build JupyterHub singleuser images for each notebook stack

[ci skip] Add badges to stack READMEs

instead of master

and include checksummed download

to give it higher priority

Rebase all-spark-notebook on pyspark-notebook

To speed up build, reduce accidental disparity, ease maintenance

(c) Copyright IBM Corp. 2016

(c) Copyright IBM Corp. 2016

For every `*-notebook` stack, a `*-singleuser` image is also built

Switch to Mesos Jessie repository, and upgrade to 0.25

use current command for Python 2 kernel install

Add cacert to .curlrc in home directory

Specify `cacert=/etc/ssl/certs/ca-certificates.crt` in `/home/$NB_USER/.curlrc` so curl can locate the certificates.

Rebase datascience-notebook on scipy-notebook

README: Add badge for Travis CI

Fixes #139

(c) Copyright IBM Corp. 2016

Fixes #147

Add and install from the r channel with conda

Modified to install Apache Toree using pip

(c) Copyright IBM Corp. 2016

Clean apt-get list

Add dockerignores

Added instructions for connecting to spark on standalone mode.

Run `apt-get clean` after installing jq

Use Travis CI

This is done to more explicitly track what version of Debian Jessie is
being used as a base image. It will also ensure that it is properly
updated on the VM even if we forget. This also should help CI and VM
builds stay speedy by using the cache even when there is a newer version
of Debian Jessie. In the long run, we may wish to re-evaluate this
strategy and fix our CI and deployment systems so as to be able to use
the latest version of Debian Jessie with important CVE and other fixes.

Comment to get Mesos for Debian Jessie