Causalnex

* Remove Boston housing dataset from "sklearn tutorial", see 91 for more information.
* Update pylint version to 2.7
* Improve speed and non-stochasticity of tests

* Fixed bug where the sklearn tutorial documentation wasn't rendering.
* Weaken pandas requirements to >=1.0, <2.0 (was ~=1.1).

* Removed Python 3.5 support and add Python 3.8 support.
* Updated core dependencies, supporting pandas 1.1, networkx 2.5, pgmpy 0.1.12.
* Added PyTorch to requirements (i.e. not optional anymore).
* Allows sklearn imports via `from causalnex.structure import DAGRegressor, DAGClassifier`.
* Added multiclass support to pytorch sklearn wrapper.
* Added multi-parameter collapsed graph as graph attribute.
* Added poisson regression support to sklearn wrapper.
* Added distribution support for structure learning:
* Added ordinal distributed data support for pytorch NOTEARS.
* Added categorical distributed data support for pytorch NOTEARS.
* Added poisson distributed data support for pytorch NOTEARS.
* Added dist type schema tutorial to docs.
* Updated sklearn tutorial in docs to show new features.
* Added constructive ImportError for pygraphviz.
* Added matplotlib and ipython display convenience functions.

* Added `DAGClassifier` sklearn interface using the Pytorch NOTEARS implementation. Supports binary classification.
* Added binary distributed data support for pytorch NOTEARS.
* Added a "distribution type" schema system for pytorch NOTEARS (`pytorch.dist_type`).
* Rename "data type" to "distribution type" in internal language.
* Fixed uniform discretiser (`Discretiser(method='uniform')`) where all bins have identical widths.
* Fixed and updated sklearn tutorial in docs.

* Added DYNOTEARS (`from_numpy_dynamic`, an algorithm for structure learning on Dynamic Bayesian Networks).
* Added Pytorch implementation for NOTEARS MLP (`pytorch.from_numpy`) which is much faster and allows nonlinear modelling.
* Added `DAGRegressor` sklearn interface using the Pytorch NOTEARS implementation.
* Added non-linear data generators for multiple data types.
* Added a count data type to the data generator using a zero-inflated Poisson.
* Set bounds/max class imbalance for binary features for the data generators.
* Bugfix to resolve issue when applying NOTEARS on data containing NaN.
* Bugfix for data_gen system. Fixes issues with root node initialization.

* Added plotting tutorial to the documentation
* Updated `viz.draw` syntax in tutorial notebooks
* Bugfix on notears lasso (`from_numpy_lasso` and `from_pandas_lasso`) where the non-negativity constraint was not being set
* Added DAG-based synthetic data generator for mixed types (binary, categorical, continuous) using a linear SEM approach.
* Unpinned some requirements