Update AutoSearchBase to use dynamically generated preprocessing pipelines

[angela97lin]

Closes #844

Notes:
self.allowed_pipelines isn't instantiated until search() if not supplied as a parameter, is that weird? (same with self.allowed_model_families)

[codecov]

Codecov Report

Merging #870 into master will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff            @@
##           master     #870    +/-   ##
========================================
  Coverage   99.76%   99.77%            
========================================
  Files         193      193            
  Lines        8603     8830   +227     
========================================
+ Hits         8583     8810   +227     
  Misses         20       20            

Impacted Files	Coverage Δ
...lml/automl/automl_algorithm/iterative_algorithm.py	100.00% <100.00%> (ø)
evalml/automl/automl_search.py	98.36% <100.00%> (+0.04%)	⬆️
evalml/pipelines/utils.py	100.00% <100.00%> (ø)
evalml/tests/automl_tests/test_automl.py	100.00% <100.00%> (ø)
.../automl_tests/test_automl_search_classification.py	100.00% <100.00%> (ø)
...ests/automl_tests/test_automl_search_regression.py	100.00% <100.00%> (ø)
evalml/tests/conftest.py	100.00% <100.00%> (ø)
evalml/tests/pipeline_tests/test_pipelines.py	99.81% <100.00%> (+<0.01%)	⬆️

---

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 1e2fcb5...eab1100. Read the comment docs.

[kmax12]

i tested on a local dataset and it work and did a quick pass through the code and left some comments. has my approval, but id get one more.

[dsherry]

@angela97lin I didn't review the tests yet, will do. But I left comments on the impl. Looks good! Mostly minor stuff. Ping me whenever you feel this is ready for re-review and I'll read the whole thing.

[dsherry]

Looks good! Left suggestions on splitting up the tests, parametrize, cleaning up dead code in the impl and logging. I'd say the only blocking thing is deleting the dead code, but would be great to get the other stuff in too, either in this PR or separate

[angela97lin]

@dsherry FYI I split the tests up as you suggested, and parameterized some but not all, my reason being that it got really unwieldy to add in all of the different parameters (dataset, mocks, mock pipelines) to support binary/multiclass and this seemed cleaner. We can always revisit though!