Add NoSplit data splitter

[eccabay]

Closes #2956

First step of adding unsupervised learning to evalml, this code is lifted directly from my innovation days branch.

[codecov]

Codecov Report

Merging #2958 (1c6aa23) into main (ffb7334) will decrease coverage by 0.8%.
The diff coverage is 100.0%.

@@           Coverage Diff           @@
##            main   #2958     +/-   ##
=======================================
- Coverage   99.7%   99.0%   -0.7%     
=======================================
  Files        307     309      +2     
  Lines      29288   29312     +24     
=======================================
- Hits       29197   29003    -194     
- Misses        91     309    +218     

Impacted Files	Coverage Δ
evalml/preprocessing/__init__.py	100.0% <100.0%> (ø)
evalml/preprocessing/data_splitters/__init__.py	100.0% <100.0%> (ø)
evalml/preprocessing/data_splitters/no_split.py	100.0% <100.0%> (ø)
evalml/tests/preprocessing_tests/test_no_split.py	100.0% <100.0%> (ø)
evalml/automl/pipeline_search_plots.py	17.9% <0.0%> (-82.1%)	⬇️
...l/tests/automl_tests/test_pipeline_search_plots.py	25.6% <0.0%> (-74.4%)	⬇️
...ests/automl_tests/test_automl_search_regression.py	84.6% <0.0%> (-15.4%)	⬇️
.../automl_tests/test_automl_search_classification.py	89.7% <0.0%> (-10.3%)	⬇️
evalml/tests/automl_tests/test_automl_utils.py	90.6% <0.0%> (-9.4%)	⬇️
...lml/tests/automl_tests/test_iterative_algorithm.py	94.7% <0.0%> (-5.3%)	⬇️
... and 3 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 ffb7334...1c6aa23. Read the comment docs.

[angela97lin]

@eccabay Curious, is there a way to incorporate the idea of no splitting without introducing another splitter? That is, is there a way to change AutoMLSearch s.t. to take in a splitter, and if the splitter is None then we don't split?

[eccabay]

@angela97lin great question! It is possible, I just spent some time playing around with my innovation days branch and proving that it could work. It involves a little refactoring of engine_base.py::train_and_score_pipeline. I originally went with the data splitter object because it seemed simpler at the time, but if it makes more sense to do it the other way, I'm more than happy to update this PR!

[chukarsten]

The data splitter that doesn't split! What functionality! So brave. Wow. All joking aside, this looks like this does what you want to do! Definitely would be good to see where this fits into an unsupervised learning epic, but I certainly trust you. I agree with Freddy that returning an empty array better conveys what this splitter is doing. I think the get_n_splits() should also probably return 1!

[angela97lin]

@eccabay Gotcha, that makes sense. I think this is great, and love how it makes it simple to not split data within our current structure without much refactoring. If this us to unsupervised learning support more quickly, I'm all on board! We can always support not having a splitter and refactor afterwards 😁

[chukarsten]

Thanks for making the changes!