* Add sampling support for AutoGL, including nodewise, layerwise and subgraph sampling.
* Refactor part of the model and trainer.

Merge from refactor_test branch and already fixed bugs detected through tests for refactor, and temporarily override  method of dataset provided by PyTorch-Geometric to maintain compabillity with PyTorch-Geometric>=1.7.0

add link prediction into documents

change model function name decode to lp_decode in response to PR #31

* add link prediction trainer, solver, and model adaptation

Random search

Modify the GCN model to support diverse dropout probabilities in different GCN layers.
Improve the sampling speed for BasicTargetDependentSampler base class.
Apply GC before training for every instance of trainer.

Fix bugs of LADIES, caused by carelessness.
Improve the Performance of LayerWiseDependentImpotanceSampler by utilizing scipy.sparse.