[MRG] EHN: support SparseDataFrame when creating a dataset

[glemaitre]

Add support for SparseDataFrame when creating a matrix.

[glemaitre]

@amueller@mfeurer@janvanrijn

Here come the support for sparse dataframe during dataset creation.

[glemaitre]

ping @amueller@mfeurer@janvanrijn any feedback?

[janvanrijn]

Personally, I am not super familiar with sparse dataframes. I can have a look at your test-cases if you'd like and see whether they cover interesting test-cases.

[mfeurer]

This looks good to me. Could you add this to the example, too?

[codecov-io]

Codecov Report

❗ No coverage uploaded for pull request base (develop@c69b0a6). Click here to learn what that means.
The diff coverage is 100%.

@@ Coverage Diff @@## develop #583 +/- ##
==========================================
Coverage ? 89.94% ==========================================
Files ? 32 Lines ? 2995 Branches ? 0 ==========================================
Hits ? 2694 Misses ? 301 Partials ? 0

Impacted Files	Coverage Δ
openml/datasets/functions.py	92% <100%> (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 c69b0a6...a9a7e2d. Read the comment docs.

[glemaitre]

@mfeurer I added a new section in the tutorial

[mfeurer]

Thanks a lot for adding the section to the documentation. I think this can be merged if

1. you resolve the conflict
2. you fix the flake8 error

[glemaitre]

Done!