Option to return dataframes for listing functions

[Neeratyoy]

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

Addresses #364.
An optional parameter has been added to the get_X functions called output_format. Which is by default dict so that nothing breaks. When output_format='dataframe' the output is a pandas data frame.

How should this PR be tested?

The following commands should suffice in demonstrating the change:
openml.datasets.list_datasets(size=10)
openml.datasets.list_datasets(size=10, output_format='dataframe')
openml.tasks.list_tasks(task_type_id=1, size = 10)
openml.tasks.list_tasks(task_type_id=1, size = 10, output_format='dataframe')
openml.evaluations.list_evaluations('predictive_accuracy', size = 10)
openml.evaluations.list_evaluations('predictive_accuracy', size = 10, output_format='object')
openml.evaluations.list_evaluations('predictive_accuracy', size = 10, output_format='dataframe')
openml.flows.list_flows(size=10)
openml.flows.list_flows(size=10, output_format='dataframe')
openml.runs.list_runs(size=10)
openml.runs.list_runs(size=10, output_format='dataframe')
openml.setups.list_setups(size=10)
openml.setups.list_setups(size=10, output_format='object')
openml.setups.list_setups(size=10, output_format='dataframe')

Any other comments?

• For design parity, all the __list_X functions have been edited to produce the desired output format. In effect, every list function has an added parameter called output_format including the _list_all in utils.py.
• The defaults have been kept as 'dict' and 'object' where applicable to ensure nothing breaks.
• Examples which had listing as an example, has been updated.
• A small unit test is added for all the listing calls that already being tested

[codecov-io]

Codecov Report

Merging #662 into develop will decrease coverage by 0.15%.
The diff coverage is 89.38%.

@@ Coverage Diff @@## develop #662 +/- ##
===========================================
- Coverage 90.5% 90.35% -0.16% 
===========================================
Files 36 36 Lines 3730 3785 +55 ===========================================
+ Hits 3376 3420 +44 - Misses 354 365 +11

Impacted Files	Coverage Δ
openml/utils.py	91.66% <100%> (+0.39%)	⬆️
openml/study/functions.py	88.23% <66.66%> (-1.69%)	⬇️
openml/evaluations/functions.py	90.74% <82.35%> (-4.82%)	⬇️
openml/datasets/functions.py	95.66% <90%> (-0.28%)	⬇️
openml/runs/functions.py	81.48% <90.9%> (-0.03%)	⬇️
openml/tasks/functions.py	84.93% <90.9%> (-0.16%)	⬇️
openml/flows/functions.py	87.78% <90.9%> (-0.31%)	⬇️
openml/setups/functions.py	93.84% <96.87%> (-0.18%)	⬇️

---

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 72f131a...0a947af. Read the comment docs.