Adding sklearn min. dependencies for all versions

[Neeratyoy]

Reference Issue

Addresses openml/openml-sklearn#1.

What does this PR implement/fix? Explain your changes.

Allows exact minimum dependency of packages to be listed in a sklearn flow created under the dependencies attribute.

How should this PR be tested?

Doesn't break any existing unit tests.

Any other comments?

Need to assess how existing flows with different dependency string is affected.

[codecov-io]

Codecov Report

Merging #1022 (463f8dd) into develop (80ae046) will decrease coverage by 0.20%.
The diff coverage is 80.00%.

@@ Coverage Diff @@## develop openml/openml-python#1022 +/- ##
===========================================
- Coverage 88.23% 88.02% -0.21% 
===========================================
Files 36 36 Lines 4571 4577 +6 ===========================================
- Hits 4033 4029 -4 - Misses 538 548 +10 

Impacted Files	Coverage Δ
openml/extensions/sklearn/extension.py	90.71% <80.00%> (-0.14%)	⬇️
openml/config.py	79.67% <0.00%> (-4.42%)	⬇️
openml/testing.py	83.53% <0.00%> (-1.72%)	⬇️
openml/datasets/functions.py	94.42% <0.00%> (-0.04%)	⬇️
openml/utils.py	91.33% <0.00%> (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 80ae046...8442576. Read the comment docs.