict for the majority class in the oversampler case or the minority class in the undersampler case is a little confusing that they don't do anything. Personally, I just struggle with the dictionary representation, but I think that's a personal problem X-D.

[chukarsten]

Cool, I like this test. Very readable.

[ParthivNaresh]

Great job with this, excellent test coverage

[ParthivNaresh]

Very cool! I don't think you need an elif here, it can just be an else since I don't think we'd have another logic path but total nitpick.

Also I like the way you set the logic for the elif block. You could do the same to the if block and move setting the self._frozen_pipeline_parameters after:

if self._sampler_name not in parameters:
     parameters[self._sampler_name] = {"sampling_ratio": self.sampler_balanced_ratio}
else:
     parameters[self._sampler_name].update({"sampling_ratio": self.sampler_balanced_ratio})
self._frozen_pipeline_parameters[self._sampler_name] = parameters[self._sampler_name]