Update JHub compatibility note

Fixes #334

README.md: added link on how to create the custom password hashed

update scikit-learn to latest release (v0.18*)

Enhancement

Upgrade to r-base 3.3.2

Pin spylon-kernel version

Add spylon-kernel to all-spark

Add issue template

Remove pinning for plotting fix and upgrade other R packages

debian, miniconda, notebook version, option updates

* Upgrade to latest debian base image
* Upgrade to Notebook 4.3
* Upgrade to Miniconda 4.2.12
* Remove USE_HTTPS env var in favor of command line options for key and cert
* Add GEN_CERT env var for generating a self-signed certificate
* Remove PASSWORD env var in favor of the new Notebook 4.3 default token auth
  or the more secure a hashed password command line option

Install Toree to conda env instead of user home

TensorFlow 0.11.0

Upgrade Spark to 2.0.2

Fixes issues with python version mismatches when using Spark

Dockerfile version for power ppc64le

adding a patch file to make Dockerfile.ppc64le more maintainable and updating the Makefile to make use of the patch file

JupyterHub 0.7

and get it from conda (conda-forge)

single-user protocol should still be compatible, so this shouldn't change supported Hub versions

Add vincent, beautifulsoup, and xlrd to scipy-notebook