New search space def

[perib]

[please review the Contribution Guidelines prior to submitting your pull request. go ahead and delete this line if you've already reviewed said guidelines.]

What does this PR do?

fixes and tweaks several things related to the new search space API

1. graph and tree search spaces are now randomly initialized with random number of nodes.
2. added a function to flatten all search spaces into a graphpipeline.
3. implemented the new search space API into the steady-state estimator
4. edited the configuration spaces to include conditional hyperparameters. tweaked some search spaces. I changed some search spaces to match the search spaces in autosklearn.
5. added a function parameter to EstimatorNode that can parse parameter dictionaries and return a new dictionary to be used for the estimator.
6. implemented some search spaces using a wrapper search spaces. Perhaps we can include more meta-estimators this way? (Gradient boosting, etc)