Install REP with Docker [Linux]

To install REP it may take some time and courage since it depends on plenty of stuff. To get to your data analytics experience quickly we've loaded docker images with all the necessary packages, so you just need to pull and run them.

Basic usage:

1. install Docker,

2. start docker daemon (sudo service docker start), and make sure it is running (docker version)

3. make sure no application is using port 8888 (this is the default port that can be changed): (netstat -anl|grep 8888 should return an empty result)

4. run docker

   # by default, use notebooks folder in current dir.
   export NOTEBOOKS=$(pwd)/notebooks
   # for python3, use yandex/rep:0.6.6_py3
   export DOCKER_IMAGE=yandex/rep:0.6.6
   export PORT=8888
   docker run -i -t --rm --volume ${NOTEBOOKS}:/notebooks -p ${PORT}:8888 ${DOCKER_IMAGE}
   

5. open browser at localhost:8888 (or other port you used in $PORT)

Building container yourself

You can build container yourself after cloning repository

git clone https://github.com/yandex/rep.git --depth 1
cd rep
sudo make [-e NOTEBOOKS=~/notebooks][-e PYTHON=3][-e PORT=8000] run

• make stop -- stop running container
• make remove -- remove container
• make help -- more information on commands and arguments

Troubleshooting

• Diagnostics is saved in /notebooks/jupyter.log. So in case you've specified <my_dir> as an argument to make, it will be saved in that directory.
• make sure my_dir is accessible from under docker user (you might need check permissions or disable SELinux: setenforce 0)
• you need sudo to operate with docker. If you don't want to bother with it, just add your user to docker group (or dockerroot - depending on your Linux distribution) (e.g. usermod -G docker `whoami` )