Add option to return indices in RandomOverSampler

[hgascon]

As in RandomUnderSampler, this commit adds the option to return the indices used for over sampling in the corresponding classes. This can be useful, for example, if the same sampling is to be used again in a different data structure with the original shape and class distribution.

[pep8speaks]

Hello @hgascon! Thanks for updating the PR.

Cheers ! There are no PEP8 issues in this Pull Request. 🍻

Comment last updated on August 22, 2018 at 20:31 Hours UTC

[codecov]

Codecov Report

Merging #439 into master will decrease coverage by 0.08%.
The diff coverage is 80%.

@@            Coverage Diff             @@
##           master     #439      +/-   ##
==========================================
- Coverage   98.77%   98.68%   -0.09%     
==========================================
  Files          75       70       -5     
  Lines        4410     4191     -219     
==========================================
- Hits         4356     4136     -220     
- Misses         54       55       +1

Impacted Files	Coverage Δ
imblearn/over_sampling/random_over_sampler.py	95.83% <80%> (-4.17%)	⬇️
imblearn/over_sampling/smote.py	91.82% <0%> (-2.44%)	⬇️
imblearn/utils/estimator_checks.py	96.69% <0%> (-0.04%)	⬇️
imblearn/over_sampling/__init__.py	100% <0%> (ø)	⬆️
imblearn/over_sampling/tests/test_smote.py	100% <0%> (ø)	⬆️
imblearn/keras/_generator.py
imblearn/tensorflow/_generator.py
imblearn/tensorflow/__init__.py
imblearn/keras/tests/test_generator.py
imblearn/keras/__init__.py

---

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 7c48491...6244ccf. Read the comment docs.

[glemaitre]

Could you add an additional test to have full coverage. You should check the output of the samples indices returned.

Please add an entry to the change log at doc/whats_new/v0.**.rst under enhancement. Like the other entries there, please reference this pull request with :issue: and credit yourself (and other contributors if applicable) with :user:

[glemaitre]

I finished up the PR. Thanks @hgascon