Maintain pipeline threshold when returning searched pipelines

[eccabay]

Closes #2844

I went with @freddyaboulton's first suggestion on how to fix the underlying problem, since it seemed like the smoothest way to maintain logical behavior under the hood.

[codecov]

Codecov Report

Merging #2948 (0e9188d) into main (910fbd0) will increase coverage by 0.1%.
The diff coverage is 100.0%.

@@           Coverage Diff           @@
##            main   #2948     +/-   ##
=======================================
+ Coverage   99.7%   99.7%   +0.1%     
=======================================
  Files        307     307             
  Lines      29197   29215     +18     
=======================================
+ Hits       29106   29124     +18     
  Misses        91      91             

Impacted Files	Coverage Δ
evalml/automl/automl_search.py	99.9% <100.0%> (+0.1%)	⬆️
evalml/pipelines/pipeline_base.py	98.4% <100.0%> (+0.1%)	⬆️
.../automl_tests/test_automl_search_classification.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 910fbd0...0e9188d. Read the comment docs.

[chukarsten]

This looks gucci to me. I am curious if we need to necessarily switch off of whether the pipeline's problem_type is binary to set the threshold or whether the threshold can always just be set to the cloning target's threshold. The testing looks solid and seems to check the boxes for what we set out to do!

[eccabay]

@chukarsten Thanks for the review! To answer your curiosity about whether the threshold can always just be set to the cloning target's threshold or not, it cannot. Only binary pipelines even have the threshold attribute, so this code throws AttributeErrors on non-binary pipelines without the is_binary switch.