Added excluded_model_families parameter to AutoMLSearch()

[christopherbunn]

Resolves #4197

[codecov]

Codecov Report

Merging #4196 (16395bd) into main (1591ca7) will decrease coverage by 0.0%.
The diff coverage is 98.7%.

@@           Coverage Diff           @@
##            main   #4196     +/-   ##
=======================================
- Coverage   99.7%   99.7%   -0.0%     
=======================================
  Files        349     349             
  Lines      38159   38229     +70     
=======================================
+ Hits       38042   38111     +69     
- Misses       117     118      +1     

Impacted Files	Coverage Δ
...lml/automl/automl_algorithm/iterative_algorithm.py	97.5% <ø> (-<0.1%)	⬇️
...valml/automl/automl_algorithm/default_algorithm.py	99.7% <75.0%> (-0.3%)	⬇️
evalml/automl/automl_algorithm/automl_algorithm.py	100.0% <100.0%> (ø)
evalml/automl/automl_search.py	99.7% <100.0%> (+0.1%)	⬆️
evalml/pipelines/components/utils.py	96.5% <100.0%> (+0.2%)	⬆️
evalml/tests/automl_tests/test_automl.py	99.6% <100.0%> (+0.1%)	⬆️
...ts/automl_tests/test_automl_iterative_algorithm.py	100.0% <100.0%> (ø)
...valml/tests/automl_tests/test_default_algorithm.py	100.0% <100.0%> (ø)
...lml/tests/automl_tests/test_iterative_algorithm.py	100.0% <100.0%> (ø)
evalml/tests/pipeline_tests/test_pipeline_utils.py	99.6% <100.0%> (+0.1%)	⬆️

[eccabay]

@christopherbunn what are your thoughts on removing allowed_model_families as part of this work?

If we don't, it may be possible to introduce some weird conflicts using the iterative algorithm where a model family is simultaneously allowed and disallowed. It's been a weird, semi-deprecated argument for a while anyways, since it's only usable in the iterative algorithm case but we still allow users to set it in the default algorithm case. Open to discussion, of course.

[christopherbunn]

@eccabay re:allowed_model_families, I think we can keep it and add logic to make sure that both of these parameters are not set. I can see cases where I would want to have this set, even if we don't currently use it anywhere else.

[eccabay]

LGTM, just have some nits about our handling of the "not set" case

[jeremyliweishih]

LGTM just wondering if we should centralize the validation of allowing and excluding model families. LMK what you think!