_find_neighbor_and_lambda off-by-one error

[ainkov]

Hi there, I believe there is an off-by-one error in taking the point to predict core distance.
Currently it's taking the 6th neighbor (if min_samples is 5) as the core distance.
It's what solved my issue in #381 , although you are the best to judge whether it's logically correct depending on the rest of the implementation.
Thanks and regards.

[pep8speaks]

Hello @ainkov, Thank you for submitting the Pull Request !

• In the file hdbscan/prediction.py:

Line 139:21: W503 line break before binary operator
Line 247:22: W503 line break before binary operator
Line 308:44: W503 line break before binary operator
Line 310:56: W503 line break before binary operator