add unit test with all identical dataset about cleanlab HOT 3 CLOSED

have a second dataset that is the same as this one, but has one extra non-different datapoint in it.

• verify scores are super similar for all the overlapping datapoints between the two datasets.
• verify the outlier score for this extra different datapoint is high and this data point is flagged

from cleanlab.

To fix the current outlier scores issue, we'll want to:

replace np.finfo(np.float_).eps everywhere in the package where it is being used as a divisor, eg:
https://github.com/cleanlab/cleanlab/blob/master/cleanlab/outlier.py#L463

with a much larger value such as 1e-8 or 100*np.finfo(np.float_).eps

Add an extra line in transform_distances_to_scores():

inds = np.close(avg_distances, 0.0)
ood_features_scores[inds] = 1.0

from cleanlab.

verify nothing analogous is needed for near duplicates

from cleanlab.