irnidhi / ds-deploy-walkthru-iris

Simple walkthrough of Dockerized Data Science Dev + Deploy pipeline using Iris data

Getting Started

• If data exists on Local, copy it to the server with scp

  • If data sits on the cloud, download it inside the server

• Optional: Copy SSH keys scp -rp ~/.ssh user@remote-server:/tmp/.ssh

• SSH into the Server

• Pull a Docker Container from https://hub.docker.com/u/eadlab/dashboard/

docker pull <image>
docker images

• Get the code for your project if it exists, or clone this repository

git clone git-url
# git clone https://github.com/dushyantkhosla/ds-cookiecutter.git ./project-01

• Start a Docker Container using the appropriate image, map ports and mirror directories

docker run -it -v $pwd:/home \
               -p 8080:8080 \
               -p 5000:5000 \
               -p 3128:3128 \
               your-docker-image

• PS: If Docker commands do not work, run systemctl start docker as root and try again

• Configure Git

git config --global http.proxy http://proxy-url:port
git config --global user.email '[email protected]'
git config --global user.name 'FirstName LastName'

Directory Structure

./
DOCKERFILE      
Jenkinsfile     
README.md       
environment.yml

./data:
  interim                 
  processed 
  raw

./figures:

./models:

./notebooks:
  01-Obtain.ipynb    
  02-Scrub.ipynb     
  03-Explore.ipynb   
  04-Model.ipynb     
  05-iNterpret.ipynb

./references:

./reports:

./src:
  __init__.py 
  obtain      
  scrub       
  explore     
  model 
  interpret   
  
  tests